id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
sequence |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2506.13045 | Haiyang Guo | Haiyang Guo, Fanhu Zeng, Fei Zhu, Jiayi Wang, Xukai Wang, Jingang Zhou, Hongbo Zhao, Wenzhuo Liu, Shijie Ma, Da-Han Wang, Xu-Yao Zhang, Cheng-Lin Liu | A Comprehensive Survey on Continual Learning in Generative Models | Preprint | null | null | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rapid advancement of generative models has enabled modern AI systems to comprehend and produce highly sophisticated content, even achieving human-level performance in specific domains. However, these models remain fundamentally constrained by catastrophic forgetting - a persistent challenge where adapting to new tasks typically leads to significant degradation in performance on previously learned tasks. To address this practical limitation, numerous approaches have been proposed to enhance the adaptability and scalability of generative models in real-world applications. In this work, we present a comprehensive survey of continual learning methods for mainstream generative models, including large language models, multimodal large language models, vision language action models, and diffusion models. Drawing inspiration from the memory mechanisms of the human brain, we systematically categorize these approaches into three paradigms: architecture-based, regularization-based, and replay-based methods, while elucidating their underlying methodologies and motivations. We further analyze continual learning setups for different generative models, including training objectives, benchmarks, and core backbones, offering deeper insights into the field. The project page of this paper is available at https://github.com/Ghy0501/Awesome-Continual-Learning-in-Generative-Models. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 02:27:25 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 15:06:34 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Guo",
"Haiyang",
""
],
[
"Zeng",
"Fanhu",
""
],
[
"Zhu",
"Fei",
""
],
[
"Wang",
"Jiayi",
""
],
[
"Wang",
"Xukai",
""
],
[
"Zhou",
"Jingang",
""
],
[
"Zhao",
"Hongbo",
""
],
[
"Liu",
"Wenzhuo",
""
],
[
"Ma",
"Shijie",
""
],
[
"Wang",
"Da-Han",
""
],
[
"Zhang",
"Xu-Yao",
""
],
[
"Liu",
"Cheng-Lin",
""
]
] |
2506.13061 | Zhengjiang Lin | Daniel Zhengyu Huang, Jiaoyang Huang, Zhengjiang Lin | Fast Convergence for High-Order ODE Solvers in Diffusion Probabilistic Models | 64 pages, 7 figures | null | null | null | cs.LG cs.NA math.CA math.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Diffusion probabilistic models generate samples by learning to reverse a noise-injection process that transforms data into noise. Reformulating this reverse process as a deterministic probability flow ordinary differential equation (ODE) enables efficient sampling using high-order solvers, often requiring only $\mathcal{O}(10)$ steps. Since the score function is typically approximated by a neural network, analyzing the interaction between its regularity, approximation error, and numerical integration error is key to understanding the overall sampling accuracy. In this work, we continue our analysis of the convergence properties of the deterministic sampling methods derived from probability flow ODEs [25], focusing on $p$-th order (exponential) Runge-Kutta schemes for any integer $p \geq 1$. Under the assumption that the first and second derivatives of the approximate score function are bounded, we develop $p$-th order (exponential) Runge-Kutta schemes and demonstrate that the total variation distance between the target distribution and the generated data distribution can be bounded above by \begin{align*}
O\bigl(d^{\frac{7}{4}}\varepsilon_{\text{score}}^{\frac{1}{2}} +d(dH_{\max})^p\bigr), \end{align*} where $\varepsilon^2_{\text{score}}$ denotes the $L^2$ error in the score function approximation, $d$ is the data dimension and $H_{\max}$ represents the maximum step size used in the solver. We numerically verify the regularity assumption on benchmark datasets, confirming that the first and second derivatives of the approximate score function remain bounded in practice. Our theoretical guarantees hold for general forward processes with arbitrary variance schedules. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 03:09:25 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 14:18:09 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Huang",
"Daniel Zhengyu",
""
],
[
"Huang",
"Jiaoyang",
""
],
[
"Lin",
"Zhengjiang",
""
]
] |
2506.13097 | Ziqing Zhou | Ziqing Zhou, Bin-Bin Gao, Yurui Pan, Lidong Wang, Wenbing Zhu, Yong Liu, Jun Liu, Mingmin Chi, Dong Wu, Bo Peng, Chengjie Wang | Pro-AD: Learning Comprehensive Prototypes with Prototype-based Constraint for Multi-class Unsupervised Anomaly Detection | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Prototype-based reconstruction methods for unsupervised anomaly detection utilize a limited set of learnable prototypes which only aggregates insufficient normal information, resulting in undesirable reconstruction. However, increasing the number of prototypes may lead to anomalies being well reconstructed through the attention mechanism, which we refer to as the "Soft Identity Mapping" problem. In this paper, we propose Pro-AD to address these issues and fully utilize the prototypes to boost the performance of anomaly detection. Specifically, we first introduce an expanded set of learnable prototypes to provide sufficient capacity for semantic information. Then we employ a Dynamic Bidirectional Decoder which integrates the process of the normal information aggregation and the target feature reconstruction via prototypes, with the aim of allowing the prototypes to aggregate more comprehensive normal semantic information from different levels of the image features and the target feature reconstruction to not only utilize its contextual information but also dynamically leverage the learned comprehensive prototypes. Additionally, to prevent the anomalies from being well reconstructed using sufficient semantic information through the attention mechanism, Pro-AD introduces a Prototype-based Constraint that applied within the target feature reconstruction process of the decoder, which further improves the performance of our approach. Extensive experiments on multiple challenging benchmarks demonstrate that our Pro-AD achieve state-of-the-art performance, highlighting its superior robustness and practical effectiveness for Multi-class Unsupervised Anomaly Detection task. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 05:04:12 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 09:23:19 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Zhou",
"Ziqing",
""
],
[
"Gao",
"Bin-Bin",
""
],
[
"Pan",
"Yurui",
""
],
[
"Wang",
"Lidong",
""
],
[
"Zhu",
"Wenbing",
""
],
[
"Liu",
"Yong",
""
],
[
"Liu",
"Jun",
""
],
[
"Chi",
"Mingmin",
""
],
[
"Wu",
"Dong",
""
],
[
"Peng",
"Bo",
""
],
[
"Wang",
"Chengjie",
""
]
] |
2506.13180 | Jingjing Xu | Jingjing Xu, Zijian Yang, Albert Zeyer, Eugen Beck, Ralf Schlueter, Hermann Ney | Dynamic Acoustic Model Architecture Optimization in Training for ASR | Accepted by Interspeech 2025 | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Architecture design is inherently complex. Existing approaches rely on either handcrafted rules, which demand extensive empirical expertise, or automated methods like neural architecture search, which are computationally intensive. In this paper, we introduce DMAO, an architecture optimization framework that employs a grow-and-drop strategy to automatically reallocate parameters during training. This reallocation shifts resources from less-utilized areas to those parts of the model where they are most beneficial. Notably, DMAO only introduces negligible training overhead at a given model complexity. We evaluate DMAO through experiments with CTC on LibriSpeech, TED-LIUM-v2 and Switchboard datasets. The results show that, using the same amount of training resources, our proposed DMAO consistently improves WER by up to 6% relatively across various architectures, model sizes, and datasets. Furthermore, we analyze the pattern of parameter redistribution and uncover insightful findings. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 07:47:34 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 08:02:29 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Xu",
"Jingjing",
""
],
[
"Yang",
"Zijian",
""
],
[
"Zeyer",
"Albert",
""
],
[
"Beck",
"Eugen",
""
],
[
"Schlueter",
"Ralf",
""
],
[
"Ney",
"Hermann",
""
]
] |
2506.13244 | Francesco Emanuele Stradi | Francesco Emanuele Stradi, Matteo Castiglioni, Alberto Marchesi, Nicola Gatti, Christian Kroer | No-Regret Learning Under Adversarial Resource Constraints: A Spending Plan Is All You Need! | null | null | null | null | cs.LG cs.AI stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study online decision making problems under resource constraints, where both reward and cost functions are drawn from distributions that may change adversarially over time. We focus on two canonical settings: $(i)$ online resource allocation where rewards and costs are observed before action selection, and $(ii)$ online learning with resource constraints where they are observed after action selection, under full feedback or bandit feedback. It is well known that achieving sublinear regret in these settings is impossible when reward and cost distributions may change arbitrarily over time. To address this challenge, we analyze a framework in which the learner is guided by a spending plan--a sequence prescribing expected resource usage across rounds. We design general (primal-)dual methods that achieve sublinear regret with respect to baselines that follow the spending plan. Crucially, the performance of our algorithms improves when the spending plan ensures a well-balanced distribution of the budget across rounds. We additionally provide a robust variant of our methods to handle worst-case scenarios where the spending plan is highly imbalanced. To conclude, we study the regret of our algorithms when competing against benchmarks that deviate from the prescribed spending plan. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 08:42:31 GMT"
},
{
"version": "v2",
"created": "Tue, 17 Jun 2025 14:48:28 GMT"
},
{
"version": "v3",
"created": "Wed, 18 Jun 2025 14:04:08 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Stradi",
"Francesco Emanuele",
""
],
[
"Castiglioni",
"Matteo",
""
],
[
"Marchesi",
"Alberto",
""
],
[
"Gatti",
"Nicola",
""
],
[
"Kroer",
"Christian",
""
]
] |
2506.13288 | Mariaveronica De Angelis | Giovanni Montani, Mariaveronica De Angelis, Maria Giovanna Dainotti | Decay of dark energy into dark matter in a metric $f(R)$ gravity: effective running Hubble constant | 8 pages, 5 figures | Phys. Dark Univ. 49 (2025) 101969 | 10.1016/j.dark.2025.101969 | null | astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We examine a modified late-Universe dynamics where dark energy decays into dark matter, within the framework of metric $f(R)$-gravity in the Jordan frame. After a detailed analysis of the modified $\Lambda \text{CDM}$ model, we introduce a theoretical diagnostic tool to capture the emergence of an effective running Hubble constant as a function of redshift. We then compare this theoretical model with the 40-bin analysis of the Supernova Pantheon sample. This comparison allows us to determine the value of the additional free parameter that appears in our model, beyond those of the standard $\Lambda \text{CDM}$ model. Our modified late Universe dynamics provides a good-quality fit to the binned data, improving upon the previous phenomenological interpretation based on a power-law decay. However, unlike the power-law model, our approach cannot be extrapolated to the recombination redshift to match the Hubble constant measured by the Planck satellite. In fact, the dynamics resulting from the binned Pantheon sample analysis address only weakly the Hubble tension between the SH0ES and the Planck Collaboration values of the Hubble constant. Here we provide a convincing representation of the observed deviation of the cosmological dynamics from the $\Lambda$CDM-one, as it out-stands from the low redshift observed sources. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 09:30:34 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Montani",
"Giovanni",
""
],
[
"De Angelis",
"Mariaveronica",
""
],
[
"Dainotti",
"Maria Giovanna",
""
]
] |
2506.13300 | Bo Li | Bo Li, Chengben Xu, Wufeng Zhang | Seewo's Submission to MLC-SLM: Lessons learned from Speech Reasoning Language Models | null | null | null | null | cs.CL cs.AI cs.SD eess.AS | http://creativecommons.org/licenses/by/4.0/ | This paper presents Seewo's systems for both tracks of the Multilingual Conversational Speech Language Model Challenge (MLC-SLM), addressing automatic speech recognition (ASR) and speaker diarization with ASR (SD-ASR). We introduce a multi-stage training pipeline that explicitly enhances reasoning and self-correction in speech language models for ASR. Our approach combines curriculum learning for progressive capability acquisition, Chain-of-Thought data augmentation to foster intermediate reflection, and Reinforcement Learning with Verifiable Rewards (RLVR) to further refine self-correction through reward-driven optimization. This approach achieves substantial improvements over the official challenge baselines. On the evaluation set, our best system attains a WER/CER of 11.57% for Track 1 and a tcpWER/tcpCER of 17.67% for Track 2. Comprehensive ablation studies demonstrate the effectiveness of each component under challenge constraints. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 09:42:05 GMT"
},
{
"version": "v2",
"created": "Tue, 17 Jun 2025 08:15:17 GMT"
},
{
"version": "v3",
"created": "Wed, 18 Jun 2025 06:57:58 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Li",
"Bo",
""
],
[
"Xu",
"Chengben",
""
],
[
"Zhang",
"Wufeng",
""
]
] |
2506.13353 | Bartosz Ko{\l}odziejek | Piotr Graczyk and Bartosz Ko{\l}odziejek and Hideto Nakashima and Maciej Wilczy\'nski | From Graphical Lasso to Atomic Norms: High-Dimensional Pattern Recovery | 41 pages, 4 figures | null | null | null | math.ST stat.TH | http://creativecommons.org/licenses/by/4.0/ | Estimating high-dimensional precision matrices is a fundamental problem in modern statistics, with the graphical lasso and its $\ell_1$-penalty being a standard approach for recovering sparsity patterns. However, many statistical models, e.g. colored graphical models, exhibit richer structures like symmetry or equality constraints, which the $\ell_1$-norm cannot adequately capture. This paper addresses the gap by extending the high-dimensional analysis of pattern recovery to a general class of atomic norm penalties, particularly those whose unit balls are polytopes, where patterns correspond to the polytope's facial structure. We establish theoretical guarantees for recovering the true pattern induced by these general atomic norms in precision matrix estimation.
Our framework builds upon and refines the primal-dual witness methodology of Ravikumar et al. (2011). Our analysis provides conditions on the deviation between sample and true covariance matrices for successful pattern recovery, given a novel, generalized irrepresentability condition applicable to any atomic norm. When specialized to the $\ell_1$-penalty, our results offer improved conditions -- including weaker deviation requirements and a less restrictive irrepresentability condition -- leading to tighter bounds and better asymptotic performance than prior work. The proposed general irrepresentability condition, based on a new thresholding concept, provides a unified perspective on model selection consistency. Numerical examples demonstrate the tightness of the derived theoretical bounds. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 10:45:51 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 07:00:56 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Graczyk",
"Piotr",
""
],
[
"Kołodziejek",
"Bartosz",
""
],
[
"Nakashima",
"Hideto",
""
],
[
"Wilczyński",
"Maciej",
""
]
] |
2506.13366 | Yaxin Fan | Didi Zhang, Yaxin Fan, Peifeng Li, Qiaoming Zhu | Enhancing Goal-oriented Proactive Dialogue Systems via Consistency Reflection and Correction | Accepted by ACL'25 (main conference) | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Goal-oriented proactive dialogue systems are designed to guide user conversations seamlessly towards specific objectives by planning a goal-oriented path. However, previous research has focused predominantly on optimizing these paths while neglecting the inconsistencies that may arise between generated responses and dialogue contexts, including user profiles, dialogue history, domain knowledge, and subgoals. To address this issue, we introduce a model-agnostic two-stage Consistency Reflection and Correction (CRC) framework. Specifically, in the consistency reflection stage, the model is prompted to reflect on the discrepancies between generated responses and dialogue contexts, identifying inconsistencies and suggesting possible corrections. In the consistency correction stage, the model generates responses that are more consistent with the dialogue context based on these reflection results. We conducted experiments on various model architectures with different parameter sizes, including encoder-decoder models (BART, T5) and decoder-only models (GPT-2, DialoGPT, Phi3, Mistral and LLaMA3), and the experimental results on three datasets demonstrate that our CRC framework significantly improves the consistency between generated responses and dialogue contexts. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 11:15:21 GMT"
},
{
"version": "v2",
"created": "Tue, 17 Jun 2025 02:34:00 GMT"
},
{
"version": "v3",
"created": "Wed, 18 Jun 2025 06:17:35 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Zhang",
"Didi",
""
],
[
"Fan",
"Yaxin",
""
],
[
"Li",
"Peifeng",
""
],
[
"Zhu",
"Qiaoming",
""
]
] |
2506.13443 | Qiegen Liu | Kang Chen, Bin Huang, Xuebin Yang, Junyan Zhang, Qiegen Liu | PRO: Projection Domain Synthesis for CT Imaging | null | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Synthesizing high quality CT projection data remains a significant challenge due to the limited availability of annotated data and the complex nature of CT imaging. In this work, we present PRO, a projection domain synthesis foundation model for CT imaging. To the best of our knowledge, this is the first study that performs CT synthesis in the projection domain. Unlike previous approaches that operate in the image domain, PRO learns rich structural representations from raw projection data and leverages anatomical text prompts for controllable synthesis. This projection domain strategy enables more faithful modeling of underlying imaging physics and anatomical structures. Moreover, PRO functions as a foundation model, capable of generalizing across diverse downstream tasks by adjusting its generative behavior via prompt inputs. Experimental results demonstrated that incorporating our synthesized data significantly improves performance across multiple downstream tasks, including low-dose and sparse-view reconstruction. These findings underscore the versatility and scalability of PRO in data generation for various CT applications. These results highlight the potential of projection domain synthesis as a powerful tool for data augmentation and robust CT imaging. Our source code is publicly available at: https://github.com/yqx7150/PRO. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 12:57:29 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 07:33:50 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Chen",
"Kang",
""
],
[
"Huang",
"Bin",
""
],
[
"Yang",
"Xuebin",
""
],
[
"Zhang",
"Junyan",
""
],
[
"Liu",
"Qiegen",
""
]
] |
2506.13496 | Kshitij Kavimandan | Kshitij Kavimandan, Angelos Nalmpantis, Emma Beauxis-Aussalet, Robert-Jan Sips | Hierarchical Multi-Positive Contrastive Learning for Patent Image Retrieval | 5 pages, 3 figures, Accepted as a short paper at the 6th Workshop on Patent Text Mining and Semantic Technologies (PatentSemTech 2025), co-located with SIGIR 2025 | null | null | null | cs.CV cs.IR cs.LG | http://creativecommons.org/licenses/by/4.0/ | Patent images are technical drawings that convey information about a patent's innovation. Patent image retrieval systems aim to search in vast collections and retrieve the most relevant images. Despite recent advances in information retrieval, patent images still pose significant challenges due to their technical intricacies and complex semantic information, requiring efficient fine-tuning for domain adaptation. Current methods neglect patents' hierarchical relationships, such as those defined by the Locarno International Classification (LIC) system, which groups broad categories (e.g., "furnishing") into subclasses (e.g., "seats" and "beds") and further into specific patent designs. In this work, we introduce a hierarchical multi-positive contrastive loss that leverages the LIC's taxonomy to induce such relations in the retrieval process. Our approach assigns multiple positive pairs to each patent image within a batch, with varying similarity scores based on the hierarchical taxonomy. Our experimental analysis with various vision and multimodal models on the DeepPatent2 dataset shows that the proposed method enhances the retrieval results. Notably, our method is effective with low-parameter models, which require fewer computational resources and can be deployed on environments with limited hardware. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 13:53:02 GMT"
},
{
"version": "v2",
"created": "Tue, 17 Jun 2025 19:39:04 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Kavimandan",
"Kshitij",
""
],
[
"Nalmpantis",
"Angelos",
""
],
[
"Beauxis-Aussalet",
"Emma",
""
],
[
"Sips",
"Robert-Jan",
""
]
] |
2506.13538 | Mohammed Mehedi Hasan | Mohammed Mehedi Hasan and Hao Li and Emad Fallahzadeh and Gopi Krishnan Rajbahadur and Bram Adams and Ahmed E. Hassan | Model Context Protocol (MCP) at First Glance: Studying the Security and Maintainability of MCP Servers | null | null | null | null | cs.SE cs.ET | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Although Foundation Models (FMs), such as GPT-4, are increasingly used in domains like finance and software engineering, reliance on textual interfaces limits these models' real-world interaction. To address this, FM providers introduced tool calling-triggering a proliferation of frameworks with distinct tool interfaces. In late 2024, Anthropic introduced the Model Context Protocol (MCP) to standardize this tool ecosystem, which has become the de facto standard with over eight million weekly SDK downloads. Despite its adoption, MCP's AI-driven, non-deterministic control flow introduces new risks to sustainability, security, and maintainability, warranting closer examination.
Towards this end, we present the first large-scale empirical study of MCP servers. Using state-of-the-art health metrics and a hybrid analysis pipeline, combining a general-purpose static analysis tool with an MCP-specific scanner, we evaluate 1,899 open-source MCP servers to assess their health, security, and maintainability. Despite MCP servers demonstrating strong health metrics, we identify eight distinct vulnerabilities -- only three overlapping with traditional software vulnerabilities. Additionally, 7.2% of servers contain general vulnerabilities and 5.5% exhibit MCP-specific tool poisoning. Regarding maintainability, while 66% exhibit code smells, 14.4% contain ten bug patterns overlapping with traditional open-source software projects. These findings highlight the need for MCP-specific vulnerability detection techniques while reaffirming the value of traditional analysis and refactoring practices. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 14:26:37 GMT"
},
{
"version": "v2",
"created": "Tue, 17 Jun 2025 04:11:24 GMT"
},
{
"version": "v3",
"created": "Wed, 18 Jun 2025 15:02:53 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Hasan",
"Mohammed Mehedi",
""
],
[
"Li",
"Hao",
""
],
[
"Fallahzadeh",
"Emad",
""
],
[
"Rajbahadur",
"Gopi Krishnan",
""
],
[
"Adams",
"Bram",
""
],
[
"Hassan",
"Ahmed E.",
""
]
] |
2506.13584 | Daniel Anadria | Daniel Anadria, Roel Dobbe, Anastasia Giachanou, Ruurd Kuiper, Richard Bartels, Wouter van Amsterdam, \'I\~nigo Mart\'inez de Rituerto de Troya, Carmen Z\"urcher, Daniel Oberski | From Data-Driven to Purpose-Driven Artificial Intelligence: Systems Thinking for Data-Analytic Automation of Patient Care | The work is under review at ACM Health | null | null | null | cs.AI cs.LG cs.SY eess.SY math.ST stat.ME stat.TH | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In this work, we reflect on the data-driven modeling paradigm that is gaining ground in AI-driven automation of patient care. We argue that the repurposing of existing real-world patient datasets for machine learning may not always represent an optimal approach to model development as it could lead to undesirable outcomes in patient care. We reflect on the history of data analysis to explain how the data-driven paradigm rose to popularity, and we envision ways in which systems thinking and clinical domain theory could complement the existing model development approaches in reaching human-centric outcomes. We call for a purpose-driven machine learning paradigm that is grounded in clinical theory and the sociotechnical realities of real-world operational contexts. We argue that understanding the utility of existing patient datasets requires looking in two directions: upstream towards the data generation, and downstream towards the automation objectives. This purpose-driven perspective to AI system development opens up new methodological opportunities and holds promise for AI automation of patient care. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 15:07:44 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 08:20:25 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Anadria",
"Daniel",
""
],
[
"Dobbe",
"Roel",
""
],
[
"Giachanou",
"Anastasia",
""
],
[
"Kuiper",
"Ruurd",
""
],
[
"Bartels",
"Richard",
""
],
[
"van Amsterdam",
"Wouter",
""
],
[
"de Troya",
"Íñigo Martínez de Rituerto",
""
],
[
"Zürcher",
"Carmen",
""
],
[
"Oberski",
"Daniel",
""
]
] |
2506.13588 | Chavdar Dutsov | Chavdar Dutsov, Timothy Hume, Maxim Pospelov and Philipp Schmidt-Wellenburg | Experimental search for electric dipole moments of light radioactive nuclei | to be submitted to Physical Review X | null | null | null | hep-ex hep-ph | http://creativecommons.org/licenses/by/4.0/ | We discuss a search for the electric dipole moment (EDM) of a light beta-radioactive ion using a compact ion trap by adapting the "frozen-spin" method. The measurement will be done on ions stripped of their valence electrons, thereby bypassing the significant Schiff screening that hinders the application of successful contemporary EDM searches using heavy neutral atoms and molecules to light nuclei. We identified $^8$Li as the most promising candidate for a proof-of-concept EDM search and we estimate that the current indirect proton EDM limit of a few $10^{-25} e\cdot$cm set by $^{199}$Hg measurements can be surpassed with a week of measurement time at existing facilities. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 15:16:08 GMT"
},
{
"version": "v2",
"created": "Tue, 17 Jun 2025 20:49:21 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Dutsov",
"Chavdar",
""
],
[
"Hume",
"Timothy",
""
],
[
"Pospelov",
"Maxim",
""
],
[
"Schmidt-Wellenburg",
"Philipp",
""
]
] |
2506.13589 | Zhucun Xue | Zhucun Xue, Jiangning Zhang, Xurong Xie, Yuxuan Cai, Yong Liu, Xiangtai Li, Dacheng Tao | AdaVideoRAG: Omni-Contextual Adaptive Retrieval-Augmented Efficient Long Video Understanding | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multimodal Large Language Models (MLLMs) struggle with long videos due to fixed context windows and weak long-term dependency modeling. Existing Retrieval-Augmented Generation (RAG) methods for videos use static retrieval strategies, leading to inefficiencies for simple queries and information loss for complex tasks. To address this, we propose AdaVideoRAG, a novel framework that dynamically adapts retrieval granularity based on query complexity using a lightweight intent classifier. Our framework employs an Omni-Knowledge Indexing module to build hierarchical databases from text (captions, ASR, OCR), visual features, and semantic graphs, enabling optimal resource allocation across tasks. We also introduce the HiVU benchmark for comprehensive evaluation. Experiments demonstrate improved efficiency and accuracy for long-video understanding, with seamless integration into existing MLLMs. AdaVideoRAG establishes a new paradigm for adaptive retrieval in video analysis. Codes will be open-sourced at https://github.com/xzc-zju/AdaVideoRAG. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 15:18:15 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 02:46:20 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Xue",
"Zhucun",
""
],
[
"Zhang",
"Jiangning",
""
],
[
"Xie",
"Xurong",
""
],
[
"Cai",
"Yuxuan",
""
],
[
"Liu",
"Yong",
""
],
[
"Li",
"Xiangtai",
""
],
[
"Tao",
"Dacheng",
""
]
] |
2506.13678 | Yi Wang | Yi Wang, Zhenghong Wang, Fan Zhang, Chengling Tang, Chaogui Kang, Di Zhu, Zhongfu Ma, Sijie Ruan, Weiyu Zhang, Yu Zheng, Philip S. Yu, Yu Liu | A Gravity-informed Spatiotemporal Transformer for Human Activity Intensity Prediction | 18 pages, 13 figures, under review | null | null | null | cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Human activity intensity prediction is a crucial to many location-based services. Although tremendous progress has been made to model dynamic spatiotemporal patterns of human activity, most existing methods, including spatiotemporal graph neural networks (ST-GNNs), overlook physical constraints of spatial interactions and the over-smoothing phenomenon in spatial correlation modeling. To address these limitations, this work proposes a physics-informed deep learning framework, namely Gravity-informed Spatiotemporal Transformer (Gravityformer) by refining transformer attention to integrate the universal law of gravitation and explicitly incorporating constraints from spatial interactions. Specifically, it (1) estimates two spatially explicit mass parameters based on inflow and outflow, (2) models the likelihood of cross-unit interaction using closed-form solutions of spatial interactions to constrain spatial modeling randomness, and (3) utilizes the learned spatial interaction to guide and mitigate the over-smoothing phenomenon in transformer attention matrices. The underlying law of human activity can be explicitly modeled by the proposed adaptive gravity model. Moreover, a parallel spatiotemporal graph convolution transformer structure is proposed for achieving a balance between coupled spatial and temporal learning. Systematic experiments on six real-world large-scale activity datasets demonstrate the quantitative and qualitative superiority of our approach over state-of-the-art benchmarks. Additionally, the learned gravity attention matrix can be disentangled and interpreted based on geographical laws. This work provides a novel insight into integrating physical laws with deep learning for spatiotemporal predictive learning. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 16:32:51 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 16:04:08 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Wang",
"Yi",
""
],
[
"Wang",
"Zhenghong",
""
],
[
"Zhang",
"Fan",
""
],
[
"Tang",
"Chengling",
""
],
[
"Kang",
"Chaogui",
""
],
[
"Zhu",
"Di",
""
],
[
"Ma",
"Zhongfu",
""
],
[
"Ruan",
"Sijie",
""
],
[
"Zhang",
"Weiyu",
""
],
[
"Zheng",
"Yu",
""
],
[
"Yu",
"Philip S.",
""
],
[
"Liu",
"Yu",
""
]
] |
2506.13701 | Cristiano Longarini | Cristiano Longarini, Daniel Price, Kaitlin M. Kratter, Giuseppe Lodato and Cathie J. Clarke | Infall-driven gravitational instability in accretion discs | Accepted for publication in MNRAS | null | null | null | astro-ph.EP | http://creativecommons.org/licenses/by/4.0/ | Gravitational instability (GI) is typically studied in cooling-dominated discs, often modelled using simplified prescriptions such as $\beta$-cooling. In this paper, we investigate the onset and evolution of GI in accretion discs subject to continuous mass injection, combining 1D and 3D numerical simulations. We explore an alternative self-regulation mechanism in which mass replenishment drives the system toward marginal stability $Q\sim 1$. In this regime, the disc establishes a steady-state disc-to-star mass ratio, balancing the mass transported to the central object with that added to the disc. Our 3D simulations reveal that the general scaling predicted from the linear theory are respected, however there are important difference compared to the cooling case in terms of morphology and pattern speed. Unlike the flocculent spirals seen in cooling-driven instability, the power is concentrated towards the dominant modes in infall-driven spirals. Additionally, spiral waves generate at the mass injection location, and propagate at constant pattern speed, unlike in the cooling case. This suggests a fundamental difference in how mass-regulated and cooling-regulated discs behave and transport angular momentum. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 17:05:42 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 15:09:14 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Longarini",
"Cristiano",
""
],
[
"Price",
"Daniel",
""
],
[
"Kratter",
"Kaitlin M.",
""
],
[
"Lodato",
"Giuseppe",
""
],
[
"Clarke",
"Cathie J.",
""
]
] |
2506.13776 | Kevin Wei | Kevin L. Wei, Patricia Paskov, Sunishchal Dev, Michael J. Byun, Anka Reuel, Xavier Roberts-Gaal, Rachel Calcott, Evie Coxon, Chinmay Deshpande | Recommendations and Reporting Checklist for Rigorous & Transparent Human Baselines in Model Evaluations | A version of this paper has been accepted to ICML 2025 as a position paper (spotlight), with the title: "Position: Human Baselines in Model Evaluations Need Rigor and Transparency (With Recommendations & Reporting Checklist)." | null | null | null | cs.AI cs.CY cs.HC | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In this position paper, we argue that human baselines in foundation model evaluations must be more rigorous and more transparent to enable meaningful comparisons of human vs. AI performance, and we provide recommendations and a reporting checklist towards this end. Human performance baselines are vital for the machine learning community, downstream users, and policymakers to interpret AI evaluations. Models are often claimed to achieve "super-human" performance, but existing baselining methods are neither sufficiently rigorous nor sufficiently well-documented to robustly measure and assess performance differences. Based on a meta-review of the measurement theory and AI evaluation literatures, we derive a framework with recommendations for designing, executing, and reporting human baselines. We synthesize our recommendations into a checklist that we use to systematically review 115 human baselines (studies) in foundation model evaluations and thus identify shortcomings in existing baselining methods; our checklist can also assist researchers in conducting human baselines and reporting results. We hope our work can advance more rigorous AI evaluation practices that can better serve both the research community and policymakers. Data is available at: https://github.com/kevinlwei/human-baselines | [
{
"version": "v1",
"created": "Mon, 09 Jun 2025 04:08:16 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Wei",
"Kevin L.",
""
],
[
"Paskov",
"Patricia",
""
],
[
"Dev",
"Sunishchal",
""
],
[
"Byun",
"Michael J.",
""
],
[
"Reuel",
"Anka",
""
],
[
"Roberts-Gaal",
"Xavier",
""
],
[
"Calcott",
"Rachel",
""
],
[
"Coxon",
"Evie",
""
],
[
"Deshpande",
"Chinmay",
""
]
] |
2506.13790 | Tapio Pitk\"aranta Mr | Tapio Pitk\"aranta | The NordDRG AI Benchmark for Large Language Models | 15 pages, 4 figures | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) are already being piloted for clinical coding and decision support. However, until now, no open benchmark has targeted the hospital funding layer where Diagnosis-Related Groups (DRG) determine reimbursement across many countries. We release NordDRG-AI-Benchmark, the first public test-bed that captures a complete DRG rule set and evaluates an LLM's ability to reason over multilingual diagnosis, procedure, and tariff logic.
The benchmark bundles three classes of artefacts: (i) definition tables with 20 interlinked tables covering DRG logic, ICD and NCSP codes, age/sex splits, and country flags; (ii) expert manuals and changelog templates describing real governance workflows; and (iii) a prompt pack of 14 CaseMix tasks that span code lookup, cross-table inference, multilingual terminology, and quality-assurance audits.
All artefacts are available at: https://github.com/longshoreforrest/norddrg-ai-benchmark
A baseline demonstration shows that five state-of-the-art LLMs perform very differently on the nine automatically verifiable tasks: o3 (OpenAI) scores 9 out of 9, GPT-4o and o4-mini-high score 7 out of 9, while Gemini 2.5 Pro and Gemini 2.5 Flash solve only 5 out of 9 and 3 out of 9, respectively. These results confirm that NordDRG-AI-Benchmark highlights domain-specific strengths and weaknesses that remain hidden in generic LLM benchmarks, offering a reproducible baseline for research on trustworthy automation in hospital funding. | [
{
"version": "v1",
"created": "Wed, 11 Jun 2025 11:40:11 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 01:54:22 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Pitkäranta",
"Tapio",
""
]
] |
2506.13801 | Laurentiu Leustean | Lauren\c{t}iu Leu\c{s}tean and Dafina Trufa\c{s} | Matching logic -- proof system $\mathcal{G}^c$ | null | null | null | null | cs.LO math.LO | http://creativecommons.org/licenses/by/4.0/ | We propose in these notes a new proof system for first-order matching logic with application, obtained by adapting to matching logic G\"{o}del's proof system for first-order intuitionistic logic. | [
{
"version": "v1",
"created": "Fri, 13 Jun 2025 11:09:36 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 16:04:44 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Leuştean",
"Laurenţiu",
""
],
[
"Trufaş",
"Dafina",
""
]
] |
2506.13832 | Jingzhe Ding | Hongda Zhu, Yiwen Zhang, Bing Zhao, Jingzhe Ding, Siyao Liu, Tong Liu, Dandan Wang, Yanan Liu, Zhaojian Li | FrontendBench: A Benchmark for Evaluating LLMs on Front-End Development via Automatic Evaluation | null | null | null | null | cs.SE cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Large Language Models (LLMs) have made significant strides in front-end code generation. However, existing benchmarks exhibit several critical limitations: many tasks are overly simplistic, test cases often lack rigor, and end-to-end validation is absent. These issues hinder the accurate assessment of model performance. To address these challenges, we present FrontendBench, a benchmark co-developed by humans and LLMs. FrontendBench categorizes tasks based on code functionality and incorporates interactive test scenarios, enabling a more comprehensive and practical evaluation of front-end code generation capabilities. The benchmark comprises 148 meticulously crafted prompt-test case pairs spanning five levels of web components, from basic UI elements to complex interactive features. Each task reflects realistic front-end development challenges. Furthermore, we introduce an automatic evaluation framework that executes generated code within a sandbox environment and assesses outcomes using predefined test scripts. This framework achieves a 90.54% agreement rate with expert human evaluations, demonstrating high reliability. We benchmark several state-of-the-art LLMs on FrontendBench and observe substantial performance disparities in handling real-world front-end tasks. These results highlight FrontendBench as a reliable and scalable benchmark, supporting consistent multimodal evaluation and providing a robust foundation for future research in front-end code generation. Our data and code will be released soon. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 03:20:31 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 13:10:14 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Zhu",
"Hongda",
""
],
[
"Zhang",
"Yiwen",
""
],
[
"Zhao",
"Bing",
""
],
[
"Ding",
"Jingzhe",
""
],
[
"Liu",
"Siyao",
""
],
[
"Liu",
"Tong",
""
],
[
"Wang",
"Dandan",
""
],
[
"Liu",
"Yanan",
""
],
[
"Li",
"Zhaojian",
""
]
] |
2506.13913 | Helder Rojas | Helder Rojas | Stochastic Processes and Diffusion Equations | null | null | null | null | math.PR | http://creativecommons.org/licenses/by/4.0/ | In these lecture notes, we explore the mathematical preliminaries and foundational concepts that connect stochastic processes with partial differential equations. We begin by investigating Brownian motion, which serves as a model for random fluctuations and is deeply connected to the heat equation. This connection forms the basis for understanding diffusion phenomena, where the probability distribution of Brownian motion evolves according to the heat equation over time. To extend this classical result to more general stochastic systems, we introduce the It\^o calculus, a powerful framework that allows us to analyze processes driven by both deterministic drift and stochastic fluctuations. This mathematical tool is essential for understanding the dynamics of more complex diffusion processes, where randomness is no longer purely Brownian, but also depends on the underlying system's state. Building on these concepts, we turn to the study of diffusion processes, which generalize Brownian motion by incorporating the Fokker-Planck equation. This equation describes how the probability density of a diffusion process evolves over time and serves as an extension of the heat equation to more complex stochastic systems. By using It\^o calculus, we can rigorously study how these processes behave and connect the microscopic randomness of individual particles to their macroscopic description via partial differential equations. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 18:48:16 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 04:18:48 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Rojas",
"Helder",
""
]
] |
2506.13914 | Yingchen Liu | Yingchen Liu, Haixin Jiang | The Impact of Social Isolation on Subjective Cognitive Decline in Older Adults: A Study Based on Network Analysis and Longitudinal Model | 29 pages, 5 figures, 2 tables | null | null | null | stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Social isolation (SI) in older adults has emerged as a critical mental health concern, with established links to cognitive decline. While depression is hypothesized to mediate this relationship, the longitudinal mechanisms remain unclear. This study employed network analysis and cross-lagged modeling to examine these relationships during pandemic-related social restrictions. We collected data from 1,230 older adults (Mage=64.49, SD=3.84) across three timepoints (during lockdown, immediately post-lockdown, and 6 months later). Network analysis identified depressive symptoms (particularly PHQ-9 item 9) as central nodes bridging SI and subjective cognitive decline (SCD). Longitudinal analyses revealed: 1) T1 SI predicted T2 depression; 2) T2 depression predicted T3 SCD. These patterns held for both online and offline SI, with excellent model fit. Our findings demonstrate depression's mediating role in the "SI-depression-SCD" pathway, highlighting how reduced social connections may gradually impair cognitive self-assessment through depressive symptoms. The results underscore the universal mental health impact of SI, regardless of interaction modality (virtual/in-person). These insights suggest that combined interventions targeting both social connection and depression management may be particularly effective for mitigating age-related cognitive decline. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 18:49:05 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 09:54:56 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Liu",
"Yingchen",
""
],
[
"Jiang",
"Haixin",
""
]
] |
2506.13916 | Arturo Jaramillo | Isaias Banales, Arturo Jaramillo, Heli Ricalde Guerrero | Branching Stein Variational Gradient Descent for sampling multimodal distributions | null | null | null | null | cs.LG stat.CO stat.ML | http://creativecommons.org/licenses/by/4.0/ | We propose a novel particle-based variational inference method designed to work with multimodal distributions. Our approach, referred to as Branched Stein Variational Gradient Descent (BSVGD), extends the classical Stein Variational Gradient Descent (SVGD) algorithm by incorporating a random branching mechanism that encourages the exploration of the state space. In this work, a theoretical guarantee for the convergence in distribution is presented, as well as numerical experiments to validate the suitability of our algorithm. Performance comparisons between the BSVGD and the SVGD are presented using the Wasserstein distance between samples and the corresponding computational times. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 18:51:44 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Banales",
"Isaias",
""
],
[
"Jaramillo",
"Arturo",
""
],
[
"Guerrero",
"Heli Ricalde",
""
]
] |
2506.13926 | Sergei A. Grebenev | G. Yu. Mozgunov (1), A. S. Pozanenko (1,2), P. Yu. Minaev (1), I. V. Chelovekov (1), S. A. Grebenev (1), A. G. Demin (3), A. V. Ridnaya (3), D. S. Svinkin (3), Yu. R. Temiraev (4), and D. D. Frederiks (3) ((1) Space Research Institute, Moscow, (2) National Research University `Higher School of Economics', Moscow, (3) Ioffe Institute, St. Petersburg, (4) GlowByte Company, Moscow) | Search for Astrophysical Transients on Limiting Time Scales and Their Classification Based on INTEGRAL Data | 22 pages, 23 figures, 5 tables | Astronomy Letters, 50 (2024), 756-779 | 10.1134/S1063773725700136 | null | astro-ph.HE astro-ph.IM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We have searched for ultra-long (> 100 s) gamma-ray transients in the data from the anticoincidence shield (ACS) of the SPI gamma-ray spectrometer onboard the INTEGRAL orbital observatory and classified them by machine learning methods. We have found about 4364 candidates for such events in the SPI-ACS data by the `blind' threshold search method. We have developed an algorithm for automatic processing of their light curves that distinguishes a candidate for transients on various time scales and allows its duration and fluence to be determined. The algorithm has been applied to calculate (and compare) the fluxes in the light curves recorded by various INTEGRAL detectors: IREM, SPI-ACS, SPI, ISGRI, and PICsIT. These fluxes have been used to train the classifier based on gradient boosting. Subsequently, we have performed a cluster analysis of the candidates found by the dimensionality reduction and clustering methods. In conclusion we have compared the remaining candidates with the data from the Konus-WIND gamma-ray detectors. Thus, we have confirmed 16 candidates for astrophysical transients, including four candidates for ultra-long gamma-ray bursts from the events detected by the SPI-ACS detector. Out of the probable events, but unconfirmed by other experiments, up to 270 events can be classified as real gamma-ray bursts. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 19:05:47 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Mozgunov",
"G. Yu.",
"",
"Space Research Institute, Moscow"
],
[
"Pozanenko",
"A. S.",
"",
"Space Research Institute, Moscow",
"National Research University `Higher School of Economics', Moscow"
],
[
"Minaev",
"P. Yu.",
"",
"Space Research Institute, Moscow"
],
[
"Chelovekov",
"I. V.",
"",
"Space Research Institute, Moscow"
],
[
"Grebenev",
"S. A.",
"",
"Space Research Institute, Moscow"
],
[
"Demin",
"A. G.",
"",
"Ioffe Institute, St. Petersburg"
],
[
"Ridnaya",
"A. V.",
"",
"Ioffe Institute, St. Petersburg"
],
[
"Svinkin",
"D. S.",
"",
"Ioffe Institute, St. Petersburg"
],
[
"Temiraev",
"Yu. R.",
"",
"GlowByte Company, Moscow"
],
[
"Frederiks",
"D. D.",
"",
"Ioffe Institute, St. Petersburg"
]
] |
2506.13973 | Harrison Katz | Harrison Katz, Liz Medina, Robert E. Weiss | Sensitivity Analysis of Priors in the Bayesian Dirichlet Auto-Regressive Moving Average Model | null | null | null | null | stat.ME stat.AP | http://creativecommons.org/licenses/by/4.0/ | Prior choice can strongly influence Bayesian Dirichlet ARMA (B-DARMA) inference for compositional time-series. Using simulations with (i) correct lag order, (ii) overfitting, and (iii) underfitting, we assess five priors: weakly-informative, horseshoe, Laplace, mixture-of-normals, and hierarchical. With the true lag order, all priors achieve comparable RMSE, though horseshoe and hierarchical slightly reduce bias. Under overfitting, aggressive shrinkage-especially the horseshoe-suppresses noise and improves forecasts, yet no prior rescues a model that omits essential VAR or VMA terms.
We then fit B-DARMA to daily SP 500 sector weights using an intentionally large lag structure. Shrinkage priors curb spurious dynamics, whereas weakly-informative priors magnify errors in volatile sectors. Two lessons emerge: (1) match shrinkage strength to the degree of overparameterization, and (2) prioritize correct lag selection, because no prior repairs structural misspecification. These insights guide prior selection and model complexity management in high-dimensional compositional time-series applications. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 20:26:29 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 15:47:17 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Katz",
"Harrison",
""
],
[
"Medina",
"Liz",
""
],
[
"Weiss",
"Robert E.",
""
]
] |
2506.13998 | Michael Anoprenko | Michael Anoprenko, Andrei Tonkikh, Alexander Spiegelman, Petr Kuznetsov | DAGs for the Masses | null | null | null | null | cs.DC | http://creativecommons.org/licenses/by/4.0/ | A recent approach to building consensus protocols on top of Directed Acyclic Graphs (DAGs) shows much promise due to its simplicity and stable throughput. However, as each node in the DAG typically includes a linear number of references to the nodes in the previous round, prior DAG protocols only scale up to a certain point when the overhead of maintaining the graph becomes the bottleneck.
To enable large-scale deployments of DAG-based protocols, we propose a sparse DAG architecture, where each node includes only a constant number of references to random nodes in the previous round. We present a sparse version of Bullshark -- one of the most prominent DAG-based consensus protocols -- and demonstrate its improved scalability.
Remarkably, unlike other protocols that use random sampling to reduce communication complexity, we manage to avoid sacrificing resilience: the protocol can tolerate up to $f<n/3$ Byzantine faults (where $n$ is the number of participants), same as its less scalable deterministic counterpart. The proposed ``sparse'' methodology can be applied to any protocol that maintains disseminated system updates and causal relations between them in a graph-like structure. Our simulations show that the considerable reduction of transmitted metadata in sparse DAGs results in more efficient network utilization and better scalability. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 20:55:30 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 15:16:15 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Anoprenko",
"Michael",
""
],
[
"Tonkikh",
"Andrei",
""
],
[
"Spiegelman",
"Alexander",
""
],
[
"Kuznetsov",
"Petr",
""
]
] |
2506.14036 | Jaesung Lee | Tatthapong Srikitrungruang, Matthew Lemon, Sina Aghaee Dabaghan Fard, Jaesung Lee, Yuxiao Zhou | Robust Physics-Informed Neural Network Approach for Estimating Heterogeneous Elastic Properties from Noisy Displacement Data | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Accurately estimating spatially heterogeneous elasticity parameters, particularly Young's modulus and Poisson's ratio, from noisy displacement measurements remains significantly challenging in inverse elasticity problems. Existing inverse estimation techniques are often limited by instability, pronounced sensitivity to measurement noise, and difficulty in recovering absolute-scale Young's modulus. This work presents a novel Inverse Elasticity Physics-Informed Neural Network (IE-PINN) specifically designed to robustly reconstruct heterogeneous distributions of elasticity parameters from noisy displacement data based on linear elasticity physics. IE-PINN integrates three distinct neural network architectures dedicated to separately modeling displacement fields, strain fields, and elasticity distributions, thereby significantly enhancing stability and accuracy against measurement noise. Additionally, a two-phase estimation strategy is introduced: the first phase recovers relative spatial distributions of Young's modulus and Poisson's ratio, and the second phase calibrates the absolute scale of Young's modulus using imposed loading boundary conditions. Additional methodological innovations, including positional encoding, sine activation functions, and a sequential pretraining protocol, further enhance the model's performance and robustness. Extensive numerical experiments demonstrate that IE-PINN effectively overcomes critical limitations encountered by existing methods, delivering accurate absolute-scale elasticity estimations even under severe noise conditions. This advancement holds substantial potential for clinical imaging diagnostics and mechanical characterization, where measurements typically encounter substantial noise. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 22:20:44 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 17:08:53 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Srikitrungruang",
"Tatthapong",
""
],
[
"Lemon",
"Matthew",
""
],
[
"Fard",
"Sina Aghaee Dabaghan",
""
],
[
"Lee",
"Jaesung",
""
],
[
"Zhou",
"Yuxiao",
""
]
] |
2506.14048 | Madox McGrae-Menge | Madox C. McGrae-Menge, Jacob R. Pierce, Frederico Fiuza, E. Paulo Alves | Embedding physical symmetries into machine-learned reduced plasma physics models via data augmentation | Corrected Figure 1 | null | null | null | physics.plasm-ph | http://creativecommons.org/licenses/by/4.0/ | Machine learning is offering powerful new tools for the development and discovery of reduced models of nonlinear, multiscale plasma dynamics from the data of first-principles kinetic simulations. However, ensuring the physical consistency of such models requires embedding fundamental symmetries of plasma dynamics. In this work, we explore a symmetry-embedding strategy based on data augmentation, where symmetry-preserving transformations (e.g., Lorentz and Galilean boosts) are applied to simulation data. Using both sparse regression and neural networks, we show that models trained on symmetry-augmented data more accurately infer the plasma fluid equations and pressure tensor closures from fully kinetic particle-in-cell simulations of magnetic reconnection. We show that this approach suppresses spurious inertial-frame-dependent correlations between dynamical variables, improves data efficiency, and significantly outperforms models trained without symmetry-augmented data, as well as commonly used theoretical pressure closure models. Our results establish symmetry-based data augmentation as a broadly applicable method for incorporating physical structure into machine-learned reduced plasma models. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 22:53:36 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 17:26:58 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"McGrae-Menge",
"Madox C.",
""
],
[
"Pierce",
"Jacob R.",
""
],
[
"Fiuza",
"Frederico",
""
],
[
"Alves",
"E. Paulo",
""
]
] |
2506.14061 | Congzhou Sha | Congzhou M Sha | Mitigating collusion in Gold Humanism Honor Society nominations | 18 pages, 6 figures | null | null | null | physics.soc-ph math.PR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | An anecdotally common complaint regarding induction into the Gold Humanism Honor Society is the bias toward close friends during the initial nomination process. In this work, we numerically simulate the nomination process under different assumptions, demonstrate that collusion can be detected, and propose a simple strategy to correct for bias in the nomination process. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 23:44:23 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 01:03:37 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Sha",
"Congzhou M",
""
]
] |
2506.14095 | Parikshit Ram | Parikshit Ram, Kenneth L. Clarkson, Tim Klinger, Shashanka Ubaru, Alexander G. Gray | Transformers Learn Faster with Semantic Focus | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Various forms of sparse attention have been explored to mitigate the quadratic computational and memory cost of the attention mechanism in transformers. We study sparse transformers not through a lens of efficiency but rather in terms of learnability and generalization. Empirically studying a range of attention mechanisms, we find that input-dependent sparse attention models appear to converge faster and generalize better than standard attention models, while input-agnostic sparse attention models show no such benefits -- a phenomenon that is robust across architectural and optimization hyperparameter choices. This can be interpreted as demonstrating that concentrating a model's "semantic focus" with respect to the tokens currently being considered (in the form of input-dependent sparse attention) accelerates learning. We develop a theoretical characterization of the conditions that explain this behavior. We establish a connection between the stability of the standard softmax and the loss function's Lipschitz properties, then show how sparsity affects the stability of the softmax and the subsequent convergence and generalization guarantees resulting from the attention mechanism. This allows us to theoretically establish that input-agnostic sparse attention does not provide any benefits. We also characterize conditions when semantic focus (input-dependent sparse attention) can provide improved guarantees, and we validate that these conditions are in fact met in our empirical evaluations. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 01:19:28 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 12:59:54 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Ram",
"Parikshit",
""
],
[
"Clarkson",
"Kenneth L.",
""
],
[
"Klinger",
"Tim",
""
],
[
"Ubaru",
"Shashanka",
""
],
[
"Gray",
"Alexander G.",
""
]
] |
2506.14168 | Hu Yu | Hu Yu, Biao Gong, Hangjie Yuan, DanDan Zheng, Weilong Chai, Jingdong Chen, Kecheng Zheng, Feng Zhao | VideoMAR: Autoregressive Video Generatio with Continuous Tokens | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Masked-based autoregressive models have demonstrated promising image generation capability in continuous space. However, their potential for video generation remains under-explored. In this paper, we propose \textbf{VideoMAR}, a concise and efficient decoder-only autoregressive image-to-video model with continuous tokens, composing temporal frame-by-frame and spatial masked generation. We first identify temporal causality and spatial bi-directionality as the first principle of video AR models, and propose the next-frame diffusion loss for the integration of mask and video generation. Besides, the huge cost and difficulty of long sequence autoregressive modeling is a basic but crucial issue. To this end, we propose the temporal short-to-long curriculum learning and spatial progressive resolution training, and employ progressive temperature strategy at inference time to mitigate the accumulation error. Furthermore, VideoMAR replicates several unique capacities of language models to video generation. It inherently bears high efficiency due to simultaneous temporal-wise KV cache and spatial-wise parallel generation, and presents the capacity of spatial and temporal extrapolation via 3D rotary embeddings. On the VBench-I2V benchmark, VideoMAR surpasses the previous state-of-the-art (Cosmos I2V) while requiring significantly fewer parameters ($9.3\%$), training data ($0.5\%$), and GPU resources ($0.2\%$). | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 04:08:18 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 09:44:09 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Yu",
"Hu",
""
],
[
"Gong",
"Biao",
""
],
[
"Yuan",
"Hangjie",
""
],
[
"Zheng",
"DanDan",
""
],
[
"Chai",
"Weilong",
""
],
[
"Chen",
"Jingdong",
""
],
[
"Zheng",
"Kecheng",
""
],
[
"Zhao",
"Feng",
""
]
] |
2506.14175 | Chenglong Wang | Chenglong Wang, Yang Gan, Yifu Huo, Yongyu Mu, Qiaozhi He, Murun Yang, Bei Li, Tong Xiao, Chunliang Zhang, Tongran Liu, Jingbo Zhu | GRAM: A Generative Foundation Reward Model for Reward Generalization | Accepted by ICML 2025 | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In aligning large language models (LLMs), reward models have played an important role, but are standardly trained as discriminative models and rely only on labeled human preference data. In this paper, we explore methods that train reward models using both unlabeled and labeled data. Building on the generative models in LLMs, we develop a generative reward model that is first trained via large-scale unsupervised learning and then fine-tuned via supervised learning. We also show that by using label smoothing, we are in fact optimizing a regularized pairwise ranking loss. This result, in turn, provides a new view of training reward models, which links generative models and discriminative models under the same class of training objectives. The outcome of these techniques is a foundation reward model, which can be applied to a wide range of tasks with little or no further fine-tuning effort. Extensive experiments show that this model generalizes well across several tasks, including response ranking, reinforcement learning from human feedback, and task adaptation with fine-tuning, achieving significant performance improvements over several strong baseline models. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 04:34:27 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 04:31:51 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Wang",
"Chenglong",
""
],
[
"Gan",
"Yang",
""
],
[
"Huo",
"Yifu",
""
],
[
"Mu",
"Yongyu",
""
],
[
"He",
"Qiaozhi",
""
],
[
"Yang",
"Murun",
""
],
[
"Li",
"Bei",
""
],
[
"Xiao",
"Tong",
""
],
[
"Zhang",
"Chunliang",
""
],
[
"Liu",
"Tongran",
""
],
[
"Zhu",
"Jingbo",
""
]
] |
2506.14230 | Kohei Kobayashi | Kohei Kobayashi | Norm Inequality for Perturbed Quantum Evolutions and Its Application to Grover's Algorithm | 5pages, 1figure | null | null | null | quant-ph | http://creativecommons.org/licenses/by-sa/4.0/ | We investigate the impact of coherent control errors on quantum state evolution by deriving a general norm inequality based on Gronwall's lemma. This inequality provides an explicit upper bound on the deviation between an ideal quantum state and one subject to arbitrary coherent perturbations, including both time-dependent and time-independent cases. The framework is broadly applicable, requiring no assumptions about the detailed structure of the perturbation and full dynamics of the quantum system. We apply this approach to analyze the robustness of Grover's search algorithm in the presence of coherent errors. Our results characterizes quantitative scaling relations between the error strength, algorithm runtime, and success probability, offering practical guidelines for the design of resilient quantum protocols. We further compare the effects of time-dependent and time-independent perturbations, showing the distinctive ways in which coherent errors accumulate in quantum dynamics. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 06:35:51 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 11:51:50 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Kobayashi",
"Kohei",
""
]
] |
2506.14253 | Kecai Deng | Kecai Deng, Hongyuan Qiu | A general solution for the 1-2 conjecture using real numbers | null | null | null | null | math.CO | http://creativecommons.org/licenses/by/4.0/ | The 1-2 conjecture posits that for every graph, weights from $\{1,2\}$ can be assigned to its vertices and edges so that adjacent vertices have distinct weighted degrees. This paper demonstrates that any set of two distinct real numbers $\{a,b\}$ can achieve the same goal, thereby confirming the 1-2 conjecture in a broader context. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 07:16:30 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 12:51:57 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Deng",
"Kecai",
""
],
[
"Qiu",
"Hongyuan",
""
]
] |
2506.14259 | David Damanik | Artur Avila (Universit\"at Z\"urich and IMPA), David Damanik (Rice University) | Dense Phenomena for Ergodic Schr\"odinger Operators: I. Spectrum, Integrated Density of States, and Lyapunov Exponent | 12 pages | null | null | null | math.SP math-ph math.DS math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider Schr\"odinger operators in $\ell^2(\Z)$ whose potentials are defined via continuous sampling along the orbits of a homeomorphism on a compact metric space. We show that for each non-atomic ergodic measure $\mu$, there is a dense set of sampling functions such that the associated almost sure spectrum has finitely many gaps, the integrated density of states is smooth, and the Lyapunov exponent is smooth and positive. As a byproduct we answer a question of Walters about the existence of non-uniform $\SL(2,\R)$ cocycles in the affirmative. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 07:20:42 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 08:04:02 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Avila",
"Artur",
"",
"Universität Zürich and IMPA"
],
[
"Damanik",
"David",
"",
"Rice University"
]
] |
2506.14261 | Rohan Gupta | Rohan Gupta, Erik Jenner | RL-Obfuscation: Can Language Models Learn to Evade Latent-Space Monitors? | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Latent-space monitors aim to detect undesirable behaviours in large language models by leveraging internal model representations rather than relying solely on black-box outputs. These methods have shown promise in identifying behaviours such as deception and unsafe completions, but a critical open question remains: can LLMs learn to evade such monitors? To study this, we introduce RL-Obfuscation, in which LLMs are finetuned via reinforcement learning to bypass latent-space monitors while maintaining coherent generations. We apply RL-Obfuscation to LLMs ranging from 7B to 14B parameters and evaluate evasion success against a suite of monitors. We find that token-level latent-space monitors are highly vulnerable to this attack. More holistic monitors, such as max-pooling or attention-based probes, remain robust. Moreover, we show that adversarial policies trained to evade a single static monitor generalise to unseen monitors of the same type. Finally, we study how the policy learned by RL bypasses these monitors and find that the model can also learn to repurpose tokens to mean something different internally. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 07:22:20 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 03:32:59 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Gupta",
"Rohan",
""
],
[
"Jenner",
"Erik",
""
]
] |
2506.14315 | Jinyan Yuan | Jinyan Yuan, Bangbang Yang, Keke Wang, Panwang Pan, Lin Ma, Xuehai Zhang, Xiao Liu, Zhaopeng Cui, Yuewen Ma | ImmerseGen: Agent-Guided Immersive World Generation with Alpha-Textured Proxies | Project webpage: https://immersegen.github.io | null | null | null | cs.GR cs.CV | http://creativecommons.org/licenses/by/4.0/ | Automatic creation of 3D scenes for immersive VR presence has been a significant research focus for decades. However, existing methods often rely on either high-poly mesh modeling with post-hoc simplification or massive 3D Gaussians, resulting in a complex pipeline or limited visual realism. In this paper, we demonstrate that such exhaustive modeling is unnecessary for achieving compelling immersive experience. We introduce ImmerseGen, a novel agent-guided framework for compact and photorealistic world modeling. ImmerseGen represents scenes as hierarchical compositions of lightweight geometric proxies, i.e., simplified terrain and billboard meshes, and generates photorealistic appearance by synthesizing RGBA textures onto these proxies. Specifically, we propose terrain-conditioned texturing for user-centric base world synthesis, and RGBA asset texturing for midground and foreground scenery. This reformulation offers several advantages: (i) it simplifies modeling by enabling agents to guide generative models in producing coherent textures that integrate seamlessly with the scene; (ii) it bypasses complex geometry creation and decimation by directly synthesizing photorealistic textures on proxies, preserving visual quality without degradation; (iii) it enables compact representations suitable for real-time rendering on mobile VR headsets. To automate scene creation from text prompts, we introduce VLM-based modeling agents enhanced with semantic grid-based analysis for improved spatial reasoning and accurate asset placement. ImmerseGen further enriches scenes with dynamic effects and ambient audio to support multisensory immersion. Experiments on scene generation and live VR showcases demonstrate that ImmerseGen achieves superior photorealism, spatial coherence and rendering efficiency compared to prior methods. Project webpage: https://immersegen.github.io. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 08:50:05 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 07:15:43 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Yuan",
"Jinyan",
""
],
[
"Yang",
"Bangbang",
""
],
[
"Wang",
"Keke",
""
],
[
"Pan",
"Panwang",
""
],
[
"Ma",
"Lin",
""
],
[
"Zhang",
"Xuehai",
""
],
[
"Liu",
"Xiao",
""
],
[
"Cui",
"Zhaopeng",
""
],
[
"Ma",
"Yuewen",
""
]
] |
2506.14397 | Yeonkyoung So | Yeonkyoung So, Gyuseong Lee, Sungmok Jung, Joonhak Lee, JiA Kang, Sangho Kim, Jaejin Lee | Thunder-NUBench: A Benchmark for LLMs' Sentence-Level Negation Understanding | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Negation is a fundamental linguistic phenomenon that poses persistent challenges for Large Language Models (LLMs), particularly in tasks requiring deep semantic understanding. Existing benchmarks often treat negation as a side case within broader tasks like natural language inference, resulting in a lack of benchmarks that exclusively target negation understanding. In this work, we introduce Thunder-NUBench, a novel benchmark explicitly designed to assess sentence-level negation understanding in LLMs. Thunder-NUBench goes beyond surface-level cue detection by contrasting standard negation with structurally diverse alternatives such as local negation, contradiction, and paraphrase. The benchmark consists of manually curated sentence-negation pairs and a multiple-choice dataset that enables in-depth evaluation of models' negation understanding. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:51:39 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 01:18:11 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"So",
"Yeonkyoung",
""
],
[
"Lee",
"Gyuseong",
""
],
[
"Jung",
"Sungmok",
""
],
[
"Lee",
"Joonhak",
""
],
[
"Kang",
"JiA",
""
],
[
"Kim",
"Sangho",
""
],
[
"Lee",
"Jaejin",
""
]
] |
2506.14424 | Richard Schussnig | Richard Schussnig, Niklas Fehn, Douglas Ramalho Queiroz Pacheco, Martin Kronbichler | Higher-Oder Splitting Schemes for Fluids with Variable Viscosity | null | null | null | null | cs.CE math-ph math.MP physics.comp-ph physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | This article investigates matrix-free higher-order discontinuous Galerkin (DG) discretizations of the Navier-Stokes equations for incompressible flows with variable viscosity. The viscosity field may be prescribed analytically or governed by a rheological law, as often found in biomedical or industrial applications. The DG discretization of the adapted second-order viscous terms is carried out via the symmetric interior penalty Galerkin method, obviating auxiliary variables. Based on this spatial discretization, we compare several linearized variants of saddle point block systems and projection-based splitting time integration schemes in terms of their computational performance. Compared to the velocity-pressure block-system for the former, the splitting scheme allows solving a sequence of simple problems such as mass, convection-diffusion and Poisson equations. We investigate under which conditions the improved temporal stability of fully implicit schemes and resulting expensive nonlinear solves outperform the splitting schemes and linearized variants that are stable under hyperbolic time step restrictions.
The key aspects of this work are i) a higher-order DG discretization for incompressible flows with variable viscosity, ii) accelerated nonlinear solver variants and suitable linearizations adopting a matrix-free $hp$-multigrid solver, and iii) a detailed comparison of the monolithic and projection-based solvers in terms of their (non-)linear solver performance.
The presented schemes are evaluated in a series of numerical examples verifying their spatial and temporal accuracy, and the preconditioner performance under increasing viscosity contrasts, while their efficiency is showcased in the backward-facing step benchmark. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:36:52 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 09:30:56 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Schussnig",
"Richard",
""
],
[
"Fehn",
"Niklas",
""
],
[
"Pacheco",
"Douglas Ramalho Queiroz",
""
],
[
"Kronbichler",
"Martin",
""
]
] |
2506.14540 | Gerardo Flores | Gerardo A. Flores and Alyssa H. Smith and Julia A. Fukuyama and Ashia C. Wilson | Aligning Evaluation with Clinical Priorities: Calibration, Label Shift, and Error Costs | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Machine learning-based decision support systems are increasingly deployed in clinical settings, where probabilistic scoring functions are used to inform and prioritize patient management decisions. However, widely used scoring rules, such as accuracy and AUC-ROC, fail to adequately reflect key clinical priorities, including calibration, robustness to distributional shifts, and sensitivity to asymmetric error costs. In this work, we propose a principled yet practical evaluation framework for selecting calibrated thresholded classifiers that explicitly accounts for the uncertainty in class prevalences and domain-specific cost asymmetries often found in clinical settings. Building on the theory of proper scoring rules, particularly the Schervish representation, we derive an adjusted variant of cross-entropy (log score) that averages cost-weighted performance over clinically relevant ranges of class balance. The resulting evaluation is simple to apply, sensitive to clinical deployment conditions, and designed to prioritize models that are both calibrated and robust to real-world variations. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 14:01:39 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 02:06:16 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Flores",
"Gerardo A.",
""
],
[
"Smith",
"Alyssa H.",
""
],
[
"Fukuyama",
"Julia A.",
""
],
[
"Wilson",
"Ashia C.",
""
]
] |
2506.14625 | Zheyu Zhang | Chenchen Yuan, Zheyu Zhang, Shuo Yang, Bardh Prenkaj, Gjergji Kasneci | Probabilistic Aggregation and Targeted Embedding Optimization for Collective Moral Reasoning in Large Language Models | Accepted to ACL 2025 (Findings) | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) have shown impressive moral reasoning abilities. Yet they often diverge when confronted with complex, multi-factor moral dilemmas. To address these discrepancies, we propose a framework that synthesizes multiple LLMs' moral judgments into a collectively formulated moral judgment, realigning models that deviate significantly from this consensus. Our aggregation mechanism fuses continuous moral acceptability scores (beyond binary labels) into a collective probability, weighting contributions by model reliability. For misaligned models, a targeted embedding-optimization procedure fine-tunes token embeddings for moral philosophical theories, minimizing JS divergence to the consensus while preserving semantic integrity. Experiments on a large-scale social moral dilemma dataset show our approach builds robust consensus and improves individual model fidelity. These findings highlight the value of data-driven moral alignment across multiple models and its potential for safer, more consistent AI systems. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 15:22:21 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 13:21:13 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Yuan",
"Chenchen",
""
],
[
"Zhang",
"Zheyu",
""
],
[
"Yang",
"Shuo",
""
],
[
"Prenkaj",
"Bardh",
""
],
[
"Kasneci",
"Gjergji",
""
]
] |
2506.14634 | Leah Von Der Heyde | Leah von der Heyde, Anna-Carolina Haensch, Bernd Wei{\ss}, Jessica Daikeler | AIn't Nothing But a Survey? Using Large Language Models for Coding German Open-Ended Survey Responses on Survey Motivation | to appear in Survey Research Methods | null | null | null | cs.CL cs.AI cs.CY | http://creativecommons.org/licenses/by/4.0/ | The recent development and wider accessibility of LLMs have spurred discussions about how they can be used in survey research, including classifying open-ended survey responses. Due to their linguistic capacities, it is possible that LLMs are an efficient alternative to time-consuming manual coding and the pre-training of supervised machine learning models. As most existing research on this topic has focused on English-language responses relating to non-complex topics or on single LLMs, it is unclear whether its findings generalize and how the quality of these classifications compares to established methods. In this study, we investigate to what extent different LLMs can be used to code open-ended survey responses in other contexts, using German data on reasons for survey participation as an example. We compare several state-of-the-art LLMs and several prompting approaches, and evaluate the LLMs' performance by using human expert codings. Overall performance differs greatly between LLMs, and only a fine-tuned LLM achieves satisfactory levels of predictive performance. Performance differences between prompting approaches are conditional on the LLM used. Finally, LLMs' unequal classification performance across different categories of reasons for survey participation results in different categorical distributions when not using fine-tuning. We discuss the implications of these findings, both for methodological research on coding open-ended responses and for their substantive analysis, and for practitioners processing or substantively analyzing such data. Finally, we highlight the many trade-offs researchers need to consider when choosing automated methods for open-ended response classification in the age of LLMs. In doing so, our study contributes to the growing body of research about the conditions under which LLMs can be efficiently, accurately, and reliably leveraged in survey research. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 15:28:53 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 09:56:49 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"von der Heyde",
"Leah",
""
],
[
"Haensch",
"Anna-Carolina",
""
],
[
"Weiß",
"Bernd",
""
],
[
"Daikeler",
"Jessica",
""
]
] |
2506.14635 | Changkai Chen | Changkai Chen, Jiliang Jing, Zhoujian Cao, Mengjie Wang | Complete quasinormal modes of Type-D black holes | 36 pages, 13 figures | null | null | null | gr-qc | http://creativecommons.org/licenses/by/4.0/ | Quasinormal mode (QNM) spectra of black holes exhibit two open problems [Conf. Proc. C 0405132, 145 (2004); CQG 26, 163001 (2009)]: (i) the discontinuity in highly damped QNMs between Schwarzschild and Kerr solutions as $a \to 0$, and (ii) the unexplained spectral proximity between QNMs and algebraically special (AS) frequencies, particularly the anomalous multiplet splitting for Kerr $\ell=2$, $m \geq 0$ modes. We develop a novel method to compute complete QNM spectra for Type-D black holes, solving both problems and establishing a mathematical framework for boundary value problems of dissipative systems. Using analytic continuation of radial eigenvalue equations, our method eliminates the dependence on auxiliary parameters in the connection formulas for confluent Heun solutions. This breakthrough overcomes the long-standing challenge of calculating QNMs that cross or lie on the negative imaginary axis (NIA). For Schwarzschild and Kerr spacetimes ($0 \leq \hat{n} \leq 41$, $2 \leq \ell \leq 16$), we present complete spectra validated through scattering amplitudes with errors $<10^{-10}$. The results provide definitive solutions to both open problems: (i) The inability of conventional methods to compute QNMs crossing or residing on the NIA leads to apparent discontinuities in Kerr spectra as $a \to 0$. (ii) When a QNM exactly coincides with the AS frequency, an additional QNM (unconventional mode) appears nearby. For the Kerr case with $\ell=2$, overtone sequences from both unconventional and AS modes exhibit precisely $2\ell+1$ branches, without multiplets or supersymmetry breaking at AS frequencies. Our method confirms Leung's conjecture of high-$\ell$ unconventional mode deviations from the NIA through the first $\ell=3$ calculation. Moreover, this paradigm surpasses the state-of-the-art Cook's continued fraction and isomonodromic methods in computational efficiency. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 15:29:36 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 09:53:26 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Chen",
"Changkai",
""
],
[
"Jing",
"Jiliang",
""
],
[
"Cao",
"Zhoujian",
""
],
[
"Wang",
"Mengjie",
""
]
] |
2506.14639 | Ariel Barton | Ariel Barton, Svitlana Mayboroda, and Alberto Pacati | The Poisson-Dirichlet problem in domains with Ahlfors regular boundary | 25 pages, 4 figures | null | null | null | math.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present an announcement of some recent results concerning well-posedness of the Poisson-Dirichlet problem with boundary data in Besov spaces with fractional smoothness. This is a far-reaching generalization as previously known theorems concerning well-posedness of the Poisson problem in such intermediate smoothness classes were mostly restricted to the context of Lipschitz domains and coefficients satisfying strong regularity assumptions. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 15:37:52 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 15:46:24 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Barton",
"Ariel",
""
],
[
"Mayboroda",
"Svitlana",
""
],
[
"Pacati",
"Alberto",
""
]
] |
2506.14665 | Rianne Van Den Berg | Giulia Luise, Chin-Wei Huang, Thijs Vogels, Derk P. Kooi, Sebastian Ehlert, Stephanie Lanius, Klaas J. H. Giesbertz, Amir Karton, Deniz Gunceler, Megan Stanley, Wessel P. Bruinsma, Lin Huang, Xinran Wei, Jos\'e Garrido Torres, Abylay Katbashev, B\'alint M\'at\'e, S\'ekou-Oumar Kaba, Roberto Sordillo, Yingrong Chen, David B. Williams-Young, Christopher M. Bishop, Jan Hermann, Rianne van den Berg, Paola Gori-Giorgi | Accurate and scalable exchange-correlation with deep learning | Main: 13 pages plus references, 11 figures and tables. Supplementary information: 19 pages, 12 figures and tables. v2 update: fix rendering of figure 1 and part of figure 5 in Safari PDF viewer | null | null | null | physics.chem-ph cs.AI cs.CE cs.LG physics.comp-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Density Functional Theory (DFT) is the most widely used electronic structure method for predicting the properties of molecules and materials. Although DFT is, in principle, an exact reformulation of the Schr\"odinger equation, practical applications rely on approximations to the unknown exchange-correlation (XC) functional. Most existing XC functionals are constructed using a limited set of increasingly complex, hand-crafted features that improve accuracy at the expense of computational efficiency. Yet, no current approximation achieves the accuracy and generality for predictive modeling of laboratory experiments at chemical accuracy -- typically defined as errors below 1 kcal/mol. In this work, we present Skala, a modern deep learning-based XC functional that bypasses expensive hand-designed features by learning representations directly from data. Skala achieves chemical accuracy for atomization energies of small molecules while retaining the computational efficiency typical of semi-local DFT. This performance is enabled by training on an unprecedented volume of high-accuracy reference data generated using computationally intensive wavefunction-based methods. Notably, Skala systematically improves with additional training data covering diverse chemistry. By incorporating a modest amount of additional high-accuracy data tailored to chemistry beyond atomization energies, Skala achieves accuracy competitive with the best-performing hybrid functionals across general main group chemistry, at the cost of semi-local DFT. As the training dataset continues to expand, Skala is poised to further enhance the predictive power of first-principles simulations. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 15:56:56 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 08:39:15 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Luise",
"Giulia",
""
],
[
"Huang",
"Chin-Wei",
""
],
[
"Vogels",
"Thijs",
""
],
[
"Kooi",
"Derk P.",
""
],
[
"Ehlert",
"Sebastian",
""
],
[
"Lanius",
"Stephanie",
""
],
[
"Giesbertz",
"Klaas J. H.",
""
],
[
"Karton",
"Amir",
""
],
[
"Gunceler",
"Deniz",
""
],
[
"Stanley",
"Megan",
""
],
[
"Bruinsma",
"Wessel P.",
""
],
[
"Huang",
"Lin",
""
],
[
"Wei",
"Xinran",
""
],
[
"Torres",
"José Garrido",
""
],
[
"Katbashev",
"Abylay",
""
],
[
"Máté",
"Bálint",
""
],
[
"Kaba",
"Sékou-Oumar",
""
],
[
"Sordillo",
"Roberto",
""
],
[
"Chen",
"Yingrong",
""
],
[
"Williams-Young",
"David B.",
""
],
[
"Bishop",
"Christopher M.",
""
],
[
"Hermann",
"Jan",
""
],
[
"Berg",
"Rianne van den",
""
],
[
"Gori-Giorgi",
"Paola",
""
]
] |
2506.14673 | Mikael M{\o}ller H{\o}gsgaard | Mikael M{\o}ller H{\o}gsgaard, Andrea Paudice | Uniform Mean Estimation for Heavy-Tailed Distributions via Median-of-Means | null | null | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by/4.0/ | The Median of Means (MoM) is a mean estimator that has gained popularity in the context of heavy-tailed data. In this work, we analyze its performance in the task of simultaneously estimating the mean of each function in a class $\mathcal{F}$ when the data distribution possesses only the first $p$ moments for $p \in (1,2]$. We prove a new sample complexity bound using a novel symmetrization technique that may be of independent interest. Additionally, we present applications of our result to $k$-means clustering with unbounded inputs and linear regression with general losses, improving upon existing works. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 16:07:36 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 06:49:11 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Høgsgaard",
"Mikael Møller",
""
],
[
"Paudice",
"Andrea",
""
]
] |
2506.14696 | Dahang Wan | Dahang Wan, Rongsheng Lu, Yang Fang, Xianli Lang, Shuangbao Shu, Jingjing Chen, Siyuan Shen, Ting Xu, Zecong Ye | YOLOv11-RGBT: Towards a Comprehensive Single-Stage Multispectral Object Detection Framework | 29 pages, 8 figures . The errors in the first version have been corrected, and no new version will be submitted in the near future. The next version will include more experiments | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multispectral object detection, which integrates information from multiple bands, can enhance detection accuracy and environmental adaptability, holding great application potential across various fields. Although existing methods have made progress in cross-modal interaction, low-light conditions, and model lightweight, there are still challenges like the lack of a unified single-stage framework, difficulty in balancing performance and fusion strategy, and unreasonable modality weight allocation. To address these, based on the YOLOv11 framework, we present YOLOv11-RGBT, a new comprehensive multimodal object detection framework. We designed six multispectral fusion modes and successfully applied them to models from YOLOv3 to YOLOv12 and RT-DETR. After reevaluating the importance of the two modalities, we proposed a P3 mid-fusion strategy and multispectral controllable fine-tuning (MCF) strategy for multispectral models. These improvements optimize feature fusion, reduce redundancy and mismatches, and boost overall model performance. Experiments show our framework excels on three major open-source multispectral object detection datasets, like LLVIP and FLIR. Particularly, the multispectral controllable fine-tuning strategy significantly enhanced model adaptability and robustness. On the FLIR dataset, it consistently improved YOLOv11 models' mAP by 3.41%-5.65%, reaching a maximum of 47.61%, verifying the framework and strategies' effectiveness. The code is available at: https://github.com/wandahangFY/YOLOv11-RGBT. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 16:37:00 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 13:54:55 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Wan",
"Dahang",
""
],
[
"Lu",
"Rongsheng",
""
],
[
"Fang",
"Yang",
""
],
[
"Lang",
"Xianli",
""
],
[
"Shu",
"Shuangbao",
""
],
[
"Chen",
"Jingjing",
""
],
[
"Shen",
"Siyuan",
""
],
[
"Xu",
"Ting",
""
],
[
"Ye",
"Zecong",
""
]
] |
2506.14731 | Jia Guo | Ling Team, Bin Hu, Cai Chen, Deng Zhao, Ding Liu, Dingnan Jin, Feng Zhu, Hao Dai, Hongzhi Luan, Jia Guo, Jiaming Liu, Jiewei Wu, Jun Mei, Jun Zhou, Junbo Zhao, Junwu Xiong, Kaihong Zhang, Kuan Xu, Lei Liang, Liang Jiang, Liangcheng Fu, Longfei Zheng, Qiang Gao, Qing Cui, Quan Wan, Shaomian Zheng, Shuaicheng Li, Tongkai Yang, Wang Ren, Xiaodong Yan, Xiaopei Wan, Xiaoyun Feng, Xin Zhao, Xinxing Yang, Xinyu Kong, Xuemin Yang, Yang Li, Yingting Wu, Yongkang Liu, Zhankai Xu, Zhenduo Zhang, Zhenglei Zhou, Zhenyu Huang, Zhiqiang Zhang, Zihao Wang, Zujie Wen | Ring-lite: Scalable Reasoning via C3PO-Stabilized Reinforcement Learning for LLMs | Technical Report | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present Ring-lite, a Mixture-of-Experts (MoE)-based large language model optimized via reinforcement learning (RL) to achieve efficient and robust reasoning capabilities. Built upon the publicly available Ling-lite model, a 16.8 billion parameter model with 2.75 billion activated parameters, our approach matches the performance of state-of-the-art (SOTA) small-scale reasoning models on challenging benchmarks (e.g., AIME, LiveCodeBench, GPQA-Diamond) while activating only one-third of the parameters required by comparable models. To accomplish this, we introduce a joint training pipeline integrating distillation with RL, revealing undocumented challenges in MoE RL training. First, we identify optimization instability during RL training, and we propose Constrained Contextual Computation Policy Optimization(C3PO), a novel approach that enhances training stability and improves computational throughput via algorithm-system co-design methodology. Second, we empirically demonstrate that selecting distillation checkpoints based on entropy loss for RL training, rather than validation metrics, yields superior performance-efficiency trade-offs in subsequent RL training. Finally, we develop a two-stage training paradigm to harmonize multi-domain data integration, addressing domain conflicts that arise in training with mixed dataset. We will release the model, dataset, and code. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 17:12:34 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Jun 2025 02:53:14 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Ling Team",
"",
""
],
[
"Hu",
"Bin",
""
],
[
"Chen",
"Cai",
""
],
[
"Zhao",
"Deng",
""
],
[
"Liu",
"Ding",
""
],
[
"Jin",
"Dingnan",
""
],
[
"Zhu",
"Feng",
""
],
[
"Dai",
"Hao",
""
],
[
"Luan",
"Hongzhi",
""
],
[
"Guo",
"Jia",
""
],
[
"Liu",
"Jiaming",
""
],
[
"Wu",
"Jiewei",
""
],
[
"Mei",
"Jun",
""
],
[
"Zhou",
"Jun",
""
],
[
"Zhao",
"Junbo",
""
],
[
"Xiong",
"Junwu",
""
],
[
"Zhang",
"Kaihong",
""
],
[
"Xu",
"Kuan",
""
],
[
"Liang",
"Lei",
""
],
[
"Jiang",
"Liang",
""
],
[
"Fu",
"Liangcheng",
""
],
[
"Zheng",
"Longfei",
""
],
[
"Gao",
"Qiang",
""
],
[
"Cui",
"Qing",
""
],
[
"Wan",
"Quan",
""
],
[
"Zheng",
"Shaomian",
""
],
[
"Li",
"Shuaicheng",
""
],
[
"Yang",
"Tongkai",
""
],
[
"Ren",
"Wang",
""
],
[
"Yan",
"Xiaodong",
""
],
[
"Wan",
"Xiaopei",
""
],
[
"Feng",
"Xiaoyun",
""
],
[
"Zhao",
"Xin",
""
],
[
"Yang",
"Xinxing",
""
],
[
"Kong",
"Xinyu",
""
],
[
"Yang",
"Xuemin",
""
],
[
"Li",
"Yang",
""
],
[
"Wu",
"Yingting",
""
],
[
"Liu",
"Yongkang",
""
],
[
"Xu",
"Zhankai",
""
],
[
"Zhang",
"Zhenduo",
""
],
[
"Zhou",
"Zhenglei",
""
],
[
"Huang",
"Zhenyu",
""
],
[
"Zhang",
"Zhiqiang",
""
],
[
"Wang",
"Zihao",
""
],
[
"Wen",
"Zujie",
""
]
] |
2506.14771 | Haiwei Dong | Mengyuan Wang, Yang Liu, Haopeng Wang, Haiwei Dong, Abdulmotaleb El Saddik | Empirical Studies of Large Scale Environment Scanning by Consumer Electronics | Accepted by IEEE Consumer Electronics Magazine | null | 10.1109/MCE.2025.3549585 | null | eess.IV cs.CV cs.ET cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents an empirical evaluation of the Matterport Pro3, a consumer-grade 3D scanning device, for large-scale environment reconstruction. We conduct detailed scanning (1,099 scanning points) of a six-floor building (17,567 square meters) and assess the device's effectiveness, limitations, and performance enhancements in diverse scenarios. Challenges encountered during the scanning are addressed through proposed solutions, while we also explore advanced methods to overcome them more effectively. Comparative analysis with another consumer-grade device (iPhone) highlights the Pro3's balance between cost-effectiveness and performance. The Matterport Pro3 achieves a denser point cloud with 1,877,324 points compared to the iPhone's 506,961 points and higher alignment accuracy with an RMSE of 0.0118 meters. The cloud-to-cloud (C2C) average distance error between the two point cloud models is 0.0408 meters, with a standard deviation of 0.0715 meters. The study demonstrates the Pro3's ability to generate high-quality 3D models suitable for large-scale applications, leveraging features such as LiDAR and advanced alignment techniques. | [
{
"version": "v1",
"created": "Thu, 27 Mar 2025 15:27:35 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Wang",
"Mengyuan",
""
],
[
"Liu",
"Yang",
""
],
[
"Wang",
"Haopeng",
""
],
[
"Dong",
"Haiwei",
""
],
[
"Saddik",
"Abdulmotaleb El",
""
]
] |
2506.14772 | Jakob De Moor | Jakob De Moor, Hans Weytjens, Johannes De Smedt, Jochen De Weerdt | SimBank: from Simulation to Solution in Prescriptive Process Monitoring | null | null | null | null | cs.DB cs.LG | http://creativecommons.org/licenses/by/4.0/ | Prescriptive Process Monitoring (PresPM) is an emerging area within Process Mining, focused on optimizing processes through real-time interventions for effective decision-making. PresPM holds significant promise for organizations seeking enhanced operational performance. However, the current literature faces two key limitations: a lack of extensive comparisons between techniques and insufficient evaluation approaches. To address these gaps, we introduce SimBank: a simulator designed for accurate benchmarking of PresPM methods. Modeled after a bank's loan application process, SimBank enables extensive comparisons of both online and offline PresPM methods. It incorporates a variety of intervention optimization problems with differing levels of complexity and supports experiments on key causal machine learning challenges, such as assessing a method's robustness to confounding in data. SimBank additionally offers a comprehensive evaluation capability: for each test case, it can generate the true outcome under each intervention action, which is not possible using recorded datasets. The simulator incorporates parallel activities and loops, drawing from common logs to generate cases that closely resemble real-life process instances. Our proof of concept demonstrates SimBank's benchmarking capabilities through experiments with various PresPM methods across different interventions, highlighting its value as a publicly available simulator for advancing research and practice in PresPM. | [
{
"version": "v1",
"created": "Fri, 28 Mar 2025 16:12:09 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"De Moor",
"Jakob",
""
],
[
"Weytjens",
"Hans",
""
],
[
"De Smedt",
"Johannes",
""
],
[
"De Weerdt",
"Jochen",
""
]
] |
2506.14774 | Burcu Sayin G\"unel | Burcu Sayin and Ipek Baris Schlicht and Ngoc Vo Hong and Sara Allievi and Jacopo Staiano and Pasquale Minervini and Andrea Passerini | MedSyn: Enhancing Diagnostics with Human-AI Collaboration | Accepted to the Trustworthy and Collaborative Artificial Intelligence Workshop 2025 (TCAI 2025) in the 4th International Conference Series on Hybrid Human-Artificial Intelligence (HHAI 2025) | null | null | null | cs.LG cs.AI cs.HC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Clinical decision-making is inherently complex, often influenced by cognitive biases, incomplete information, and case ambiguity. Large Language Models (LLMs) have shown promise as tools for supporting clinical decision-making, yet their typical one-shot or limited-interaction usage may overlook the complexities of real-world medical practice. In this work, we propose a hybrid human-AI framework, MedSyn, where physicians and LLMs engage in multi-step, interactive dialogues to refine diagnoses and treatment decisions. Unlike static decision-support tools, MedSyn enables dynamic exchanges, allowing physicians to challenge LLM suggestions while the LLM highlights alternative perspectives. Through simulated physician-LLM interactions, we assess the potential of open-source LLMs as physician assistants. Results show open-source LLMs are promising as physician assistants in the real world. Future work will involve real physician interactions to further validate MedSyn's usefulness in diagnostic accuracy and patient outcomes. | [
{
"version": "v1",
"created": "Wed, 07 May 2025 09:37:18 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Sayin",
"Burcu",
""
],
[
"Schlicht",
"Ipek Baris",
""
],
[
"Hong",
"Ngoc Vo",
""
],
[
"Allievi",
"Sara",
""
],
[
"Staiano",
"Jacopo",
""
],
[
"Minervini",
"Pasquale",
""
],
[
"Passerini",
"Andrea",
""
]
] |
2506.14775 | Tobias Labarta | Tobias Labarta, Nhi Hoang, Katharina Weitz, Wojciech Samek, Sebastian Lapuschkin, Leander Weber | See What I Mean? CUE: A Cognitive Model of Understanding Explanations | 10 pages, 5 figures (main text), 4 tables, 455-participant user study | null | null | null | cs.HC cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | As machine learning systems increasingly inform critical decisions, the need for human-understandable explanations grows. Current evaluations of Explainable AI (XAI) often prioritize technical fidelity over cognitive accessibility which critically affects users, in particular those with visual impairments. We propose CUE, a model for Cognitive Understanding of Explanations, linking explanation properties to cognitive sub-processes: legibility (perception), readability (comprehension), and interpretability (interpretation). In a study (N=455) testing heatmaps with varying colormaps (BWR, Cividis, Coolwarm), we found comparable task performance but lower confidence/effort for visually impaired users. Unlike expected, these gaps were not mitigated and sometimes worsened by accessibility-focused color maps like Cividis. These results challenge assumptions about perceptual optimization and support the need for adaptive XAI interfaces. They also validate CUE by demonstrating that altering explanation legibility affects understandability. We contribute: (1) a formalized cognitive model for explanation understanding, (2) an integrated definition of human-centered explanation properties, and (3) empirical evidence motivating accessible, user-tailored XAI. | [
{
"version": "v1",
"created": "Fri, 09 May 2025 22:05:20 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Labarta",
"Tobias",
""
],
[
"Hoang",
"Nhi",
""
],
[
"Weitz",
"Katharina",
""
],
[
"Samek",
"Wojciech",
""
],
[
"Lapuschkin",
"Sebastian",
""
],
[
"Weber",
"Leander",
""
]
] |
2506.14776 | Bin Shen | Yijie Miao and Bin Shen | Shi-type and Hamilton-type gradient estimates for a general parabolic equation under compact Finsler $CD(-K,N)$ geometric flows | null | null | null | null | math.DG math.AP | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Recently, the Li-Yau-type gradient estimates for positive solutions to parabolic equations
\begin{equation}
\partial_t u=\Delta u+\mathcal{R}_1u+\mathcal{R}_2u^{\alpha}+\mathcal{R}_3u(\log u)^{\beta},\notag
\end{equation}
under the general compact Finsler $CD(-K,N)$ geometric flow are studied. Here $\mathcal{R}_1$,$\mathcal{R}_2$,$\mathcal{R}_3$ $\in$ $ C^{1}(M,[0,T])$, $\alpha$ and $\beta$ are both positive constants, $T$ is the maximal existence time for the flow.
However, compared with the Riemannian case, the curvature conditions impose stricter derivative bounds on the development term in the geometric flow, as well as on the derivative bounds of the distortion of the manifold. In this manuscript, we present Shi-type and Hamilton-type gradient estimates to demonstrate the possibility of removing such conditions. | [
{
"version": "v1",
"created": "Tue, 13 May 2025 06:55:02 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Miao",
"Yijie",
""
],
[
"Shen",
"Bin",
""
]
] |
2506.14777 | Jules Leguy | Jules Leguy, Pierre-Antoine Jean, Felipe Torres Figueroa, S\'ebastien Harispe | WebXAII: an open-source web framework to study human-XAI interaction | null | null | null | null | cs.HC cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | This article introduces WebXAII, an open-source web framework designed to facilitate research on human interaction with eXplainable Artificial Intelligence (XAI) systems. The field of XAI is rapidly expanding, driven by the growing societal implications of the widespread adoption of AI (and in particular machine learning) across diverse applications. Researchers who study the interaction between humans and XAI techniques typically develop ad hoc interfaces in order to conduct their studies. These interfaces are usually not shared alongside the results of the studies, which limits their reusability and the reproducibility of experiments. In response, we design and implement WebXAII, a web-based platform that can embody full experimental protocols, meaning that it can present all aspects of the experiment to human participants and record their responses. The experimental protocols are translated into a composite architecture of generic views and modules, which offers a lot of flexibility. The architecture is defined in a structured configuration file, so that protocols can be implemented with minimal programming skills. We demonstrate that WebXAII can effectively embody relevant protocols, by reproducing the protocol of a state-of-the-art study of the literature. The framework is available at https://github.com/PAJEAN/WebXAII. | [
{
"version": "v1",
"created": "Fri, 16 May 2025 10:12:53 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Leguy",
"Jules",
""
],
[
"Jean",
"Pierre-Antoine",
""
],
[
"Figueroa",
"Felipe Torres",
""
],
[
"Harispe",
"Sébastien",
""
]
] |
2506.14778 | Mihail Cocos | Mihail Cocos | A Riemannian Characterization of Compact Affine Manifolds with Parallel Volume | null | null | null | null | math.DG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We establish that any affine manifold $(M,\nabla)$ endowed with a parallel volume form $\omega,$ admits, in any conformal class of Riemannian metrics, a representative $H$ for which $\nabla$ is the Levi-Civita connection. This provides a constructive proof that such manifolds are necessarily complete, generalizing the "if" direction of Markus' conjecture \cite{markus1962}. Moreover, our result demonstrates that these structures are intrinsically Riemannian-flat, a stronger conclusion than the affine completeness asserted by Markus. The metric $H$ arises naturally from the Hessian of volume-normalized distance functions and is shown to be globally smooth and $\nabla$-parallel, extending results of \cite{goldman1982} and \cite{benzecri1955} to higher dimensions with additional geometric structure. The construction proceeds through three technically novel steps: (1) local parallel metric normalization using the given volume form, (2) explicit Hessian calculations in adapted coordinates, and (3) gluing via affine transition maps that preserve the volumetric geometry. This approach reveals an unexpected rigidity in flat affine manifolds with compatible volume that goes beyond the topological constraints studied in \cite{fried1980}. | [
{
"version": "v1",
"created": "Thu, 22 May 2025 23:11:14 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Cocos",
"Mihail",
""
]
] |
2506.14779 | Veronica Pasquarella | Veronica Pasquarella | Homological Mirror Symmetry Course at SIMIS: Introduction and Applications | 127 pages, plus references. Course taught in the Spring Semester 2025 | null | null | null | math.AG math-ph math.CT math.MP math.SG | http://creativecommons.org/licenses/by/4.0/ | The present work consists of topics covered through a course currently taught by the author at SIMIS. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 00:58:04 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Pasquarella",
"Veronica",
""
]
] |
2506.14780 | Reda Chhaibi | Reda Chhaibi, Serge Gratton, Samuel Vaiter | Faster Computation of Entropic Optimal Transport via Stable Low Frequency Modes | 22 pages, 5 figures | null | null | null | math.NA cs.NA math.OC math.PR stat.CO stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we propose an accelerated version for the Sinkhorn algorithm, which is the reference method for computing the solution to Entropic Optimal Transport.
Its main draw-back is the exponential slow-down of convergence as the regularization weakens $\varepsilon \rightarrow 0$.
Thanks to spectral insights on the behavior of the Hessian, we propose to mitigate the problem via an original spectral warm-start strategy. This leads to faster convergence compared to the reference method, as also demonstrated in our numerical experiments. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 11:39:19 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Chhaibi",
"Reda",
""
],
[
"Gratton",
"Serge",
""
],
[
"Vaiter",
"Samuel",
""
]
] |
2506.14781 | Corentin Delacour | Corentin Delacour, M Mahmudul Hasan Sajeeb, Joao P. Hespanha, Kerem Y. Camsari | Two-dimensional Parallel Tempering for Constrained Optimization | null | null | null | null | cs.LG cond-mat.stat-mech math.OC stat.ML | http://creativecommons.org/licenses/by/4.0/ | Sampling Boltzmann probability distributions plays a key role in machine learning and optimization, motivating the design of hardware accelerators such as Ising machines. While the Ising model can in principle encode arbitrary optimization problems, practical implementations are often hindered by soft constraints that either slow down mixing when too strong, or fail to enforce feasibility when too weak. We introduce a two-dimensional extension of the powerful parallel tempering algorithm (PT) that addresses this challenge by adding a second dimension of replicas interpolating the penalty strengths. This scheme ensures constraint satisfaction in the final replicas, analogous to low-energy states at low temperature. The resulting two-dimensional parallel tempering algorithm (2D-PT) improves mixing in heavily constrained replicas and eliminates the need to explicitly tune the penalty strength. In a representative example of graph sparsification with copy constraints, 2D-PT achieves near-ideal mixing, with Kullback-Leibler divergence decaying as O(1/t). When applied to sparsified Wishart instances, 2D-PT yields orders of magnitude speedup over conventional PT with the same number of replicas. The method applies broadly to constrained Ising problems and can be deployed on existing Ising machines. | [
{
"version": "v1",
"created": "Sat, 24 May 2025 20:41:45 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Delacour",
"Corentin",
""
],
[
"Sajeeb",
"M Mahmudul Hasan",
""
],
[
"Hespanha",
"Joao P.",
""
],
[
"Camsari",
"Kerem Y.",
""
]
] |
2506.14782 | Geraci Joseph | Joseph Geraci, Bessi Qorri, Christian Cumbaa, Mike Tsay, Paul Leonczyk, Luca Pani | Integrating Dynamical Systems Learning with Foundational Models: A Meta-Evolutionary AI Framework for Clinical Trials | 27 pages | null | null | null | cs.LG q-bio.QM | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Artificial intelligence (AI) has evolved into an ecosystem of specialized "species," each with unique strengths. We analyze two: DeepSeek-V3, a 671-billion-parameter Mixture of Experts large language model (LLM) exemplifying scale-driven generality, and NetraAI, a dynamical system-based framework engineered for stability and interpretability on small clinical trial datasets. We formalize NetraAI's foundations, combining contraction mappings, information geometry, and evolutionary algorithms to identify predictive patient cohorts. Features are embedded in a metric space and iteratively contracted toward stable attractors that define latent subgroups. A pseudo-temporal embedding and long-range memory enable exploration of higher-order feature interactions, while an internal evolutionary loop selects compact, explainable 2-4-variable bundles ("Personas").
To guide discovery, we introduce an LLM Strategist as a meta-evolutionary layer that observes Persona outputs, prioritizes promising variables, injects domain knowledge, and assesses robustness. This two-tier architecture mirrors the human scientific process: NetraAI as experimentalist, the LLM as theorist, forming a self-improving loop.
In case studies (schizophrenia, depression, pancreatic cancer), NetraAI uncovered small, high-effect-size subpopulations that transformed weak baseline models (AUC ~0.50-0.68) into near-perfect classifiers using only a few features. We position NetraAI at the intersection of dynamical systems, information geometry, and evolutionary learning, aligned with emerging concept-level reasoning paradigms such as LeCun's Joint Embedding Predictive Architecture (JEPA). By prioritizing reliable, explainable knowledge, NetraAI offers a new generation of adaptive, self-reflective AI to accelerate clinical discovery. | [
{
"version": "v1",
"created": "Sun, 25 May 2025 03:34:33 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Geraci",
"Joseph",
""
],
[
"Qorri",
"Bessi",
""
],
[
"Cumbaa",
"Christian",
""
],
[
"Tsay",
"Mike",
""
],
[
"Leonczyk",
"Paul",
""
],
[
"Pani",
"Luca",
""
]
] |
2506.14783 | Mohamed Masry | Mohamed Masry, Mohamed Amen, Mohamed Elzyat, Mohamed Hamed, Norhan Magdy, Maram Khaled | ETS: Open Vocabulary Electroencephalography-To-Text Decoding and Sentiment Classification | Graduation project report submitted at Faculty of Computer Science and Artificial Intelligence, Helwan University | null | null | null | cs.LG cs.CL cs.HC | http://creativecommons.org/licenses/by/4.0/ | Decoding natural language from brain activity using non-invasive electroencephalography (EEG) remains a significant challenge in neuroscience and machine learning, particularly for open-vocabulary scenarios where traditional methods struggle with noise and variability. Previous studies have achieved high accuracy on small-closed vocabularies, but it still struggles on open vocabularies. In this study, we propose ETS, a framework that integrates EEG with synchronized eye-tracking data to address two critical tasks: (1) open-vocabulary text generation and (2) sentiment classification of perceived language. Our model achieves a superior performance on BLEU and Rouge score for EEG-To-Text decoding and up to 10% F1 score on EEG-based ternary sentiment classification, which significantly outperforms supervised baselines. Furthermore, we show that our proposed model can handle data from various subjects and sources, showing great potential for high performance open vocabulary eeg-to-text system. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 10:58:13 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Masry",
"Mohamed",
""
],
[
"Amen",
"Mohamed",
""
],
[
"Elzyat",
"Mohamed",
""
],
[
"Hamed",
"Mohamed",
""
],
[
"Magdy",
"Norhan",
""
],
[
"Khaled",
"Maram",
""
]
] |
2506.14784 | Emre Yilmaz | Emre Yilmaz and Philipp Bekemeyer | Predicting Onflow Parameters Using Transfer Learning for Domain and Task Adaptation | null | null | null | null | cs.LG physics.flu-dyn | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Determining onflow parameters is crucial from the perspectives of wind tunnel testing and regular flight and wind turbine operations. These parameters have traditionally been predicted via direct measurements which might lead to challenges in case of sensor faults. Alternatively, a data-driven prediction model based on surface pressure data can be used to determine these parameters. It is essential that such predictors achieve close to real-time learning as dictated by practical applications such as monitoring wind tunnel operations or learning the variations in aerodynamic performance of aerospace and wind energy systems. To overcome the challenges caused by changes in the data distribution as well as in adapting to a new prediction task, we propose a transfer learning methodology to predict the onflow parameters, specifically angle of attack and onflow speed. It requires first training a convolutional neural network (ConvNet) model offline for the core prediction task, then freezing the weights of this model except the selected layers preceding the output node, and finally executing transfer learning by retraining these layers. A demonstration of this approach is provided using steady CFD analysis data for an airfoil for i) domain adaptation where transfer learning is performed with data from a target domain having different data distribution than the source domain and ii) task adaptation where the prediction task is changed. Further exploration on the influence of noisy data, performance on an extended domain, and trade studies varying sampling sizes and architectures are provided. Results successfully demonstrate the potential of the approach for adaptation to changing data distribution, domain extension, and task update while the application for noisy data is concluded to be not as effective. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 13:39:42 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Yilmaz",
"Emre",
""
],
[
"Bekemeyer",
"Philipp",
""
]
] |
2506.14785 | Shiping Zhou | Shiping Zhou, Juntao Huang and Andrew J. Christlieb | Moment-enhanced shallow water equations for non-slip boundary conditions | null | null | null | null | math.NA cs.NA math-ph math.MP physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | The shallow water equations often assume a constant velocity profile along the vertical axis. However, this assumption does not hold in many practical applications. To better approximate the vertical velocity distribution, models such as the shallow water moment expansion models have been proposed. Nevertheless, under non-slip bottom boundary conditions, both the standard shallow water equation and its moment-enhanced models struggle to accurately capture the vertical velocity profile due to the stiff source terms. In this work, we propose modified shallow water equations and corresponding moment-enhanced models that perform well under both non-slip and slip boundary conditions. The primary difference between the modified and original models lies in the treatment of the source term, which allows our modified moment expansion models to be readily generalized, while maintaining compatibility with our previous analysis on the hyperbolicity of the model. To assess the performance of both the standard and modified moment expansion models, we conduct a comprehensive numerical comparison with the incompressible Navier--Stokes equations -- a comparison that is absent from existing literature. | [
{
"version": "v1",
"created": "Mon, 26 May 2025 19:19:16 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Zhou",
"Shiping",
""
],
[
"Huang",
"Juntao",
""
],
[
"Christlieb",
"Andrew J.",
""
]
] |
2506.14786 | Haobo Li | Haobo Li, Eunseo Jung, Zixin Chen, Zhaowei Wang, Yueya Wang, Huamin Qu, Alexis Kai Hon Lau | PIPE: Physics-Informed Position Encoding for Alignment of Satellite Images and Time Series | null | null | null | null | cs.LG cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | Multimodal time series forecasting is foundational in various fields, such as utilizing satellite imagery and numerical data for predicting typhoons in climate science. However, existing multimodal approaches primarily focus on utilizing text data to help time series forecasting, leaving the visual data in existing time series datasets untouched. Furthermore, it is challenging for models to effectively capture the physical information embedded in visual data, such as satellite imagery's temporal and geospatial context, which extends beyond images themselves. To address this gap, we propose physics-informed positional encoding (PIPE), a lightweight method that embeds physical information into vision language models (VLMs). PIPE introduces two key innovations: (1) a physics-informed positional indexing scheme for mapping physics to positional IDs, and (2) a variant-frequency positional encoding mechanism for encoding frequency information of physical variables and sequential order of tokens within the embedding space. By preserving both the physical information and sequential order information, PIPE significantly improves multimodal alignment and forecasting accuracy. Through the experiments on the most representative and the largest open-sourced satellite image dataset, PIPE achieves state-of-the-art performance in both deep learning forecasting and climate domain methods, demonstrating superiority across benchmarks, including a 12% improvement in typhoon intensity forecasting over prior works. Our code is provided in the supplementary material. | [
{
"version": "v1",
"created": "Tue, 27 May 2025 04:03:41 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Li",
"Haobo",
""
],
[
"Jung",
"Eunseo",
""
],
[
"Chen",
"Zixin",
""
],
[
"Wang",
"Zhaowei",
""
],
[
"Wang",
"Yueya",
""
],
[
"Qu",
"Huamin",
""
],
[
"Lau",
"Alexis Kai Hon",
""
]
] |
2506.14787 | Yuan Tian | Funing Li, Yuan Tian, Ruben Noortwyck, Jifeng Zhou, Liming Kuang, Robert Schulz | Topology-Aware and Highly Generalizable Deep Reinforcement Learning for Efficient Retrieval in Multi-Deep Storage Systems | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In modern industrial and logistics environments, the rapid expansion of fast delivery services has heightened the demand for storage systems that combine high efficiency with increased density. Multi-deep autonomous vehicle storage and retrieval systems (AVS/RS) present a viable solution for achieving greater storage density. However, these systems encounter significant challenges during retrieval operations due to lane blockages. A conventional approach to mitigate this issue involves storing items with homogeneous characteristics in a single lane, but this strategy restricts the flexibility and adaptability of multi-deep storage systems.
In this study, we propose a deep reinforcement learning-based framework to address the retrieval problem in multi-deep storage systems with heterogeneous item configurations. Each item is associated with a specific due date, and the objective is to minimize total tardiness. To effectively capture the system's topology, we introduce a graph-based state representation that integrates both item attributes and the local topological structure of the multi-deep warehouse. To process this representation, we design a novel neural network architecture that combines a Graph Neural Network (GNN) with a Transformer model. The GNN encodes topological and item-specific information into embeddings for all directly accessible items, while the Transformer maps these embeddings into global priority assignments. The Transformer's strong generalization capability further allows our approach to be applied to storage systems with diverse layouts. Extensive numerical experiments, including comparisons with heuristic methods, demonstrate the superiority of the proposed neural network architecture and the effectiveness of the trained agent in optimizing retrieval tardiness. | [
{
"version": "v1",
"created": "Tue, 27 May 2025 08:07:38 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Li",
"Funing",
""
],
[
"Tian",
"Yuan",
""
],
[
"Noortwyck",
"Ruben",
""
],
[
"Zhou",
"Jifeng",
""
],
[
"Kuang",
"Liming",
""
],
[
"Schulz",
"Robert",
""
]
] |
2506.14788 | Masato Kimura Dr. | Md Mamun Miah, Ryuhei Wakida, Masato Kimura | Energy-consistent dynamic fracture phase field models: unilateral constraints and finite element simulations | null | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by/4.0/ | Phase field models have emerged as a powerful and flexible framework for simulating complex interface-driven phenomena across a wide range of scientific and engineering applications. In fracture mechanics, the phase field approach--formulated as a gradient flow of the Griffith fracture energy with Ambrosio-Tortorelli regularization--has gained significant attention for its ability to capture complex crack topologies. In this study, we propose a dynamic fracture phase field model (DF-PFM) based on the elastodynamic wave equation. We further extend this framework by incorporating a unilateral contact condition, yielding a refined model suitable for simulating fault rupture under high pressure. For both models, we rigorously derive energy dissipation identities under mixed boundary conditions, ensuring energy consistency of the formulations. To validate the proposed approach, we conduct numerical experiments using linear implicit time discretization and finite element methods. Our simulations demonstrate that the unilateral contact condition is essential for accurately capturing shear-dominated crack propagation and preventing non-physical interpenetration, especially under high-compression loading scenarios relevant to seismic faulting. | [
{
"version": "v1",
"created": "Tue, 27 May 2025 15:23:45 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Miah",
"Md Mamun",
""
],
[
"Wakida",
"Ryuhei",
""
],
[
"Kimura",
"Masato",
""
]
] |
2506.14789 | Saman Khamesian | Saman Khamesian, Asiful Arefeen, Bithika M. Thompson, Maria Adela Grando, and Hassan Ghasemzadeh | AZT1D: A Real-World Dataset for Type 1 Diabetes | 4 pages | null | 10.17632/gk9m674wcx.1 | null | cs.LG q-bio.QM | http://creativecommons.org/licenses/by/4.0/ | High quality real world datasets are essential for advancing data driven approaches in type 1 diabetes (T1D) management, including personalized therapy design, digital twin systems, and glucose prediction models. However, progress in this area has been limited by the scarcity of publicly available datasets that offer detailed and comprehensive patient data. To address this gap, we present AZT1D, a dataset containing data collected from 25 individuals with T1D on automated insulin delivery (AID) systems. AZT1D includes continuous glucose monitoring (CGM) data, insulin pump and insulin administration data, carbohydrate intake, and device mode (regular, sleep, and exercise) obtained over 6 to 8 weeks for each patient. Notably, the dataset provides granular details on bolus insulin delivery (i.e., total dose, bolus type, correction specific amounts) features that are rarely found in existing datasets. By offering rich, naturalistic data, AZT1D supports a wide range of artificial intelligence and machine learning applications aimed at improving clinical decision making and individualized care in T1D. | [
{
"version": "v1",
"created": "Tue, 27 May 2025 21:54:53 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Khamesian",
"Saman",
""
],
[
"Arefeen",
"Asiful",
""
],
[
"Thompson",
"Bithika M.",
""
],
[
"Grando",
"Maria Adela",
""
],
[
"Ghasemzadeh",
"Hassan",
""
]
] |
2506.14790 | Tianxiang Zhan | Tianxiang Zhan, Ming Jin, Yuanpeng He, Yuxuan Liang, Yong Deng, Shirui Pan | Continuous Evolution Pool: Taming Recurring Concept Drift in Online Time Series Forecasting | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recurring concept drift, a type of concept drift in which previously observed data patterns reappear after some time, is one of the most prevalent types of concept drift in time series. As time progresses, concept drift occurs and previously encountered concepts are forgotten, thereby leading to a decline in the accuracy of online predictions. Existing solutions employ parameter updating techniques to delay forgetting; however, this may result in the loss of some previously learned knowledge while neglecting the exploration of knowledge retention mechanisms. To retain all conceptual knowledge and fully utilize it when the concepts recur, we propose the Continuous Evolution Pool (CEP), a pooling mechanism that stores different instances of forecasters for different concepts. Our method first selects the forecaster nearest to the test sample and then learns the features from its neighboring samples - a process we refer to as the retrieval. If there are insufficient neighboring samples, it indicates that a new concept has emerged, and a new model will evolve from the current nearest sample to the pool to store the knowledge of the concept. Simultaneously, the elimination mechanism will enable outdated knowledge to be cleared to ensure the prediction effect of the forecasters. Experiments on different architectural models and eight real datasets demonstrate that CEP effectively retains the knowledge of different concepts. In the scenario of online forecasting with recurring concepts, CEP significantly enhances the prediction results. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 03:27:49 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Zhan",
"Tianxiang",
""
],
[
"Jin",
"Ming",
""
],
[
"He",
"Yuanpeng",
""
],
[
"Liang",
"Yuxuan",
""
],
[
"Deng",
"Yong",
""
],
[
"Pan",
"Shirui",
""
]
] |
2506.14791 | Yunchong Liu | Jingxuan Zhou, Yuehao Wu, Yibo Zhang, Yeyubei Zhang, Yunchong Liu, Bolin Huang, Chunhong Yuan | SemIRNet: A Semantic Irony Recognition Network for Multimodal Sarcasm Detection | 5 pages, 3 figures | null | null | null | cs.CV cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Aiming at the problem of difficulty in accurately identifying graphical implicit correlations in multimodal irony detection tasks, this paper proposes a Semantic Irony Recognition Network (SemIRNet). The model contains three main innovations: (1) The ConceptNet knowledge base is introduced for the first time to acquire conceptual knowledge, which enhances the model's common-sense reasoning ability; (2) Two cross-modal semantic similarity detection modules at the word level and sample level are designed to model graphic-textual correlations at different granularities; and (3) A contrastive learning loss function is introduced to optimize the spatial distribution of the sample features, which improves the separability of positive and negative samples. Experiments on a publicly available multimodal irony detection benchmark dataset show that the accuracy and F1 value of this model are improved by 1.64% and 2.88% to 88.87% and 86.33%, respectively, compared with the existing optimal methods. Further ablation experiments verify the important role of knowledge fusion and semantic similarity detection in improving the model performance. | [
{
"version": "v1",
"created": "Wed, 28 May 2025 16:09:34 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Zhou",
"Jingxuan",
""
],
[
"Wu",
"Yuehao",
""
],
[
"Zhang",
"Yibo",
""
],
[
"Zhang",
"Yeyubei",
""
],
[
"Liu",
"Yunchong",
""
],
[
"Huang",
"Bolin",
""
],
[
"Yuan",
"Chunhong",
""
]
] |
2506.14792 | Keaton Burns | Calum S. Skene, Keaton J. Burns | Fast automated adjoints for spectral PDE solvers | 17 pages, 6 figures | null | null | null | math.NA astro-ph.SR cs.NA physics.ao-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a general and automated approach for computing model gradients for PDE solvers built on sparse spectral methods, and implement this capability in the widely used open-source Dedalus framework. We apply reverse-mode automatic differentiation to symbolic graph representations of PDEs, efficiently constructing adjoint solvers that retain the speed and memory efficiency of this important class of modern numerical methods. This approach enables users to compute gradients and perform optimization for a wide range of time-dependent and nonlinear systems without writing additional code. The framework supports a broad class of equations, geometries, and boundary conditions, and runs efficiently in parallel using MPI. We demonstrate the flexibility and capabilities of this system using canonical problems from the literature, showing both strong performance and practical utility for a wide variety of inverse problems. By integrating automatic adjoints into a flexible high-level solver, our approach enables researchers to perform gradient-based optimization and sensitivity analyses in spectral simulations with ease and efficiency. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 16:04:44 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Skene",
"Calum S.",
""
],
[
"Burns",
"Keaton J.",
""
]
] |
2506.14793 | Aditya Ravuri | Aditya Ravuri, Neil D. Lawrence | Protein Language Model Zero-Shot Fitness Predictions are Improved by Inference-only Dropout | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Protein Language Models (PLMs) such as ESM2 have been shown to be capable of zero-shot prediction of critical scalar properties of proteins (fitness). In this work, we show that injecting a dropout layer at inference time between a PLM's featurizer/embedding layer and its transformer, and averaging its output akin to Monte-Carlo dropout increases zero-shot performance on a subset of the ProteinGym dataset. This is the case even when the model was not trained with dropouts to begin with, and does not require retraining or finetuning of the PLM. A dropout of 0.1 seems performant across all models. | [
{
"version": "v1",
"created": "Sat, 31 May 2025 17:16:59 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Ravuri",
"Aditya",
""
],
[
"Lawrence",
"Neil D.",
""
]
] |
2506.14794 | Daniel Klingmann | Henrik Klagges, Robert Dahlke, Fabian Klemm, Benjamin Merkel, Daniel Klingmann, David A. Reiss, Dan Zecha | Assembly of Experts: Linear-time construction of the Chimera LLM variants with emergent and adaptable behaviors | null | null | null | null | cs.LG cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | Requiring $10^{13}$-$10^{15}$ FLOPs to calculate one 8 bit weight in an LLM during pretraining is extremely expensive and seems inefficient. To better leverage the huge investments made into pretrained models, we develop the new "Assembly-of-Experts" (AoE) construction method to create capable child variants of existing Mixture-of-Experts parent models in linear time. Model weight tensors get interpolated individually, allowing to enhance or suppress semantic features of the parents.
Varying the proportion of weights taken from the parent models, we observe some properties of the AoE child model changing gradually, while other behavioral traits emerge with a sharp transition. Surprisingly, nearly every generated model is functional and capable, which makes searching the model space straightforward.
We construct the DeepSeek R1T "Chimera", a 671B open-weights hybrid model combining DeepSeek's V3-0324 and R1 model variants. The child inherits only the routed expert tensors of R1, but still achieves about R1-level intelligence. At the same time, it uses about 40\% fewer output tokens, close to V3 speed. Constructed without any fine-tuning or distillation, the Chimera exhibits surprisingly compact, orderly reasoning compared to its parent models. | [
{
"version": "v1",
"created": "Sat, 31 May 2025 18:23:19 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Klagges",
"Henrik",
""
],
[
"Dahlke",
"Robert",
""
],
[
"Klemm",
"Fabian",
""
],
[
"Merkel",
"Benjamin",
""
],
[
"Klingmann",
"Daniel",
""
],
[
"Reiss",
"David A.",
""
],
[
"Zecha",
"Dan",
""
]
] |
2506.14795 | Batuhan Hangun | Batuhan Hangun, Emine Akpinar, Oguz Altun, Onder Eyecioglu | Comparative Analysis of QNN Architectures for Wind Power Prediction: Feature Maps and Ansatz Configurations | 6 pages, 2 figures | null | null | null | quant-ph cs.AI cs.LG eess.SP | http://creativecommons.org/licenses/by/4.0/ | Quantum Machine Learning (QML) is an emerging field at the intersection of quantum computing and machine learning, aiming to enhance classical machine learning methods by leveraging quantum mechanics principles such as entanglement and superposition. However, skepticism persists regarding the practical advantages of QML, mainly due to the current limitations of noisy intermediate-scale quantum (NISQ) devices. This study addresses these concerns by extensively assessing Quantum Neural Networks (QNNs)-quantum-inspired counterparts of Artificial Neural Networks (ANNs), demonstrating their effectiveness compared to classical methods. We systematically construct and evaluate twelve distinct QNN configurations, utilizing two unique quantum feature maps combined with six different entanglement strategies for ansatz design. Experiments conducted on a wind energy dataset reveal that QNNs employing the Z feature map achieve up to 93% prediction accuracy when forecasting wind power output using only four input parameters. Our findings show that QNNs outperform classical methods in predictive tasks, underscoring the potential of QML in real-world applications. | [
{
"version": "v1",
"created": "Sat, 31 May 2025 19:17:53 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Hangun",
"Batuhan",
""
],
[
"Akpinar",
"Emine",
""
],
[
"Altun",
"Oguz",
""
],
[
"Eyecioglu",
"Onder",
""
]
] |
2506.14796 | Zhangyang Gao | Zhangyang Gao, Hao Wang, Cheng Tan, Chenrui Xu, Mengdi Liu, Bozhen Hu, Linlin Chao, Xiaoming Zhang, Stan Z. Li | PFMBench: Protein Foundation Model Benchmark | null | null | null | null | q-bio.BM cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This study investigates the current landscape and future directions of protein foundation model research. While recent advancements have transformed protein science and engineering, the field lacks a comprehensive benchmark for fair evaluation and in-depth understanding. Since ESM-1B, numerous protein foundation models have emerged, each with unique datasets and methodologies. However, evaluations often focus on limited tasks tailored to specific models, hindering insights into broader generalization and limitations. Specifically, researchers struggle to understand the relationships between tasks, assess how well current models perform across them, and determine the criteria in developing new foundation models. To fill this gap, we present PFMBench, a comprehensive benchmark evaluating protein foundation models across 38 tasks spanning 8 key areas of protein science. Through hundreds of experiments on 17 state-of-the-art models across 38 tasks, PFMBench reveals the inherent correlations between tasks, identifies top-performing models, and provides a streamlined evaluation protocol. Code is available at \href{https://github.com/biomap-research/PFMBench}{\textcolor{blue}{GitHub}}. | [
{
"version": "v1",
"created": "Sun, 01 Jun 2025 07:40:07 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Gao",
"Zhangyang",
""
],
[
"Wang",
"Hao",
""
],
[
"Tan",
"Cheng",
""
],
[
"Xu",
"Chenrui",
""
],
[
"Liu",
"Mengdi",
""
],
[
"Hu",
"Bozhen",
""
],
[
"Chao",
"Linlin",
""
],
[
"Zhang",
"Xiaoming",
""
],
[
"Li",
"Stan Z.",
""
]
] |
2506.14797 | Marco Nurisso | Marco Nurisso, Jesseba Fernando, Raj Deshpande, Alan Perotti, Raja Marjieh, Steven M. Frankland, Richard L. Lewis, Taylor W. Webb, Declan Campbell, Francesco Vaccarino, Jonathan D. Cohen, Giovanni Petri | Bound by semanticity: universal laws governing the generalization-identification tradeoff | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Intelligent systems must deploy internal representations that are simultaneously structured -- to support broad generalization -- and selective -- to preserve input identity. We expose a fundamental limit on this tradeoff. For any model whose representational similarity between inputs decays with finite semantic resolution $\varepsilon$, we derive closed-form expressions that pin its probability of correct generalization $p_S$ and identification $p_I$ to a universal Pareto front independent of input space geometry. Extending the analysis to noisy, heterogeneous spaces and to $n>2$ inputs predicts a sharp $1/n$ collapse of multi-input processing capacity and a non-monotonic optimum for $p_S$. A minimal ReLU network trained end-to-end reproduces these laws: during learning a resolution boundary self-organizes and empirical $(p_S,p_I)$ trajectories closely follow theoretical curves for linearly decaying similarity. Finally, we demonstrate that the same limits persist in two markedly more complex settings -- a convolutional neural network and state-of-the-art vision-language models -- confirming that finite-resolution similarity is a fundamental emergent informational constraint, not merely a toy-model artifact. Together, these results provide an exact theory of the generalization-identification trade-off and clarify how semantic resolution shapes the representational capacity of deep networks and brains alike. | [
{
"version": "v1",
"created": "Sun, 01 Jun 2025 15:56:26 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Nurisso",
"Marco",
""
],
[
"Fernando",
"Jesseba",
""
],
[
"Deshpande",
"Raj",
""
],
[
"Perotti",
"Alan",
""
],
[
"Marjieh",
"Raja",
""
],
[
"Frankland",
"Steven M.",
""
],
[
"Lewis",
"Richard L.",
""
],
[
"Webb",
"Taylor W.",
""
],
[
"Campbell",
"Declan",
""
],
[
"Vaccarino",
"Francesco",
""
],
[
"Cohen",
"Jonathan D.",
""
],
[
"Petri",
"Giovanni",
""
]
] |
2506.14798 | Siwei Tu | Siwei Tu, Jingyi Xu, Weidong Yang, Lei Bai, Ben Fei | MODS: Multi-source Observations Conditional Diffusion Model for Meteorological State Downscaling | null | null | null | null | physics.ao-ph cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate acquisition of high-resolution surface meteorological conditions is critical for forecasting and simulating meteorological variables. Directly applying spatial interpolation methods to derive meteorological values at specific locations from low-resolution grid fields often yields results that deviate significantly from the actual conditions. Existing downscaling methods primarily rely on the coupling relationship between geostationary satellites and ERA5 variables as a condition. However, using brightness temperature data from geostationary satellites alone fails to comprehensively capture all the changes in meteorological variables in ERA5 maps. To address this limitation, we can use a wider range of satellite data to make more full use of its inversion effects on various meteorological variables, thus producing more realistic results across different meteorological variables. To further improve the accuracy of downscaling meteorological variables at any location, we propose the Multi-source Observation Down-Scaling Model (MODS). It is a conditional diffusion model that fuses data from multiple geostationary satellites GridSat, polar-orbiting satellites (AMSU-A, HIRS, and MHS), and topographic data (GEBCO), as conditions, and is pre-trained on the ERA5 reanalysis dataset. During training, latent features from diverse conditional inputs are extracted separately and fused into ERA5 maps via a multi-source cross-attention module. By exploiting the inversion relationships between reanalysis data and multi-source atmospheric variables, MODS generates atmospheric states that align more closely with real-world conditions. During sampling, MODS enhances downscaling consistency by incorporating low-resolution ERA5 maps and station-level meteorological data as guidance. Experimental results demonstrate that MODS achieves higher fidelity when downscaling ERA5 maps to a 6.25 km resolution. | [
{
"version": "v1",
"created": "Mon, 02 Jun 2025 08:40:16 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Tu",
"Siwei",
""
],
[
"Xu",
"Jingyi",
""
],
[
"Yang",
"Weidong",
""
],
[
"Bai",
"Lei",
""
],
[
"Fei",
"Ben",
""
]
] |
2506.14799 | Evdoxia Taka | Evdoxia Taka, Debadyuti Bhattacharya, Joanne Garde-Hansen, Sanjay Sharma, Tanaya Guha | Analyzing Character Representation in Media Content using Multimodal Foundation Model: Effectiveness and Trust | null | null | null | null | cs.HC cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in AI has enabled automated analysis of complex media content at scale and generate actionable insights regarding character representation along such dimensions as gender and age. Past work focused on quantifying representation from audio/video/text using various ML models, but without having the audience in the loop. We ask, even if character distribution along demographic dimensions are available, how useful are they to the general public? Do they actually trust the numbers generated by AI models? Our work addresses these questions through a user study, while proposing a new AI-based character representation and visualization tool. Our tool based on the Contrastive Language Image Pretraining (CLIP) foundation model to analyze visual screen data to quantify character representation across dimensions of age and gender. We also designed effective visualizations suitable for presenting such analytics to lay audience. Next, we conducted a user study to seek empirical evidence on the usefulness and trustworthiness of the AI-generated results for carefully chosen movies presented in the form of our visualizations. We note that participants were able to understand the analytics from our visualization, and deemed the tool `overall useful'. Participants also indicated a need for more detailed visualizations to include more demographic categories and contextual information of the characters. Participants' trust in AI-based gender and age models is seen to be moderate to low, although they were not against the use of AI in this context. Our tool including code, benchmarking, and data from the user study can be found here: https://anonymous.4open.science/r/Character-Representation-Media-FF7B | [
{
"version": "v1",
"created": "Mon, 02 Jun 2025 13:46:28 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Taka",
"Evdoxia",
""
],
[
"Bhattacharya",
"Debadyuti",
""
],
[
"Garde-Hansen",
"Joanne",
""
],
[
"Sharma",
"Sanjay",
""
],
[
"Guha",
"Tanaya",
""
]
] |
2506.14800 | Soheil Firooz | Soheil Firooz, B. Daya Reddy, Paul Steinmann | A micromorphic-based artificial diffusion method for stabilized finite element approximation of convection-diffusion problems | null | null | null | null | math.NA cs.NA math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a novel artificial diffusion method to circumvent the instabilities associated with the standard finite element approximation of convection-diffusion equations. Motivated by the micromorphic approach, we introduce an auxiliary variable, which is related to the gradient of the field of interest, and which leads to a coupled problem. Conditions for well-posedness of the resulting formulation are established. We carry out a comprehensive numerical study to compare the proposed methodology against some well-established approaches in one- and two-dimensional settings. The proposed method outperforms established approaches in general in approximating accurately the solutions to pertinent and challenging problems. | [
{
"version": "v1",
"created": "Mon, 02 Jun 2025 15:47:18 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Firooz",
"Soheil",
""
],
[
"Reddy",
"B. Daya",
""
],
[
"Steinmann",
"Paul",
""
]
] |
2506.14801 | Priyam Das | Priyam Das | GLASD: A Loss-Function-Agnostic Global Optimizer for Robust Correlation Estimation under Data Contamination and Heavy Tails | null | null | null | null | stat.AP math.OC stat.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Robust correlation estimation is essential in high-dimensional settings, particularly when data are contaminated by outliers or exhibit heavy-tailed behavior. Many robust loss functions of practical interest-such as those involving truncation or redescending M-estimators-lead to objective functions that are inherently non-convex and non-differentiable. Traditional methods typically focus on a single loss function tailored to a specific contamination model and develop custom algorithms tightly coupled with that loss, limiting generality and adaptability. We introduce GLASD (Global Adaptive Stochastic Descent), a general-purpose black-box optimization algorithm designed to operate over the manifold of positive definite correlation matrices. Unlike conventional solvers, GLASD requires no gradient information and imposes no assumptions of convexity or smoothness, making it ideally suited for optimizing a wide class of loss functions-including non-convex, non-differentiable, or discontinuous objectives. This flexibility allows GLASD to serve as a unified framework for robust estimation under arbitrary user-defined criteria. We demonstrate its effectiveness through extensive simulations involving contaminated and heavy-tailed distributions, as well as a real-data application to breast cancer proteomic network inference, where GLASD successfully identifies biologically plausible interactions despite the presence of outliers. The proposed method is scalable, constraint-aware, and available as open-source software at GitHub. | [
{
"version": "v1",
"created": "Mon, 02 Jun 2025 23:34:09 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Das",
"Priyam",
""
]
] |
2506.14802 | Zuochen Ye | Zuochen Ye | ss-Mamba: Semantic-Spline Selective State-Space Model | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose ss-Mamba, a novel foundation model that enhances time series forecasting by integrating semantic-aware embeddings and adaptive spline-based temporal encoding within a selective state-space modeling framework. Building upon the recent success of Transformer architectures, ss-Mamba adopts the Mamba selective state space model as an efficient alternative that achieves comparable performance while significantly reducing computational complexity from quadratic to linear time. Semantic index embeddings, initialized from pretrained language models, allow effective generalization to previously unseen series through meaningful semantic priors. Additionally, spline-based Kolmogorov-Arnold Networks (KAN) dynamically and interpretably capture complex seasonalities and non-stationary temporal effects, providing a powerful enhancement over conventional temporal feature encodings. Extensive experimental evaluations confirm that ss-Mamba delivers superior accuracy, robustness, and interpretability, demonstrating its capability as a versatile and computationally efficient alternative to traditional Transformer-based models in time-series forecasting. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 03:26:57 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Ye",
"Zuochen",
""
]
] |
2506.14803 | Arbind Agrahari Baniya | Arbind Agrahari Baniya, Tsz-Kwan Lee, Peter W. Eklund, and Sunil Aryal | Omnidirectional Video Super-Resolution using Deep Learning | null | in IEEE Transactions on Multimedia, vol. 26, pp. 540-554, 2024 | 10.1109/TMM.2023.3267294 | null | cs.MM cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Omnidirectional Videos (or 360{\deg} videos) are widely used in Virtual Reality (VR) to facilitate immersive and interactive viewing experiences. However, the limited spatial resolution in 360{\deg} videos does not allow for each degree of view to be represented with adequate pixels, limiting the visual quality offered in the immersive experience. Deep learning Video Super-Resolution (VSR) techniques used for conventional videos could provide a promising software-based solution; however, these techniques do not tackle the distortion present in equirectangular projections of 360{\deg} video signals. An additional obstacle is the limited availability of 360{\deg} video datasets for study. To address these issues, this paper creates a novel 360{\deg} Video Dataset (360VDS) with a study of the extensibility of conventional VSR models to 360{\deg} videos. This paper further proposes a novel deep learning model for 360{\deg} Video Super-Resolution (360{\deg} VSR), called Spherical Signal Super-resolution with a Proportioned Optimisation (S3PO). S3PO adopts recurrent modelling with an attention mechanism, unbound from conventional VSR techniques like alignment. With a purpose-built feature extractor and a novel loss function addressing spherical distortion, S3PO outperforms most state-of-the-art conventional VSR models and 360{\deg}~specific super-resolution models on 360{\deg} video datasets. A step-wise ablation study is presented to understand and demonstrate the impact of the chosen architectural sub-components, targeted training and optimisation. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 05:59:21 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Baniya",
"Arbind Agrahari",
""
],
[
"Lee",
"Tsz-Kwan",
""
],
[
"Eklund",
"Peter W.",
""
],
[
"Aryal",
"Sunil",
""
]
] |
2506.14804 | Allen Lobo | Allen Lobo | On the dynamical evolution of randomness Part B: Geometrisation and the origin of convergence in LLN | null | null | null | null | physics.data-an math.PR | http://creativecommons.org/licenses/by/4.0/ | In classical probability theory, the convergence of empirical frequencies to theoretical probabilities: as captured by the Law of Large Numbers (LLN): is treated as axiomatic and emergent from statistical assumptions such as independence and identical distribution. In this work, a novel dynamical framework is constructed in which convergence arises as a consequence of structured evolution in outcome space, rather than a statistical postulate. Through this formalism, statistical convergence is derived dynamically, revealing an internal structure to randomness and exposing entanglement between successive trials. The system recovers classical LLN behaviour in the large-number limit, while predicting deviations, transient fluctuations, and geometric asymmetries in the early regime. This work inaugurates a new paradigm: dynamical probability mechanics: in which randomness is modelled not as a sequence of disconnected stochastic events, but as a physically structured, feedback-driven process. The theory provides a novel explanatory layer beneath statistical laws and opens pathways toward a mechanistic foundation of probability itself. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 07:03:04 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Lobo",
"Allen",
""
]
] |
2506.14805 | Yang Yao | Yang Yao, Lingyu Li, Jiaxin Song, Chiyu Chen, Zhenqi He, Yixu Wang, Xin Wang, Tianle Gu, Jie Li, Yan Teng, Yingchun Wang | Argus Inspection: Do Multimodal Large Language Models Possess the Eye of Panoptes? | null | null | null | null | cs.CV cs.AI cs.CL cs.LG cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As Multimodal Large Language Models (MLLMs) continue to evolve, their cognitive and reasoning capabilities have seen remarkable progress. However, challenges in visual fine-grained perception and commonsense causal inference persist. This paper introduces Argus Inspection, a multimodal benchmark with two levels of difficulty, emphasizing detailed visual recognition while incorporating real-world commonsense understanding to evaluate causal reasoning abilities. Expanding on it, we present the Eye of Panoptes framework, which integrates a binary parametric Sigmoid metric with an indicator function, enabling a more holistic evaluation of MLLMs' responses in opinion-based reasoning tasks. Experiments conducted on 26 mainstream MLLMs reveal that the highest performance in visual fine-grained reasoning reaches only 0.46, highlighting considerable potential for enhancement. Our research offers valuable perspectives for the continued refinement of MLLMs. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 13:44:14 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Yao",
"Yang",
""
],
[
"Li",
"Lingyu",
""
],
[
"Song",
"Jiaxin",
""
],
[
"Chen",
"Chiyu",
""
],
[
"He",
"Zhenqi",
""
],
[
"Wang",
"Yixu",
""
],
[
"Wang",
"Xin",
""
],
[
"Gu",
"Tianle",
""
],
[
"Li",
"Jie",
""
],
[
"Teng",
"Yan",
""
],
[
"Wang",
"Yingchun",
""
]
] |
2506.14806 | Bochen Lyu | Bochen Lyu, Xiaojing Zhang, Fangyi Zheng, He Wang, Zheng Wang, Zhanxing Zhu | Heavy-Ball Momentum Method in Continuous Time and Discretization Error Analysis | 32 pages, 7 figures | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper establishes a continuous time approximation, a piece-wise continuous differential equation, for the discrete Heavy-Ball (HB) momentum method with explicit discretization error. Investigating continuous differential equations has been a promising approach for studying the discrete optimization methods. Despite the crucial role of momentum in gradient-based optimization methods, the gap between the original discrete dynamics and the continuous time approximations due to the discretization error has not been comprehensively bridged yet. In this work, we study the HB momentum method in continuous time while putting more focus on the discretization error to provide additional theoretical tools to this area. In particular, we design a first-order piece-wise continuous differential equation, where we add a number of counter terms to account for the discretization error explicitly. As a result, we provide a continuous time model for the HB momentum method that allows the control of discretization error to arbitrary order of the step size. As an application, we leverage it to find a new implicit regularization of the directional smoothness and investigate the implicit bias of HB for diagonal linear networks, indicating how our results can be used in deep learning. Our theoretical findings are further supported by numerical experiments. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 14:59:22 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Lyu",
"Bochen",
""
],
[
"Zhang",
"Xiaojing",
""
],
[
"Zheng",
"Fangyi",
""
],
[
"Wang",
"He",
""
],
[
"Wang",
"Zheng",
""
],
[
"Zhu",
"Zhanxing",
""
]
] |
2506.14807 | Eric Ngondiep | Eric Ngondiep | An explicit computational approach for a three-dimensional system of nonlinear elastodynamic sine-Gordon problem | 21 pages, 20 figures, 6 tables | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by-sa/4.0/ | This paper proposes an explicit computational method for solving a three-dimensional system of nonlinear elastodynamic sine-Gordon equations subject to appropriate initial and boundary conditions. The time derivative is approximated by interpolation technique whereas the finite element approach is used to approximate the space derivatives. The developed numerical scheme is so-called, high-order explicit computational technique. The new algorithm efficiently treats the time derivative term and provides a suitable time step restriction for stability and convergence. Under this time step limitation, both stability and error estimates of the proposed approach are deeply analyzed using a constructed strong norm. The theoretical studies indicate that the developed approach is temporal second-order convergent and spatially third-order accurate. Some numerical examples are carried out to confirm the theory, to validate the computational efficiency and to demonstrate the practical applicability of the new computational technique. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 15:35:42 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Ngondiep",
"Eric",
""
]
] |
2506.14808 | Jenny Schmalfuss | Jenny Schmalfuss and Nadine Chang and Vibashan VS and Maying Shen and Andres Bruhn and Jose M. Alvarez | PARC: A Quantitative Framework Uncovering the Symmetries within Vision Language Models | Accepted to CVPR 2025 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision language models (VLMs) respond to user-crafted text prompts and visual inputs, and are applied to numerous real-world problems. VLMs integrate visual modalities with large language models (LLMs), which are well known to be prompt-sensitive. Hence, it is crucial to determine whether VLMs inherit this instability to varying prompts. We therefore investigate which prompt variations VLMs are most sensitive to and which VLMs are most agnostic to prompt variations. To this end, we introduce PARC (Prompt Analysis via Reliability and Calibration), a VLM prompt sensitivity analysis framework built on three pillars: (1) plausible prompt variations in both the language and vision domain, (2) a novel model reliability score with built-in guarantees, and (3) a calibration step that enables dataset- and prompt-spanning prompt variation analysis. Regarding prompt variations, PARC's evaluation shows that VLMs mirror LLM language prompt sensitivity in the vision domain, and most destructive variations change the expected answer. Regarding models, outstandingly robust VLMs among 22 evaluated models come from the InternVL2 family. We further find indications that prompt sensitivity is linked to training data. The code will be at https://github.com/NVlabs/PARC. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:42:32 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Schmalfuss",
"Jenny",
""
],
[
"Chang",
"Nadine",
""
],
[
"VS",
"Vibashan",
""
],
[
"Shen",
"Maying",
""
],
[
"Bruhn",
"Andres",
""
],
[
"Alvarez",
"Jose M.",
""
]
] |
2506.14809 | Peng Jiang | Peng Jiang and Vinicius Cezar Monteiro de Lira and Antonio Maiorino | Impact of a Deployed LLM Survey Creation Tool through the IS Success Model | null | null | null | null | cs.HC cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Surveys are a cornerstone of Information Systems (IS) research, yet creating high-quality surveys remains labor-intensive, requiring both domain expertise and methodological rigor. With the evolution of large language models (LLMs), new opportunities emerge to automate survey generation. This paper presents the real-world deployment of an LLM-powered system designed to accelerate data collection while maintaining survey quality. Deploying such systems in production introduces real-world complexity, including diverse user needs and quality control. We evaluate the system using the DeLone and McLean IS Success Model to understand how generative AI can reshape a core IS method. This study makes three key contributions. To our knowledge, this is the first application of the IS Success Model to a generative AI system for survey creation. In addition, we propose a hybrid evaluation framework combining automated and human assessments. Finally, we implement safeguards that mitigate post-deployment risks and support responsible integration into IS workflows. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:36:36 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Jiang",
"Peng",
""
],
[
"de Lira",
"Vinicius Cezar Monteiro",
""
],
[
"Maiorino",
"Antonio",
""
]
] |
2506.14810 | Qiwen Zhang | Qiwen Zhang | Intelligent Routing for Sparse Demand Forecasting: A Comparative Evaluation of Selection Strategies | 7 pages, 4 figures, conference | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Sparse and intermittent demand forecasting in supply chains presents a critical challenge, as frequent zero-demand periods hinder traditional model accuracy and impact inventory management. We propose and evaluate a Model-Router framework that dynamically selects the most suitable forecasting model-spanning classical, ML, and DL methods for each product based on its unique demand pattern. By comparing rule-based, LightGBM, and InceptionTime routers, our approach learns to assign appropriate forecasting strategies, effectively differentiating between smooth, lumpy, or intermittent demand regimes to optimize predictions. Experiments on the large-scale Favorita dataset show our deep learning (Inception Time) router improves forecasting accuracy by up to 11.8% (NWRMSLE) over strong, single-model benchmarks with 4.67x faster inference time. Ultimately, these gains in forecasting precision will drive substantial reductions in both stockouts and wasteful excess inventory, underscoring the critical role of intelligent, adaptive Al in optimizing contemporary supply chain operations. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 03:09:45 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Zhang",
"Qiwen",
""
]
] |
2506.14811 | Mikel Malag\'on | Mikel Malag\'on, Josu Ceberio, Jose A. Lozano | Self-Composing Policies for Scalable Continual Reinforcement Learning | ICML 2024 (oral) | null | null | null | cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | This work introduces a growable and modular neural network architecture that naturally avoids catastrophic forgetting and interference in continual reinforcement learning. The structure of each module allows the selective combination of previous policies along with its internal policy, accelerating the learning process on the current task. Unlike previous growing neural network approaches, we show that the number of parameters of the proposed approach grows linearly with respect to the number of tasks, and does not sacrifice plasticity to scale. Experiments conducted in benchmark continuous control and visual problems reveal that the proposed approach achieves greater knowledge transfer and performance than alternative methods. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 15:12:15 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Malagón",
"Mikel",
""
],
[
"Ceberio",
"Josu",
""
],
[
"Lozano",
"Jose A.",
""
]
] |
2506.14812 | Zhihang Xu | Zhihang Xu, Min Wang, Zhu Wang | Weak TransNet: A Petrov-Galerkin based neural network method for solving elliptic PDEs | null | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by/4.0/ | While deep learning has achieved remarkable success in solving partial differential equations (PDEs), it still faces significant challenges, particularly when the PDE solutions have low regularity or singularities. To address these issues, we propose the Weak TransNet (WTN) method, based on a Petrov-Galerkin formulation, for solving elliptic PDEs in this work, though its framework may extend to other classes of equations. Specifically, the neural feature space defined by TransNet (Zhang et al., 2023) is used as the trial space, while the test space is composed of radial basis functions. Since the solution is expressed as a linear combination of trial functions, the coefficients can be determined by minimizing the weak PDE residual via least squares. Thus, this approach could help mitigate the challenges of non-convexity and ill-conditioning that often arise in neural network training. Furthermore, the WTN method is extended to handle problems whose solutions exhibit multiscale features or possess sharp variations. Several numerical experiments are presented to demonstrate the robustness and efficiency of the proposed methods. | [
{
"version": "v1",
"created": "Thu, 05 Jun 2025 19:06:10 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Xu",
"Zhihang",
""
],
[
"Wang",
"Min",
""
],
[
"Wang",
"Zhu",
""
]
] |
2506.14813 | Peng Huang | Yuxuan Jiang, Ziming Zhou, Boyu Xu, Beijie Liu, Runhui Xu, Peng Huang | Training with Confidence: Catching Silent Errors in Deep Learning Training with Automated Proactive Checks | 19 pages, to appear in 19th USENIX Symposium on Operating Systems Design and Implementation (OSDI '25) | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Training deep learning (DL) models is a complex process, making it prone to silent errors that are challenging to detect and diagnose. This paper presents TRAINCHECK, a framework that takes a proactive checking approach to address silent training errors. TRAINCHECK automatically infers invariants tailored for DL training. It uses these invariants to proactively detect silent errors during the training process while providing debugging help. To evaluate TRAINCHECK, we reproduce 20 real-world silent training errors with diverse root causes. TRAINCHECK successfully detects 18 errors within a single training iteration. It also uncovers 6 unknown bugs in popular training libraries that lead to silent errors. | [
{
"version": "v1",
"created": "Fri, 06 Jun 2025 05:34:42 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Jiang",
"Yuxuan",
""
],
[
"Zhou",
"Ziming",
""
],
[
"Xu",
"Boyu",
""
],
[
"Liu",
"Beijie",
""
],
[
"Xu",
"Runhui",
""
],
[
"Huang",
"Peng",
""
]
] |
2506.14814 | Amit Verma Kumar | Ankita Yadav and Amit K. Verma | Semi-orthogonal Tribonacci Wavelets and Numerical Solutions of Nonlinear Singular BVPs Arising in a Chemical Reaction | null | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by/4.0/ | In this article, we introduce a semi-orthogonal tribonacci wavelet and develop a semi-orthogonal tribonacci wavelet collocation method, offering an effective numerical method for solving a class of non-linear singular BVPs. | [
{
"version": "v1",
"created": "Fri, 06 Jun 2025 05:59:12 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Yadav",
"Ankita",
""
],
[
"Verma",
"Amit K.",
""
]
] |
2506.14815 | Gyaneshwar Agrahari | Gyaneshwar Agrahari, Kiran Bist, Monika Pandey, Jacob Kapita, Zachary James, Jackson Knox, Steven Heymsfield, Sophia Ramirez, Peter Wolenski, Nadejda Drenska | Predicting Anthropometric Body Composition Variables Using 3D Optical Imaging and Machine Learning | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Accurate prediction of anthropometric body composition variables, such as Appendicular Lean Mass (ALM), Body Fat Percentage (BFP), and Bone Mineral Density (BMD), is essential for early diagnosis of several chronic diseases. Currently, researchers rely on Dual-Energy X-ray Absorptiometry (DXA) scans to measure these metrics; however, DXA scans are costly and time-consuming. This work proposes an alternative to DXA scans by applying statistical and machine learning models on biomarkers (height, volume, left calf circumference, etc) obtained from 3D optical images. The dataset consists of 847 patients and was sourced from Pennington Biomedical Research Center. Extracting patients' data in healthcare faces many technical challenges and legal restrictions. However, most supervised machine learning algorithms are inherently data-intensive, requiring a large amount of training data. To overcome these limitations, we implemented a semi-supervised model, the $p$-Laplacian regression model. This paper is the first to demonstrate the application of a $p$-Laplacian model for regression. Our $p$-Laplacian model yielded errors of $\sim13\%$ for ALM, $\sim10\%$ for BMD, and $\sim20\%$ for BFP when the training data accounted for 10 percent of all data. Among the supervised algorithms we implemented, Support Vector Regression (SVR) performed the best for ALM and BMD, yielding errors of $\sim 8\%$ for both, while Least Squares SVR performed the best for BFP with $\sim 11\%$ error when trained on 80 percent of the data. Our findings position the $p$-Laplacian model as a promising tool for healthcare applications, particularly in a data-constrained environment. | [
{
"version": "v1",
"created": "Sun, 08 Jun 2025 03:42:56 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Agrahari",
"Gyaneshwar",
""
],
[
"Bist",
"Kiran",
""
],
[
"Pandey",
"Monika",
""
],
[
"Kapita",
"Jacob",
""
],
[
"James",
"Zachary",
""
],
[
"Knox",
"Jackson",
""
],
[
"Heymsfield",
"Steven",
""
],
[
"Ramirez",
"Sophia",
""
],
[
"Wolenski",
"Peter",
""
],
[
"Drenska",
"Nadejda",
""
]
] |
2506.14816 | Alavikunhu Panthakkan Dr | Alavikunhu Panthakkan, Zubair Medammal, S M Anzar, Fatma Taher and Hussain Al-Ahmad | A Hybrid ConvNeXt-EfficientNet AI Solution for Precise Falcon Disease Detection | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Falconry, a revered tradition involving the training and hunting with falcons, requires meticulous health surveillance to ensure the health and safety of these prized birds, particularly in hunting scenarios. This paper presents an innovative method employing a hybrid of ConvNeXt and EfficientNet AI models for the classification of falcon diseases. The study focuses on accurately identifying three conditions: Normal, Liver Disease and 'Aspergillosis'. A substantial dataset was utilized for training and validating the model, with an emphasis on key performance metrics such as accuracy, precision, recall, and F1-score. Extensive testing and analysis have shown that our concatenated AI model outperforms traditional diagnostic methods and individual model architectures. The successful implementation of this hybrid AI model marks a significant step forward in precise falcon disease detection and paves the way for future developments in AI-powered avian healthcare solutions. | [
{
"version": "v1",
"created": "Sun, 08 Jun 2025 17:01:44 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Panthakkan",
"Alavikunhu",
""
],
[
"Medammal",
"Zubair",
""
],
[
"Anzar",
"S M",
""
],
[
"Taher",
"Fatma",
""
],
[
"Al-Ahmad",
"Hussain",
""
]
] |
2506.14817 | Simon Polichinel Von Der Maase PhD | Simon P. von der Maase | Next-Generation Conflict Forecasting: Unleashing Predictive Patterns through Spatiotemporal Learning | 33 pages, 9 figures, 3 tables. Presented at workshops hosted by PRIO, AFK (German Association for Peace and Conflict Studies), CCEW (Bundeswehr University Munich), Uppsala University, SODAS (University of Copenhagen) and in briefings with UN agencies including UNIDIR, OCHA, and FAO | null | null | null | stat.OT cs.AI cs.CY cs.LG stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Forecasting violent conflict at high spatial and temporal resolution remains a central challenge for both researchers and policymakers. This study presents a novel neural network architecture for forecasting three distinct types of violence -- state-based, non-state, and one-sided -- at the subnational (priogrid-month) level, up to 36 months in advance. The model jointly performs classification and regression tasks, producing both probabilistic estimates and expected magnitudes of future events. It achieves state-of-the-art performance across all tasks and generates approximate predictive posterior distributions to quantify forecast uncertainty.
The architecture is built on a Monte Carlo Dropout Long Short-Term Memory (LSTM) U-Net, integrating convolutional layers to capture spatial dependencies with recurrent structures to model temporal dynamics. Unlike many existing approaches, it requires no manual feature engineering and relies solely on historical conflict data. This design enables the model to autonomously learn complex spatiotemporal patterns underlying violent conflict.
Beyond achieving state-of-the-art predictive performance, the model is also highly extensible: it can readily integrate additional data sources and jointly forecast auxiliary variables. These capabilities make it a promising tool for early warning systems, humanitarian response planning, and evidence-based peacebuilding initiatives. | [
{
"version": "v1",
"created": "Sun, 08 Jun 2025 20:42:29 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"von der Maase",
"Simon P.",
""
]
] |
2506.14818 | Tao Liu | Tao Liu, Juhao Wu, Mark Ying | Non-degenerate Ground State of the Spin-Boson Model under Abelian Diagonalization | null | null | null | null | cond-mat.str-el quant-ph | http://creativecommons.org/licenses/by/4.0/ | By utilizing a unitary transformation, we derive the necessary and sufficient conditions for the degeneracy between the even- and odd-parity energy states of the spin-boson model (SBM). Employing the Rayleigh quotient of matrix algebra, we rigorously prove that the ground state energy of the SBM is lower than the systems lowest possible degenerate energy and possesses a definite parity. Based on the necessary and sufficient conditions for parity breaking, we provide an analytical expression for the parity-breaking critical value, which is closely related to the expansion order and computational accuracy. This expression reproduces the SBM phase diagram obtained by quantum Monte Carlo (QMC) and logarithmic discretization numerical renormalization group (NRG) methods. However, this phase diagram does not characterize the ground state of the system. | [
{
"version": "v1",
"created": "Mon, 09 Jun 2025 15:34:30 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Liu",
"Tao",
""
],
[
"Wu",
"Juhao",
""
],
[
"Ying",
"Mark",
""
]
] |
2506.14819 | Cristina Lenardi | Simone Minelli, Manuel Iozzia, Claudio Piazzoni, Alberto Vertova, Cristina Lenardi, Alessandro Minguzzi | Mixed platinum and zirconia powder as electrocatalyst for hydrogen evolution and oxidation reaction | Manuscript: 18 pages, 6 figures, 5 tables; Supplementary info: 2 pages, 2 figures | null | null | null | physics.chem-ph cond-mat.mtrl-sci | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The development of efficient and cost-effective bifunctional electrocatalysts for hydrogen-related technologies is crucial for the transition to sustainable energy. In this work, we present Pt@ZrO2 as a high-performance ceramic-based electrocatalyst for the hydrogen evolution reaction (HER) and hydrogen oxidation reaction (HOR), demonstrating superior activity and stability in alkaline conditions. Compared to Pd@ZrO2, previously investigated in our earlier work, Pt@ZrO2 exhibits significantly lower charge transfer resistance and enhanced HER/HOR kinetics, as confirmed by electrochemical impedance spectroscopy. The synergistic interaction between platinum and ZrO2 stabilizes Pt in a partially oxidized state, facilitating the Volmer step and improving charge transfer kinetics. Cyclic voltammetry further confirms the remarkable electrochemical stability of Pt@ZrO2 under stress conditions. Importantly, Pt@ZrO2 offers a cost-effective alternative to pure platinum by reducing noble metal loading while maintaining excellent catalytic performance. These findings establish Pt@ZrO2 as a better alternative to Pd@ZrO2 and a promising candidate for next-generation hydrogen energy applications. | [
{
"version": "v1",
"created": "Mon, 09 Jun 2025 16:18:05 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Minelli",
"Simone",
""
],
[
"Iozzia",
"Manuel",
""
],
[
"Piazzoni",
"Claudio",
""
],
[
"Vertova",
"Alberto",
""
],
[
"Lenardi",
"Cristina",
""
],
[
"Minguzzi",
"Alessandro",
""
]
] |
2506.14820 | Hyeon Jeon | Hyeon Jeon, Hyunwook Lee, Yun-Hsin Kuo, Taehyun Yang, Daniel Archambault, Sungahn Ko, Takanori Fujiwara, Kwan-Liu Ma, Jinwook Seo | Navigating High-Dimensional Backstage: A Guide for Exploring Literature for the Reliable Use of Dimensionality Reduction | EG/VGTC EuroVis 2025 Short paper | null | 10.2312/evs.20251087 | null | cs.HC cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Visual analytics using dimensionality reduction (DR) can easily be unreliable for various reasons, e.g., inherent distortions in representing the original data. The literature has thus proposed a wide range of methodologies to make DR-based visual analytics reliable. However, the diversity and extensiveness of the literature can leave novice analysts and researchers uncertain about where to begin and proceed. To address this problem, we propose a guide for reading papers for reliable visual analytics with DR. Relying on the previous classification of the relevant literature, our guide helps both practitioners to (1) assess their current DR expertise and (2) identify papers that will further enhance their understanding. Interview studies with three experts in DR and data visualizations validate the significance, comprehensiveness, and usefulness of our guide. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:06:54 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Jeon",
"Hyeon",
""
],
[
"Lee",
"Hyunwook",
""
],
[
"Kuo",
"Yun-Hsin",
""
],
[
"Yang",
"Taehyun",
""
],
[
"Archambault",
"Daniel",
""
],
[
"Ko",
"Sungahn",
""
],
[
"Fujiwara",
"Takanori",
""
],
[
"Ma",
"Kwan-Liu",
""
],
[
"Seo",
"Jinwook",
""
]
] |
2506.14821 | Sunil Kumar | Sunil Kumar, Bowen Zhao, Leo Dirac, Paulina Varshavskaya | Reinforcing VLMs to Use Tools for Detailed Visual Reasoning Under Resource Constraints | null | null | null | null | cs.LG cs.AI cs.CV | http://creativecommons.org/licenses/by-sa/4.0/ | Despite tremendous recent advances in large model reasoning ability, vision-language models (VLMs) still struggle with detailed visual reasoning, especially when compute resources are limited. To address this challenge, we draw inspiration from methods like Deepseek-r1 for VLMs and train smaller-scale models with Group Relative Policy Optimization (GRPO) to use external tools such as zoom. The greatest benefit is obtained with a combination of GRPO learning, a simple reward structure, a simplified tool-calling interface, allocating additional tokens to the result of the tool call, and a training data mix that over-represents visually difficult examples. Compared to similarly-sized baseline models, our method achieves better performance on some visual question-answering (VQA) tasks, thanks to the detailed visual information gathered from the external tool. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:11:44 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Kumar",
"Sunil",
""
],
[
"Zhao",
"Bowen",
""
],
[
"Dirac",
"Leo",
""
],
[
"Varshavskaya",
"Paulina",
""
]
] |
2506.14822 | Konstantin Rybakov | Tatyana A. Averina, Konstantin A. Rybakov | Analysis and conditional optimization of projection estimates for the distribution of random variable using Legendre polynomials | null | null | null | null | stat.CO cs.NA math.NA math.PR | http://creativecommons.org/licenses/by/4.0/ | Algorithms for jointly obtaining projection estimates of the density and distribution function of a random variable using the Legendre polynomials are proposed. For these algorithms, a problem of the conditional optimization is solved. Such an optimization allows one increasing the approximation accuracy with a minimum computational costs. The proposed algorithms are tested on examples with different degree of smoothness of the density. | [
{
"version": "v1",
"created": "Thu, 12 Jun 2025 16:10:02 GMT"
}
] | 2025-06-19T00:00:00 | [
[
"Averina",
"Tatyana A.",
""
],
[
"Rybakov",
"Konstantin A.",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.