id
stringlengths 9
16
| submitter
stringlengths 1
64
⌀ | authors
stringlengths 3
62.8k
| title
stringlengths 1
382
| comments
stringlengths 1
1.15k
⌀ | journal-ref
stringlengths 1
557
⌀ | doi
stringlengths 7
153
⌀ | report-no
stringlengths 1
509
⌀ | categories
stringlengths 5
125
| license
stringclasses 9
values | abstract
stringlengths 6
6.09k
| versions
listlengths 1
187
| update_date
timestamp[s]date 2007-05-23 00:00:00
2025-06-27 00:00:00
| authors_parsed
sequencelengths 1
2.83k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2506.17957 | Hao Li | Hao Li and Jie Zhu | Static Spherical Vacuum Solution to Bumblebee Gravity with Time-like VEVs | 6 pages, 0 figures | null | null | null | gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The static spherical vacuum solution in a bumblebee gravity model where the bumblebee field \(B_\mu\) has a time-like vacuum expectation value \(b_\mu\) is studied. We show that in general curved space-time solutions are not allowed and only the Minkowski space-time exists. However, it is surprising that non-trivial solutions can be obtained so long as a unique condition for the vacuum expectation \(b^2\equiv -b^\mu b_\mu=2/\kappa\), where \(\kappa=8\pi G\), is satisfied. We argue that naturally these solutions are not stable since quantum corrections would invalidate the likely numerical coincidence, unless there are some unknown fine-tuning mechanisms preventing any deviation from this condition. Nevertheless, some properties of these novel but peculiar solutions are discussed, and we show that the extremal Reissner-Nordstr{\"o}m solution is a limit of one of our solutions. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 09:22:56 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 11:40:35 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Li",
"Hao",
""
],
[
"Zhu",
"Jie",
""
]
] |
2506.17971 | Mahmoud Salim Dr. | Mahmoud M. Salim, Khaled M. Rabie, Ali H. Muqaibel | Robust Energy-Efficient DRL-Based Optimization in UAV-Mounted RIS Systems with Jitter | null | null | null | null | cs.IT eess.SP math.IT | http://creativecommons.org/licenses/by/4.0/ | In this letter, we propose an energy-efficient design for an unmanned aerial vehicle (UAV)-mounted reconfigurable intelligent surface (RIS) communication system with nonlinear energy harvesting (EH) and UAV jitter. A joint optimization problem is formulated to maximize the EH efficiency of the UAV-mounted RIS by controlling the user powers, RIS phase shifts, and time-switching factor, subject to quality of service and practical EH constraints. The problem is nonconvex and time-coupled due to UAV angular jitter and nonlinear EH dynamics, making it intractable for conventional optimization methods. To address this, we reformulate the problem as a deep reinforcement learning (DRL) environment and develop a smoothed softmax dual deep deterministic policy gradient algorithm. The proposed method incorporates action clipping, entropy regularization, and softmax-weighted Q-value estimation to improve learning stability and exploration. Simulation results show that the proposed algorithm converges reliably under various UAV jitter levels and achieves an average EH efficiency of 45.07\%, approaching the 53.09\% upper bound of exhaustive search, and outperforming other DRL baselines. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 10:06:14 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Salim",
"Mahmoud M.",
""
],
[
"Rabie",
"Khaled M.",
""
],
[
"Muqaibel",
"Ali H.",
""
]
] |
2506.18025 | Alistair Rowe | C. Fornos, N. ALyabyeva, Y. W. Ho, J. Peretti, A. C. H. Rowe, J. S. Speck, C. Weisbuch | Nanoscale imaging of reduced forward bias at V-pits in green-emitting nitride LEDs | 6 pages, 5 figures | null | null | null | physics.app-ph cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | Record wall-plug efficiencies in long-wavelength, nitride light-emitting diodes (LEDs) have recently been achieved in devices containing high V-pit densities. Numerical modeling suggests this may be due to improved electrical efficiencies (EE). In order to test this proposition, a novel scanning tunneling luminescence microscope (STLM) is used to map the local optoelectronic properties of commercial, green-emitting LED heterostructures around V-pits with nanoscale spatial resolution. Using the STLM tip as the hole injector, injection at the lips of V-pits is found to be drastically different from injection on the heterostructure's c-plane. A $\approx$ 3-fold improvement in internal quantum efficiency near V-pits is observed at low injection, and at higher injection a $\approx$1.75 V reduction in the forward bias unambiguously confirms the EE hypothesis for hole injection. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 13:12:14 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Fornos",
"C.",
""
],
[
"ALyabyeva",
"N.",
""
],
[
"Ho",
"Y. W.",
""
],
[
"Peretti",
"J.",
""
],
[
"Rowe",
"A. C. H.",
""
],
[
"Speck",
"J. S.",
""
],
[
"Weisbuch",
"C.",
""
]
] |
2506.18030 | Alok Shukla | Shashi Pandey, Swaroop Ganguly, Alok Shukla, Anurag Tripathi | Enhanced UV Photodetector Efficiency with a ZnO/Ga$_2$O$_3$ Heterojunction | 23 pages, 10 figures | ACS Appl. Electron. Mater. 7, 1173 (2025) | 10.1021/acsaelm.4c02067 | null | cond-mat.mtrl-sci physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Heterostructures comprising uncoated ZnO and coated with thin layers of Ga$_2$O$_3$ were produced using spin-coating and subsequent hydrothermal processing. X-ray diffraction examination verifies the structural integrity of the synthesized heterostructures (HTs). Optical and photoluminescence spectra were recorded to assess the variation in absorption and emission of the Ga$_2$O$_3$-coated HTs in comparison to the pristine ZnO. We conducted comparative density-functional theory (DFT) computations to corroborate the measured band gaps of both categories of HTs. To assess the stability of our devices, the transient response to on/off light switching under zero bias has been studied. The rise time $\tau_{r1}$ ($\tau_{r2}$) is 2300 (500) ms and the decay time $\tau_{d1}$ ($\tau_{d2}$) is 2700 (5000) ms have been observed for bare ZnO and ZnO/Ga$_2$O$_3$ HTs, respectively. A significant amount of change was also observed in the electrical transport properties from bare ZnO to ZnO/Ga$_2$O$_3$. To see the performance of device, responsivity (R) and detectivity (D = 1/NEP$_B$) have been measured. It is evident from observation that responsivity of a device shows maximum value in UV region while it is reducing with visible region for HTs. In case of detectivity, the maximum value reached was $145 \times 10^{14}$ Hz$^{1/2}$/W (at ~ 200 nm) and $38 \times 10^{14}$ Hz$^{1/2}$/W (at 300 nm) for Ga$_2$O$_3$ coated ZnO, and bare ZnO HTs, respectively. The maximum responsivity measured for the bare ZnO HTs is 7 (A/W) while that of Ga$_2$O$_3$ coated ZnO HTs is 38 (A/W). It suggests a simple way of designing materials for fabricating broad-range cost-effective photodetectors. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 13:20:13 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Pandey",
"Shashi",
""
],
[
"Ganguly",
"Swaroop",
""
],
[
"Shukla",
"Alok",
""
],
[
"Tripathi",
"Anurag",
""
]
] |
2506.18050 | Lyuye Zhang | Lyuye Zhang, Jian Zhang, Kaixuan Li, Chong Wang, Chengwei Liu, Jiahui Wu, Sen Chen, Yaowen Zheng, Yang Liu | VFArch\=e: A Dual-Mode Framework for Locating Vulnerable Functions in Open-Source Software | 15 pages | null | null | null | cs.SE | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Software Composition Analysis (SCA) has become pivotal in addressing vulnerabilities inherent in software project dependencies. In particular, reachability analysis is increasingly used in Open-Source Software (OSS) projects to identify reachable vulnerabilities (e.g., CVEs) through call graphs, enabling a focus on exploitable risks. Performing reachability analysis typically requires the vulnerable function (VF) to track the call chains from downstream applications. However, such crucial information is usually unavailable in modern vulnerability databases like NVD. While directly extracting VF from modified functions in vulnerability patches is intuitive, patches are not always available. Moreover, our preliminary study shows that over 26% of VF do not exist in the modified functions. Meanwhile, simply ignoring patches to search vulnerable functions suffers from overwhelming noises and lexical gaps between descriptions and source code. Given that almost half of the vulnerabilities are equipped with patches, a holistic solution that handles both scenarios with and without patches is required. To meet real-world needs and automatically localize VF, we present VFArch\=e, a dual-mode approach designed for disclosed vulnerabilities, applicable in scenarios with or without available patch links. The experimental results of VFArch\=e on our constructed benchmark dataset demonstrate significant efficacy regarding three metrics, achieving 1.3x and 1.9x Mean Reciprocal Rank over the best baselines for Patch-present and Patch-absent modes, respectively. Moreover, VFArch\=e has proven its applicability in real-world scenarios by successfully locating VF for 43 out of 50 latest vulnerabilities with reasonable efforts and significantly reducing 78-89% false positives of SCA tools. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 14:33:10 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 07:11:45 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Zhang",
"Lyuye",
""
],
[
"Zhang",
"Jian",
""
],
[
"Li",
"Kaixuan",
""
],
[
"Wang",
"Chong",
""
],
[
"Liu",
"Chengwei",
""
],
[
"Wu",
"Jiahui",
""
],
[
"Chen",
"Sen",
""
],
[
"Zheng",
"Yaowen",
""
],
[
"Liu",
"Yang",
""
]
] |
2506.18069 | Krzysztof Kutt | Klaudia Ropel, Krzysztof Kutt, Luiz do Valle Miranda, Grzegorz J. Nalepa | Unfolding the Past: A Comprehensive Deep Learning Approach to Analyzing Incunabula Pages | 10 pages, 8 figures; submitted to TPDL 2025; change in v2: updated e-mail address | null | null | null | cs.DL cs.CV | http://creativecommons.org/licenses/by/4.0/ | We developed a proof-of-concept method for the automatic analysis of the structure and content of incunabula pages. A custom dataset comprising 500 annotated pages from five different incunabula was created using resources from the Jagiellonian Digital Library. Each page was manually labeled with five predefined classes: Text, Title, Picture, Table, and Handwriting. Additionally, the publicly available DocLayNet dataset was utilized as supplementary training data. To perform object detection, YOLO11n and YOLO11s models were employed and trained using two strategies: a combined dataset (DocLayNet and the custom dataset) and the custom dataset alone. The highest performance (F1 = 0.94) was achieved by the YOLO11n model trained exclusively on the custom data. Optical character recognition was then conducted on regions classified as Text, using both Tesseract and Kraken OCR, with Tesseract demonstrating superior results. Subsequently, image classification was applied to the Picture class using a ResNet18 model, achieving an accuracy of 98.7% across five subclasses: Decorative_letter, Illustration, Other, Stamp, and Wrong_detection. Furthermore, the CLIP model was utilized to generate semantic descriptions of illustrations. The results confirm the potential of machine learning in the analysis of early printed books, while emphasizing the need for further advancements in OCR performance and visual content interpretation. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 15:33:20 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 11:19:13 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Ropel",
"Klaudia",
""
],
[
"Kutt",
"Krzysztof",
""
],
[
"Miranda",
"Luiz do Valle",
""
],
[
"Nalepa",
"Grzegorz J.",
""
]
] |
2506.18079 | Andreas Maeder | Andreas Maeder, Robert J. Chapman, Alessandra Sabatti, Giovanni Finco, Jost Kellner, Rachel Grange | Programmable Bell State Generation in an Integrated Thin Film Lithium Niobate Circuit | null | null | null | null | quant-ph physics.optics | http://creativecommons.org/licenses/by/4.0/ | Entanglement is central to quantum technologies such as cryptography, sensing, and computing. Photon pairs generated via nonlinear optical processes are excellent for preparing entangled states due to their long coherence times and compatibility with fiber optic networks. Steady progress in nanofabrication has positioned lithium niobate-on-insulator (LNOI) as a leading platform for monolithic integration of photon pair sources into optical circuits, leveraging its strong second-order nonlinearity. Here, we present a reconfigurable photonic integrated circuit on LNOI, which combines two on-chip photon pair sources with programmable interferometers, enabling generation of entangled states. The pair sources achieve a source brightness of 26 MHz nm$^{-1}$ mW$^{-1}$ while maintaining a coincidence-to-accidental ratio above 100. We successfully interfere the two sources with $99 \pm 0.7$ % visibility, demonstrating the indistinguishability required for producing entanglement on-chip. We show preparation of any of the maximally entangled Bell states with fidelity above 90 % verified by quantum state tomography. These results establish LNOI as a compelling, scalable platform to explore integrated quantum photonic technologies enabled by high-brightness sources of entangled quantum states. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 15:53:20 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Maeder",
"Andreas",
""
],
[
"Chapman",
"Robert J.",
""
],
[
"Sabatti",
"Alessandra",
""
],
[
"Finco",
"Giovanni",
""
],
[
"Kellner",
"Jost",
""
],
[
"Grange",
"Rachel",
""
]
] |
2506.18082 | Esteban Garces Arias | Esteban Garces Arias, Hannah Blocher, Julian Rodemann, Matthias A{\ss}enmacher, Christoph Jansen | Statistical Multicriteria Evaluation of LLM-Generated Text | null | null | null | null | cs.CL stat.AP | http://creativecommons.org/licenses/by/4.0/ | Assessing the quality of LLM-generated text remains a fundamental challenge in natural language processing. Current evaluation approaches often rely on isolated metrics or simplistic aggregations that fail to capture the nuanced trade-offs between coherence, diversity, fluency, and other relevant indicators of text quality. In this work, we adapt a recently proposed framework for statistical inference based on Generalized Stochastic Dominance (GSD) that addresses three critical limitations in existing benchmarking methodologies: the inadequacy of single-metric evaluation, the incompatibility between cardinal automatic metrics and ordinal human judgments, and the lack of inferential statistical guarantees. The GSD-front approach enables simultaneous evaluation across multiple quality dimensions while respecting their different measurement scales, building upon partial orders of decoding strategies, thus avoiding arbitrary weighting of the involved metrics. By applying this framework to evaluate common decoding strategies against human-generated text, we demonstrate its ability to identify statistically significant performance differences while accounting for potential deviations from the i.i.d. assumption of the sampling design. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 16:08:44 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 07:59:45 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Arias",
"Esteban Garces",
""
],
[
"Blocher",
"Hannah",
""
],
[
"Rodemann",
"Julian",
""
],
[
"Aßenmacher",
"Matthias",
""
],
[
"Jansen",
"Christoph",
""
]
] |
2506.18109 | Yuanyuan Zhao | Zhihao Niu and Yuanyuan Zhao | Soft-Clamped Perimeter Modes of Polygon Resonators | 6 pages, 4 figures | null | null | null | physics.optics physics.app-ph quant-ph | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Polygon resonators are known for their small size and high force sensitivity. We propose an analytical model to predict the mechanical quality factors (Q) of polygon perimeter modes by incorporating a stress term in the Timoshenko-Gere equation. We discuss the origin of losses and their relationship with polygon structure dimensions. Notably, we demonstrate that the torsion loss can be suppressed through the torsion angle of supporting tethers. These results offer new insights into optimizing high Q polygon resonators for cavity optomechanics, force sensing, and multimode physics of phononic structures. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 17:44:35 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Niu",
"Zhihao",
""
],
[
"Zhao",
"Yuanyuan",
""
]
] |
2506.18122 | Georgios Nikolaidis | Carlo Bellavita, \'Alvaro Miguel Moreno, Georgios Nikolaidis, Jos\'e \'Angel Pel\'aez | Fractional Volterra-type operator induced by radial weight acting on Hardy space | null | null | null | null | math.CV math.CA math.FA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given a radial doubling weight $\mu$ on the unit disc $\mathbb{D}$ of the complex plane and its odd moments $\mu_{2n+1}=\int_0^1 s^{2n+1}\mu(s)\, ds$, we consider the fractional derivative
$$
D^\mu(f)(z)=\sum_{n=0}^{\infty} \frac{\widehat{f}(n)}{\mu_{2n+1}}z^n,
$$
of a function $ f(z)=\sum_{n=0}^{\infty}\widehat{f}(n)z^n$ analytic in $\mathbb{D}$. We also consider the fractional integral operator $I^\mu(f)(z)=\sum_{n=0}^{\infty} \mu_{2n+1}\widehat{f}(n)z^n$, and the fractional Volterra-type operator
$$
V_{\mu,g}(f)(z)= I^\mu(f\cdot D^\mu(g))(z),\quad f\in\mathcal{H}(\mathbb{D}),
$$
for any fixed $g\in\mathcal{H}(\mathbb{D})$. We prove that $V_{\mu,g}$ is bounded (compact) on a Hardy space $H^p$, $0<p<\infty$, if and only if $g$ belongs to $\text{BMOA}$ ($\text{VMOA}$). Moreover, if $\int_0^1 \frac{\left(\int_r^1 \mu(s)\, ds\right)^p}{(1-r)^2}\,dr=+\infty$, we prove that $V_{\mu,g}$ belongs to the Schatten class $S_p(H^2)$ if and only if $g=0$. On the other hand, if $\frac{\left(\int_r^1 \mu(s)\, ds\right)^p}{(1-r)^2}$ is a radial doubling weight it is proved that $V_{\mu,g} \in S_p(H^2)$ if and only if $g$ belongs to the Besov space $B_p$. En route, we obtain descriptions of $H^p$, $\text{BMOA}$, $\text{VMOA}$ and $B_p$ in terms of the fractional derivative $D^\mu$. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 18:09:43 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 09:10:04 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Bellavita",
"Carlo",
""
],
[
"Moreno",
"Álvaro Miguel",
""
],
[
"Nikolaidis",
"Georgios",
""
],
[
"Peláez",
"José Ángel",
""
]
] |
2506.18154 | Zhanshan (Sam) Ma | Zhanshan Sam Ma and R. A. J. Taylor | Six Decades Post-Discovery of Taylor's Power Law: From Ecological and Statistical Universality, Through Prime Number Distributions and Tipping-Point Signals, to Heterogeneity and Stability of Complex Networks | null | null | null | null | q-bio.OT cs.CE cs.IT cs.SI math.IT q-bio.PE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | First discovered by L. R. Taylor (1961, Nature), Taylor's Power Law (TPL) correlates the mean (M) population abundances and the corresponding variances (V) across a set of insect populations using a power function (V=aM^b). TPL has demonstrated its 'universality' across numerous fields of sciences, social sciences, and humanities. This universality has inspired two main prongs of exploration: one from mathematicians and statisticians, who might instinctively respond with a convergence theorem similar to the central limit theorem of the Gaussian distribution, and another from biologists, ecologists, physicists, etc., who are more interested in potential underlying ecological or organizational mechanisms. Over the past six decades, TPL studies have produced a punctuated landscape with three relatively distinct periods (1960s-1980s; 1990s-2000s, and 2010s-2020s) across the two prongs of abstract and physical worlds. Eight themes have been identified and reviewed on this landscape, including population spatial aggregation and ecological mechanisms, TPL and skewed statistical distributions, mathematical/statistical mechanisms of TPL, sample vs. population TPL, population stability, synchrony, and early warning signals for tipping points, TPL on complex networks, TPL in macrobiomes, and in microbiomes. Three future research directions including fostering reciprocal interactions between the two prongs, heterogeneity measuring, and exploration in the context of evolution. The significance of TPL research includes practically, population fluctuations captured by TPL are relevant for agriculture, forestry, fishery, wildlife-conservation, epidemiology, tumor heterogeneity, earthquakes, social inequality, stock illiquidity, financial stability, tipping point events, etc.; theoretically, TPL is one form of power laws, which are related to phase transitions, universality, scale-invariance, etc. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 19:47:16 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Ma",
"Zhanshan Sam",
""
],
[
"Taylor",
"R. A. J.",
""
]
] |
2506.18167 | Constantin Venhoff | Constantin Venhoff, Iv\'an Arcuschin, Philip Torr, Arthur Conmy, Neel Nanda | Understanding Reasoning in Thinking Language Models via Steering Vectors | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in large language models (LLMs) have led to the development of thinking language models that generate extensive internal reasoning chains before producing responses. While these models achieve improved performance, controlling their reasoning processes remains challenging. This work presents a steering approach for thinking LLMs by analyzing and manipulating specific reasoning behaviors in DeepSeek-R1-Distill models. Through a systematic experiment on 500 tasks across 10 diverse categories, we identify several reasoning behaviors exhibited by thinking models, including expressing uncertainty, generating examples for hypothesis validation, and backtracking in reasoning chains. We demonstrate that these behaviors are mediated by linear directions in the model's activation space and can be controlled using steering vectors. By extracting and applying these vectors, we provide a method to modulate specific aspects of the model's reasoning process, such as its tendency to backtrack or express uncertainty. Our approach offers practical tools for steering reasoning processes in thinking models in a controlled and interpretable manner. We validate our steering method using three DeepSeek-R1-Distill models, demonstrating consistent control across different model architectures. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 20:45:26 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 01:53:33 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Venhoff",
"Constantin",
""
],
[
"Arcuschin",
"Iván",
""
],
[
"Torr",
"Philip",
""
],
[
"Conmy",
"Arthur",
""
],
[
"Nanda",
"Neel",
""
]
] |
2506.18172 | Irsyad Adam | Irsyad Adam, Tengyue Zhang, Shrayes Raman, Zhuyu Qiu, Brandon Taraku, Hexiang Feng, Sile Wang, Ashwath Radhachandran, Shreeram Athreya, Vedrana Ivezic, Peipei Ping, Corey Arnold, and William Speier | STACT-Time: Spatio-Temporal Cross Attention for Cine Thyroid Ultrasound Time Series Classification | null | null | null | null | eess.IV cs.AI cs.CV | http://creativecommons.org/licenses/by-sa/4.0/ | Thyroid cancer is among the most common cancers in the United States. Thyroid nodules are frequently detected through ultrasound (US) imaging, and some require further evaluation via fine-needle aspiration (FNA) biopsy. Despite its effectiveness, FNA often leads to unnecessary biopsies of benign nodules, causing patient discomfort and anxiety. To address this, the American College of Radiology Thyroid Imaging Reporting and Data System (TI-RADS) has been developed to reduce benign biopsies. However, such systems are limited by interobserver variability. Recent deep learning approaches have sought to improve risk stratification, but they often fail to utilize the rich temporal and spatial context provided by US cine clips, which contain dynamic global information and surrounding structural changes across various views. In this work, we propose the Spatio-Temporal Cross Attention for Cine Thyroid Ultrasound Time Series Classification (STACT-Time) model, a novel representation learning framework that integrates imaging features from US cine clips with features from segmentation masks automatically generated by a pretrained model. By leveraging self-attention and cross-attention mechanisms, our model captures the rich temporal and spatial context of US cine clips while enhancing feature representation through segmentation-guided learning. Our model improves malignancy prediction compared to state-of-the-art models, achieving a cross-validation precision of 0.91 (plus or minus 0.02) and an F1 score of 0.89 (plus or minus 0.02). By reducing unnecessary biopsies of benign nodules while maintaining high sensitivity for malignancy detection, our model has the potential to enhance clinical decision-making and improve patient outcomes. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 21:14:04 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Adam",
"Irsyad",
""
],
[
"Zhang",
"Tengyue",
""
],
[
"Raman",
"Shrayes",
""
],
[
"Qiu",
"Zhuyu",
""
],
[
"Taraku",
"Brandon",
""
],
[
"Feng",
"Hexiang",
""
],
[
"Wang",
"Sile",
""
],
[
"Radhachandran",
"Ashwath",
""
],
[
"Athreya",
"Shreeram",
""
],
[
"Ivezic",
"Vedrana",
""
],
[
"Ping",
"Peipei",
""
],
[
"Arnold",
"Corey",
""
],
[
"Speier",
"William",
""
]
] |
2506.18194 | Gabriel Vogel | Francesco Piccoli, Gabriel Vogel, Jana M. Weber | Joint Embedding Predictive Architecture for self-supervised pretraining on polymer molecular graphs | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Recent advances in machine learning (ML) have shown promise in accelerating the discovery of polymers with desired properties by aiding in tasks such as virtual screening via property prediction. However, progress in polymer ML is hampered by the scarcity of high-quality labeled datasets, which are necessary for training supervised ML models. In this work, we study the use of the very recent 'Joint Embedding Predictive Architecture' (JEPA), a type of architecture for self-supervised learning (SSL), on polymer molecular graphs to understand whether pretraining with the proposed SSL strategy improves downstream performance when labeled data is scarce. Our results indicate that JEPA-based self-supervised pretraining on polymer graphs enhances downstream performance, particularly when labeled data is very scarce, achieving improvements across all tested datasets. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 22:51:53 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Piccoli",
"Francesco",
""
],
[
"Vogel",
"Gabriel",
""
],
[
"Weber",
"Jana M.",
""
]
] |
2506.18204 | Youjie Zhou | Youjie Zhou and Guofeng Mei and Yiming Wang and Yi Wan and Fabio Poiesi | Multimodal Fusion SLAM with Fourier Attention | Accepted in IEEE RAL | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Visual SLAM is particularly challenging in environments affected by noise, varying lighting conditions, and darkness. Learning-based optical flow algorithms can leverage multiple modalities to address these challenges, but traditional optical flow-based visual SLAM approaches often require significant computational resources.To overcome this limitation, we propose FMF-SLAM, an efficient multimodal fusion SLAM method that utilizes fast Fourier transform (FFT) to enhance the algorithm efficiency. Specifically, we introduce a novel Fourier-based self-attention and cross-attention mechanism to extract features from RGB and depth signals. We further enhance the interaction of multimodal features by incorporating multi-scale knowledge distillation across modalities. We also demonstrate the practical feasibility of FMF-SLAM in real-world scenarios with real time performance by integrating it with a security robot by fusing with a global positioning module GNSS-RTK and global Bundle Adjustment. Our approach is validated using video sequences from TUM, TartanAir, and our real-world datasets, showcasing state-of-the-art performance under noisy, varying lighting, and dark conditions.Our code and datasets are available at https://github.com/youjie-zhou/FMF-SLAM.git. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 23:44:07 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 09:24:14 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Zhou",
"Youjie",
""
],
[
"Mei",
"Guofeng",
""
],
[
"Wang",
"Yiming",
""
],
[
"Wan",
"Yi",
""
],
[
"Poiesi",
"Fabio",
""
]
] |
2506.18236 | Nobuki Takeda | Nobuki Takeda | Differential operators on Hermitian modular forms on $\mathrm{U}(n, n)$ | null | null | null | null | math.NT | http://creativecommons.org/licenses/by/4.0/ | We construct explicit differential operators on hermitian modular forms, extending methods developed for Siegel modular forms. These differential operators are closely related to the two-variable spherical pluriharmonic polynomials. We construct explicit bases for the space of such polynomials and use them to build concrete operators. As an application, we derive an exact pullback formula for hermitian Eisenstein series. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 02:02:51 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 12:46:27 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Takeda",
"Nobuki",
""
]
] |
2506.18269 | Min Yin | Min Yin and Haoyu Liu and Boyi Lian and Chunlei Chai | Co-persona: Leveraging LLMs and Expert Collaboration to Understand User Personas through Social Media Data Analysis | 17pages,5figures,8tables | null | null | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | This study introduces Co-Persona, a methodological framework bridging large-scale social media analysis with authentic user understanding through systematic integration of Large Language Models and expert validation. Through a case study of B.Co, a Chinese manufacturer, we investigated Co-Persona application in bedside lamp development. Our methodology analyzed over 38 million posts from Xiao Hongshu, employing multi-stage data processing combining advanced NLP with expert validation. Analysis revealed five user personas derived from bedtime behaviors: Health Aficionados, Night Owls, Interior Decorators, Child-care Workers, and Workaholics-each showing unique pre-sleep activities and product preferences. Findings demonstrate Co-Persona enhances manufacturers' ability to process large datasets while maintaining user understanding. The methodology provides structured approaches for targeted marketing and product strategies. Research contributes to theoretical understanding of data-driven persona development and practical applications in consumer-driven innovation. Code and data available at https://github.com/INFPa/LLMwithPersona. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 03:54:30 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 09:12:31 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Yin",
"Min",
""
],
[
"Liu",
"Haoyu",
""
],
[
"Lian",
"Boyi",
""
],
[
"Chai",
"Chunlei",
""
]
] |
2506.18270 | Qiegen Liu | Qinrong Cai, Yu Guan, Zhibo Chen, Dong Liang, Qiuyun Fan, Qiegen Liu | Adaptive Mask-guided K-space Diffusion for Accelerated MRI Reconstruction | 10 pages, 9 figures | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As the deep learning revolution marches on, masked modeling has emerged as a distinctive approach that involves predicting parts of the original data that are proportionally masked during training, and has demonstrated exceptional performance in multiple fields. Magnetic Resonance Imaging (MRI) reconstruction is a critical task in medical imaging that seeks to recover high-quality images from under-sampled k-space data. However, previous MRI reconstruction strategies usually optimized the entire image domain or k-space, without considering the importance of different frequency regions in the k-space This work introduces a diffusion model based on adaptive masks (AMDM), which utilizes the adaptive adjustment of frequency distribution based on k-space data to develop a hybrid masks mechanism that adapts to different k-space inputs. This enables the effective separation of high-frequency and low-frequency components, producing diverse frequency-specific representations. Additionally, the k-space frequency distribution informs the generation of adaptive masks, which, in turn, guide a closed-loop diffusion process. Experimental results verified the ability of this method to learn specific frequency information and thereby improved the quality of MRI reconstruction, providing a flexible framework for optimizing k-space data using masks in the future. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 03:54:53 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Cai",
"Qinrong",
""
],
[
"Guan",
"Yu",
""
],
[
"Chen",
"Zhibo",
""
],
[
"Liang",
"Dong",
""
],
[
"Fan",
"Qiuyun",
""
],
[
"Liu",
"Qiegen",
""
]
] |
2506.18298 | Jiawei Wang | Jia-wei Wang, Xiang-Fa Zhou, Guang-Can Guo and Zheng-Wei Zhou | Thermalization of Quantum Many-Body Scars in Kinetically Constrained Systems | main manuscript with 7 pages containing 4 figures, complementary material with 8 pages containing 3 figures | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | The phenomenon of quantum many-body scars (QMBS) has been studied both theoretically and experimentally, due to its unusual violation of the eigenstate thermalization hypothesis (ETH). In this paper, we extend the ETH to a new description based on the grand canonical ensemble to depict the thermal properties of QMBS models. For this purpose, we embed the dynamics of kinetically constrained systems within the Lindblad-like master equation, and demonstrate that the violation of the ETH by scar eigenstates is related to their slow decay in the corresponding dissipative process. Within this open system description, we reformulate the ETH to demonstrate that both scar eigenstates and thermal ones exhibit thermalization governed by grand canonical statistics. Consequently, our revised ETH unifies scars and thermal states under a cohesive thermodynamic rule. Our work resolves the fundamental tension between constraint-induced non-ergodicity and thermalization paradigms, establishing a unified route to generalized thermalization for quantum many-body systems. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 05:32:47 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 13:22:48 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Wang",
"Jia-wei",
""
],
[
"Zhou",
"Xiang-Fa",
""
],
[
"Guo",
"Guang-Can",
""
],
[
"Zhou",
"Zheng-Wei",
""
]
] |
2506.18303 | Giovanni Scala | Giovanni Scala, Anindita Bera, Gniewomir Sarbicki | Entanglement detection via third-order local invariants from randomized measurements | null | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We compute all third-order local invariants accessible via randomised measurements and employ them to derive separability criteria. The reconstruction of the invariants yields experimentally accessible entanglement criteria for multipartite states with arbitrary local dimensions. The results show that third-order invariants capture inter-subsystem correlations beyond second-order spectral criteria within more feasible entanglement detection protocols than full tomography. As an example, Werner states in $d=3$ the entanglement is detected for $p>\frac 12$ at the second-order correlations, and it is improved to $p>\frac 1{\sqrt[3]{10}}$ at the third-order. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 05:41:07 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 09:34:29 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Scala",
"Giovanni",
""
],
[
"Bera",
"Anindita",
""
],
[
"Sarbicki",
"Gniewomir",
""
]
] |
2506.18310 | Jinze He | Jinze He, Junzhe Qiang, Yiying Dong, Jingyi Wang, Tian Dong, Gongcheng Yue, Rongjin Zhuang, Mingze Lv, Siyuan Yu, Zhongjin Lin, Xinlun Cai, Yuanmu Yang, Guanhao Wu and Yang Li | Programmable electro-optic frequency comb empowers integrated parallel convolution processing | null | null | null | null | physics.optics physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Integrated photonic convolution processors make optical neural networks (ONNs) a transformative solution for artificial intelligence applications such as machine vision. To enhance the parallelism, throughput, and energy efficiency of ONNs, wavelength multiplexing is widely applied. However, it often encounters the challenges of low compactness, limited scalability, and high weight reconstruction latency. Here, we proposed and demonstrated an integrated photonic processing unit with a parallel convolution computing speed of 1.62 trillion operations per second (TOPS) and a weight reconstruction speed exceeding 38 GHz. This processing unit simultaneously achieves, for the first time, multi-wavelength generation and weight mapping via a single programmable electro-optic (EO) frequency comb, featuring unprecedented compactness, device-footprint independent scalability, and near-unity optical power conversion efficiency (conversion efficiency from input optical power to output weighted comb lines). To demonstrate the reconfigurability and functionality of this processing unit, we implemented image edge detection and object classification based on EO combs obtained using the particle swarm algorithm and an EO comb neural network training framework, respectively. Our programmable EO comb-based processing framework establishes a new paradigm towards the development of low-latency monolithic photonic processors, promising real-time in-sensor learning for autonomous vehicles, intelligent robotics, and drones. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 05:54:47 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"He",
"Jinze",
""
],
[
"Qiang",
"Junzhe",
""
],
[
"Dong",
"Yiying",
""
],
[
"Wang",
"Jingyi",
""
],
[
"Dong",
"Tian",
""
],
[
"Yue",
"Gongcheng",
""
],
[
"Zhuang",
"Rongjin",
""
],
[
"Lv",
"Mingze",
""
],
[
"Yu",
"Siyuan",
""
],
[
"Lin",
"Zhongjin",
""
],
[
"Cai",
"Xinlun",
""
],
[
"Yang",
"Yuanmu",
""
],
[
"Wu",
"Guanhao",
""
],
[
"Li",
"Yang",
""
]
] |
2506.18323 | Nisar Ahmed | Muhammad Azeem Aslam, Hassan Khalid, and Nisar Ahmed | A Multi-Scale Spatial Attention-Based Zero-Shot Learning Framework for Low-Light Image Enhancement | null | null | null | null | eess.IV cs.AI cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Low-light image enhancement remains a challenging task, particularly in the absence of paired training data. In this study, we present LucentVisionNet, a novel zero-shot learning framework that addresses the limitations of traditional and deep learning-based enhancement methods. The proposed approach integrates multi-scale spatial attention with a deep curve estimation network, enabling fine-grained enhancement while preserving semantic and perceptual fidelity. To further improve generalization, we adopt a recurrent enhancement strategy and optimize the model using a composite loss function comprising six tailored components, including a novel no-reference image quality loss inspired by human visual perception. Extensive experiments on both paired and unpaired benchmark datasets demonstrate that LucentVisionNet consistently outperforms state-of-the-art supervised, unsupervised, and zero-shot methods across multiple full-reference and no-reference image quality metrics. Our framework achieves high visual quality, structural consistency, and computational efficiency, making it well-suited for deployment in real-world applications such as mobile photography, surveillance, and autonomous navigation. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 06:11:55 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Aslam",
"Muhammad Azeem",
""
],
[
"Khalid",
"Hassan",
""
],
[
"Ahmed",
"Nisar",
""
]
] |
2506.18335 | Saad Wazir | Saad Wazir and Daeyoung Kim | Rethinking Decoder Design: Improving Biomarker Segmentation Using Depth-to-Space Restoration and Residual Linear Attention | Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR), 2025, pp. 30861-30871 | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by/4.0/ | Segmenting biomarkers in medical images is crucial for various biotech applications. Despite advances, Transformer and CNN based methods often struggle with variations in staining and morphology, limiting feature extraction. In medical image segmentation, where datasets often have limited sample availability, recent state-of-the-art (SOTA) methods achieve higher accuracy by leveraging pre-trained encoders, whereas end-to-end methods tend to underperform. This is due to challenges in effectively transferring rich multiscale features from encoders to decoders, as well as limitations in decoder efficiency. To address these issues, we propose an architecture that captures multi-scale local and global contextual information and a novel decoder design, which effectively integrates features from the encoder, emphasizes important channels and regions, and reconstructs spatial dimensions to enhance segmentation accuracy. Our method, compatible with various encoders, outperforms SOTA methods, as demonstrated by experiments on four datasets and ablation studies. Specifically, our method achieves absolute performance gains of 2.76% on MoNuSeg, 3.12% on DSB, 2.87% on Electron Microscopy, and 4.03% on TNBC datasets compared to existing SOTA methods. Code: https://github.com/saadwazir/MCADS-Decoder | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 06:32:36 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Wazir",
"Saad",
""
],
[
"Kim",
"Daeyoung",
""
]
] |
2506.18362 | Hiroshi Kakinuma | Hiroshi Kakinuma, Saya Ajito, Koki Okumura, Makoto Akahoshi, Yu Takabatake, Tomohiko Omura, Motomichi Koyama, and Eiji Akiyama | In situ 2D visualization of hydrogen entry into Zn-coated steels in NaCl solutions: Roles of Zn dissolution and potential distribution | 74 pages, 24 figures | null | null | null | physics.chem-ph cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | The hydrogen entry behavior of a partially Zn-coated steel sheet in NaCl solutions was investigated employing a polyaniline-based hydrogenochromic sensor, electrochemical hydrogen permeation tests, and potential measurements using a scanning Kelvin probe. While the Zn coating mitigated corrosion of the steel substrate, it simultaneously accelerated the hydrogen entry. The hydrogen entry occurred at the bare steel surface regions exposed to the NaCl solution, with the hydrogen flux exhibiting non-uniform distribution: higher near the dissolving Zn coating. While no significant differences in Zn dissolution behavior or galvanic current were observed between 0.1 and 0.01 M NaCl solutions, the total hydrogen flux decreased with decreasing Cl ion concentration. This reduction was attributed to a potential gradient induced by differences in electrolyte conductivity. The results demonstrate that potential distribution, rather than galvanic current, is a dominant factor influencing hydrogen entry under the investigated conditions. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 07:46:26 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Kakinuma",
"Hiroshi",
""
],
[
"Ajito",
"Saya",
""
],
[
"Okumura",
"Koki",
""
],
[
"Akahoshi",
"Makoto",
""
],
[
"Takabatake",
"Yu",
""
],
[
"Omura",
"Tomohiko",
""
],
[
"Koyama",
"Motomichi",
""
],
[
"Akiyama",
"Eiji",
""
]
] |
2506.18370 | V\'ictor J. Maci\'a | V\'ictor J. Maci\'a | Galton-Watson processes, simple varieties of trees and Khinchin families | null | null | null | null | math.CV math.CO math.PR | http://creativecommons.org/licenses/by/4.0/ | In this note, we introduce a unified analytic framework that connects simple varieties of trees, Bienayme-Galton-Watson processes and Khinchin families. Using Lagrange's inversion formula, we derive new coefficient-based expressions for extinction probabilities and reinterpret them as boundary phenomena tied to the domain of the inverse of the solution to Lagrange's equation. This perspective reveals a structural bridge between combinatorial and probabilistic models, simplifying classical arguments and yielding new results. It also leads to a computationally efficient method for simulating Galton-Watson processes via power series coefficients. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 07:56:40 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 06:53:00 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Maciá",
"Víctor J.",
""
]
] |
2506.18405 | Arvind Rameshwar | V. Arvind Rameshwar and Anshoo Tandon | $(\ell,\delta)$-Diversity: Linkage-Robustness via a Composition Theorem | 10 pages, 2 tables, 2 figures | null | null | null | cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | In this paper, we consider the problem of degradation of anonymity upon linkages of anonymized datasets. We work in the setting where an adversary links together $t\geq 2$ anonymized datasets in which a user of interest participates, based on the user's known quasi-identifiers, which motivates the use of $\ell$-diversity as the notion of dataset anonymity. We first argue that in the worst case, such linkage attacks can reveal the exact sensitive attribute of the user, even when each dataset respects $\ell$-diversity, for moderately large values of $\ell$. This issue motivates our definition of (approximate) $(\ell,\delta)$-diversity -- a parallel of (approximate) $(\epsilon,\delta)$-differential privacy (DP) -- which simply requires that a dataset respect $\ell$-diversity, with high probability. We then present a mechanism for achieving $(\ell,\delta)$-diversity, in the setting of independent and identically distributed samples. Next, we establish bounds on the degradation of $(\ell,\delta)$-diversity, via a simple ``composition theorem,'' similar in spirit to those in the DP literature, thereby showing that approximate diversity, unlike standard diversity, is roughly preserved upon linkage. Finally, we describe simple algorithms for maximizing utility, measured in terms of the number of anonymized ``equivalence classes,'' and derive explicit lower bounds on the utility, for special sample distributions. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 08:42:14 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 05:27:22 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Rameshwar",
"V. Arvind",
""
],
[
"Tandon",
"Anshoo",
""
]
] |
2506.18412 | Yang Zhang | Jiangxu Li, Jian Tang, Louis Primeau, Thomas Siyuan Ding, Rahul Soni, Tiema Qian, Kenji Watanabe, Takashi Taniguchi, Ni Ni, Adrian Del Maestro, Qiong Ma, and Yang Zhang | Interaction-Driven Topological Transitions in Monolayer TaIrTe$_4$ | 12+18 pages, 8+14 figures | null | null | null | cond-mat.str-el cond-mat.mes-hall | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Discovering materials that combine topological phenomena with correlated electron behavior is a central pursuit in quantum materials research. Monolayer TaIrTe$_4$ has recently emerged as a promising platform in this context, hosting robust quantum spin Hall insulator (QSHI) phases both within a single-particle gap and within a correlation-induced gap arising from van Hove singularities (vHSs), accessed via electrostatic doping. Its intrinsic monolayer nature offers exceptional tunability and the potential to realize a versatile array of interaction-driven topological phases. In this work, we combine theory and experiment to map the phase landscape of monolayer TaIrTe$_4$. Using Hartree-Fock calculations, we investigate the interaction-driven phase diagram near the vHSs under commensurate filling conditions. By systematically tuning the dielectric screening and strain, we uncover a rich set of ground states--including QSHI, trivial insulator, higher-order topological insulator, and metallic phase--among which are interaction-driven topological phase transitions. Experimentally, we perform both local and nonlocal transport measurements across a broad set of devices, which--due to unavoidable strain variations during fabrication-realize several phases consistent with theoretical predictions. Together, our results lay the groundwork for understanding correlation-driven topological phenomena in TaIrTe$_4$ and open new directions for engineering exotic quantum phases in low-dimensional materials beyond the limitations of moir\'e superlattices. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 08:47:30 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Li",
"Jiangxu",
""
],
[
"Tang",
"Jian",
""
],
[
"Primeau",
"Louis",
""
],
[
"Ding",
"Thomas Siyuan",
""
],
[
"Soni",
"Rahul",
""
],
[
"Qian",
"Tiema",
""
],
[
"Watanabe",
"Kenji",
""
],
[
"Taniguchi",
"Takashi",
""
],
[
"Ni",
"Ni",
""
],
[
"Del Maestro",
"Adrian",
""
],
[
"Ma",
"Qiong",
""
],
[
"Zhang",
"Yang",
""
]
] |
2506.18418 | Hui-Ming Wang | Xuejian Wei, and Hui-Ming Wang | Multi-user Downlink with Reconfigurable Intelligent Metasurface Antennas (RIMSA) Array | Accepted by IEEE TVT | null | null | null | cs.IT eess.SP math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Reconfigurable Intelligent Surfaces (RIS) is a transformative technology with great potential in many applications in wireless communications and realizing the Internet of Everything at sixth generation (6G). In this study, we propose a wireless system where the RIS acts as an antenna, which we call Reconfigurable Intelligent Metasurface Antennas (RIMSA). In particular, the base station (BS) equipped with a RIMSA array performs downlink transmissions to multiple users, where each user has a single or multiple RIMSA/RF links, and we aim to solve the sum-rate maximization problem by jointly optimizing the digital processing matrix of the transceivers and the phase responses of RIMSA array at both BS and users. For the multi-user multiple-input single-output (MU-MISO) scenario, we develop an alternating optimization algorithm to slove the problem, where a fractional programming (FP) is used to optimize the digital processing matrix and a product manifold optimization (PMO) is proposed to provide the optimal phase responses of the RIMSA array at both BS and users. For the multi-user multiple-input multiple-output (MU-MIMO) scenario, we equate it to a weighted sum of mean square errors minimization problem, which can be solved by three subproblems iteratively. Both the optimal digital precoder subproblem and the optimal digital combiner subproblem have closed-form solutions, and the subproblem of RIMSA configuration is solved by the PMO algorithm as well. Simulation results demonstrate that the proposed algorithms achieve significant performance gains over conventional algorithms. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 09:00:24 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Wei",
"Xuejian",
""
],
[
"Wang",
"Hui-Ming",
""
]
] |
2506.18425 | Sk Rakeeb | Sk Rakeeb, Animesh Sharma, Sagar Dam, Ameya Parab, Amit Lad, Yash.M. Ved, Amita Das, G. Ravindra Kumar | Spatial, Spectral and Temporal Response of High Intensity Laser Plasma Mirrors- Direct Observation of the Ponderomotive Push | null | null | null | null | physics.plasm-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Plasma-based optics have emerged as a powerful platform for manipulating and amplifying ultra-intense laser pulses. However, the inherently nonlinear and dynamic nature of plasma leads to significant spatial, spectral, and temporal modulations when driven at relativistic intensities. These modifications can dramatically alter the structure of the reflected laser pulses, posing challenges for their use in applications such as vacuum ultraviolet (VUV) and X-ray generation, as well as relativistic particle acceleration. Comprehensive, multidimensional diagnostics are essential to accurately characterize these so-called `plasma mirrors' (PMs). We present a direct, \textit{in situ} measurement of the three-dimensional plasma surface evolution during femtosecond laser irradiation, achieved through simultaneous analysis of the wavefront, spectrum, and temporal profile of the reflected light. Our measurements reveal surface deformations on the order of a few hundred nanometers at relativistic intensities, in agreement with three-dimensional particle-in-cell (3D-PIC) simulations. Additionally, the PM induces substantial modifications to the pulse spectrum and temporal profile, introducing spatio-temporal couplings. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 09:04:19 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 06:42:55 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Rakeeb",
"Sk",
""
],
[
"Sharma",
"Animesh",
""
],
[
"Dam",
"Sagar",
""
],
[
"Parab",
"Ameya",
""
],
[
"Lad",
"Amit",
""
],
[
"Ved",
"Yash. M.",
""
],
[
"Das",
"Amita",
""
],
[
"Kumar",
"G. Ravindra",
""
]
] |
2506.18432 | Zhaohui Yang | Wei Xu, Zhaohui Yang, Derrick Wing Kwan Ng, Robert Schober, H. Vincent Poor, Zhaoyang Zhang, Xiaohu You | A New Pathway to Integrated Learning and Communication (ILAC): Large AI Model and Hyperdimensional Computing for Communication | 27 pages 14 figures | null | null | null | eess.SP | http://creativecommons.org/licenses/by-sa/4.0/ | The rapid evolution of forthcoming sixth-generation (6G) wireless networks necessitates the seamless integration of artificial intelligence (AI) with wireless communications to support emerging intelligent applications that demand both efficient communication and robust learning performance. This dual requirement calls for a unified framework of integrated learning and communication (ILAC), where AI enhances communication through intelligent signal processing and adaptive resource management, while wireless networks support AI model deployment by enabling efficient and reliable data exchanges. However, achieving this integration presents significant challenges in practice. Communication constraints, such as limited bandwidth and fluctuating channels, hinder learning accuracy and convergence. Simultaneously, AI-driven learning dynamics, including model updates and task-driven inference, introduce excessive burdens on communication systems, necessitating flexible, context-aware transmission strategies. Finally, we present a case study on a cost-to-performance optimization problem, where task assignments, model size selection, bandwidth allocation, and transmission power control are jointly optimized, considering computational cost, communication efficiency, and inference accuracy. Leveraging the Dinkelbach and alternating optimization algorithms, we offer a practical and effective solution to achieve an optimal balance between learning performance and communication constraints. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 09:13:54 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 06:26:34 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Xu",
"Wei",
""
],
[
"Yang",
"Zhaohui",
""
],
[
"Ng",
"Derrick Wing Kwan",
""
],
[
"Schober",
"Robert",
""
],
[
"Poor",
"H. Vincent",
""
],
[
"Zhang",
"Zhaoyang",
""
],
[
"You",
"Xiaohu",
""
]
] |
2506.18435 | Alessandro Daniele Genuardi Oquendo | Alessandro Daniele Genuardi Oquendo, Ali Nadir, Tigers Jonuzi, Siddhartha Patra, Nilotpal Kanti Sinha, Rom\'an Or\'us, Sam Mugel | Accelerating Photonic Integrated Circuit Design: Traditional, ML and Quantum Methods | We thank the Centro para el Desarrollo Tecnol\'ogico e Industria (CDTI) for their support under the Project INSPIRE via Misiones PERTE Chip 2023. Grant EXP-00162621 / MIG-20231016 | null | null | null | quant-ph physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Photonic Integrated Circuits (PICs) provide superior speed, bandwidth, and energy efficiency, making them ideal for communication, sensing, and quantum computing applications. Despite their potential, PIC design workflows and integration lag behind those in electronics, calling for groundbreaking advancements. This review outlines the state of PIC design, comparing traditional simulation methods with machine learning approaches that enhance scalability and efficiency. It also explores the promise of quantum algorithms and quantum-inspired methods to address design challenges. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 09:17:22 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Oquendo",
"Alessandro Daniele Genuardi",
""
],
[
"Nadir",
"Ali",
""
],
[
"Jonuzi",
"Tigers",
""
],
[
"Patra",
"Siddhartha",
""
],
[
"Sinha",
"Nilotpal Kanti",
""
],
[
"Orús",
"Román",
""
],
[
"Mugel",
"Sam",
""
]
] |
2506.18445 | Shantanu Desai | Kunj Panchal, Shantanu Desai | Comparison of $R_h=ct$ and $\Lambda$CDM using DESI DR1 measurements | 5 pages, 1 figure. Accepted for publication in EPL | null | 10.1209/0295-5075/ade738 | null | astro-ph.CO | http://creativecommons.org/licenses/by/4.0/ | We use DESI DR1 BAO measurements of the ratio of tranvserse comoving distance to Hubble distance in order to test the compatibility of $R_h=ct$ over flat $\Lambda$CDM. For this purpose, we used Bayesian model selection to evaluate the efficacy of these models given the observed data. When we consider the BAO measurements up to redshift of 1.3, both models are equally favored. However, when we consider the Lyman-$\alpha$ QSO measurement at redshift of 2.33, we find Bayes factors of greater than 100 for flat $\Lambda$CDM over $R_h=ct$ using two different priors for $\Omega_m$, indicating that $\Lambda$CDM is decisively favored over $R_h=ct$. The same is the case when we combine all the measurements. Therefore, the DESI DRI measurements rule out $R_h=ct$ cosmology, albeit this is driven by the Lyman-$\alpha$ QSO measurement at $z=2.33$. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 09:30:19 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Panchal",
"Kunj",
""
],
[
"Desai",
"Shantanu",
""
]
] |
2506.18483 | Utkarsh Patel | Utkarsh Patel, Sabyasachi Ghosh, Prasanta Murmu | Numerically studying Pesticide diffusion in air using Langevin formalism | 10 pages, 5 figures | null | null | null | physics.app-ph | http://creativecommons.org/licenses/by/4.0/ | The use of pesticides for enhancing crop yield and preventing infestations is a widespread agricultural practice. However, in recent years, there has been a growing shift toward traditional chemical-free organic farming. Regulatory frameworks impose specific distance requirements between organic farms and neighboring lands where chemical pesticides are used to minimize cross-contamination. In this work, we numerically analyze the spread of pesticide droplets to adjacent fields under varying weather conditions, providing a systematic analysis that highlights conditions where existing guidelines might require reassessment. We employ the formalism of the Langevin equations to model the diffusion of pesticide particles and their transport due to wind and other environmental factors. Assuming a non-relativistic, classical diffusion framework, we track the dispersion of commonly used pesticides to assess their potential contamination range. We present our key findings, discuss their implications, and, toward the end, outline possible directions for future research. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 10:37:12 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 08:52:00 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Patel",
"Utkarsh",
""
],
[
"Ghosh",
"Sabyasachi",
""
],
[
"Murmu",
"Prasanta",
""
]
] |
2506.18484 | Pascal Kl\"ockner | Pascal Kl\"ockner, Jos\'e Teixeira, Diana Montezuma, Jaime S. Cardoso, Hugo M. Horlings, Sara P. Oliveira | GANs vs. Diffusion Models for virtual staining with the HER2match dataset | null | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Virtual staining is a promising technique that uses deep generative models to recreate histological stains, providing a faster and more cost-effective alternative to traditional tissue chemical staining. Specifically for H&E-HER2 staining transfer, despite a rising trend in publications, the lack of sufficient public datasets has hindered progress in the topic. Additionally, it is currently unclear which model frameworks perform best for this particular task. In this paper, we introduce the HER2match dataset, the first publicly available dataset with the same breast cancer tissue sections stained with both H&E and HER2. Furthermore, we compare the performance of several Generative Adversarial Networks (GANs) and Diffusion Models (DMs), and implement a novel Brownian Bridge Diffusion Model for H&E-HER2 translation. Our findings indicate that, overall, GANs perform better than DMs, with only the BBDM achieving comparable results. Furthermore, we emphasize the importance of data alignment, as all models trained on HER2match produced vastly improved visuals compared to the widely used consecutive-slide BCI dataset. This research provides a new high-quality dataset ([available upon publication acceptance]), improving both model training and evaluation. In addition, our comparison of frameworks offers valuable guidance for researchers working on the topic. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 10:37:41 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Klöckner",
"Pascal",
""
],
[
"Teixeira",
"José",
""
],
[
"Montezuma",
"Diana",
""
],
[
"Cardoso",
"Jaime S.",
""
],
[
"Horlings",
"Hugo M.",
""
],
[
"Oliveira",
"Sara P.",
""
]
] |
2506.18512 | Yuting Zhang | Yuting Zhang, Kaishen Yuan, Hao Lu, Yutao Yue, Jintai Chen, Kaishun Wu | MedTVT-R1: A Multimodal LLM Empowering Medical Reasoning and Diagnosis | null | null | null | null | eess.IV cs.CL cs.CV q-bio.QM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate and interpretable multi-disease diagnosis remains a critical challenge in medical research, particularly when leveraging heterogeneous multimodal medical data. Current approaches often rely on single-modal data, limiting their ability to comprehensively understand complex diseases. To address this, we propose MedTVT-R1, a novel Multimodal Large Language Model (MLLM) framework designed to integrate clinical multimodal data for reasoning and diagnosing multiple diseases. We construct MedTVT-QA, a curated instruction dataset that provides question-answer pairs for physiological-level interpretations and disease-level diagnoses with a Chain of Evidence approach. MedTVT-R1 incorporates a modality perception layer to capture inter-modal dependencies and adaptively weight modality contributions. Additionally, we employ Group Relative Policy Optimization (GRPO)-based Reinforcement Fine-Tuning with a Jaccard Reward function to enhance diagnostic reasoning. Experimental results demonstrate MedTVT-R1's superiority in multimodal feature utilization and multi-disease diagnosis, offering significant potential for clinical applications such as diagnostic report generation and comorbidity reasoning. The dataset and code are available at https://github.com/keke-nice/MedTVT-R1. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 11:06:31 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Zhang",
"Yuting",
""
],
[
"Yuan",
"Kaishen",
""
],
[
"Lu",
"Hao",
""
],
[
"Yue",
"Yutao",
""
],
[
"Chen",
"Jintai",
""
],
[
"Wu",
"Kaishun",
""
]
] |
2506.18532 | Mengjie Qian | Mengjie Qian, Rao Ma, Stefano Bann\`o, Mark J.F. Gales, Kate M. Knill | End-to-End Spoken Grammatical Error Correction | This work has been submitted to the IEEE for possible publication | null | null | null | cs.CL cs.LG cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Grammatical Error Correction (GEC) and feedback play a vital role in supporting second language (L2) learners, educators, and examiners. While written GEC is well-established, spoken GEC (SGEC), aiming to provide feedback based on learners' speech, poses additional challenges due to disfluencies, transcription errors, and the lack of structured input. SGEC systems typically follow a cascaded pipeline consisting of Automatic Speech Recognition (ASR), disfluency detection, and GEC, making them vulnerable to error propagation across modules. This work examines an End-to-End (E2E) framework for SGEC and feedback generation, highlighting challenges and possible solutions when developing these systems. Cascaded, partial-cascaded and E2E architectures are compared, all built on the Whisper foundation model. A challenge for E2E systems is the scarcity of GEC labeled spoken data. To address this, an automatic pseudo-labeling framework is examined, increasing the training data from 77 to over 2500 hours. To improve the accuracy of the SGEC system, additional contextual information, exploiting the ASR output, is investigated. Candidate feedback of their mistakes is an essential step to improving performance. In E2E systems the SGEC output must be compared with an estimate of the fluent transcription to obtain the feedback. To improve the precision of this feedback, a novel reference alignment process is proposed that aims to remove hypothesised edits that results from fluent transcription errors. Finally, these approaches are combined with an edit confidence estimation approach, to exclude low-confidence edits. Experiments on the in-house Linguaskill (LNG) corpora and the publicly available Speak & Improve (S&I) corpus show that the proposed approaches significantly boost E2E SGEC performance. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 11:40:04 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Qian",
"Mengjie",
""
],
[
"Ma",
"Rao",
""
],
[
"Bannò",
"Stefano",
""
],
[
"Gales",
"Mark J. F.",
""
],
[
"Knill",
"Kate M.",
""
]
] |
2506.18557 | Dongjin Kim | Sung Jin Um, Dongjin Kim, Sangmin Lee, Jung Uk Kim | Object-aware Sound Source Localization via Audio-Visual Scene Understanding | Accepted at CVPR 2025 | Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR), 2025, pp. 8342-8351 | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Audio-visual sound source localization task aims to spatially localize sound-making objects within visual scenes by integrating visual and audio cues. However, existing methods struggle with accurately localizing sound-making objects in complex scenes, particularly when visually similar silent objects coexist. This limitation arises primarily from their reliance on simple audio-visual correspondence, which does not capture fine-grained semantic differences between sound-making and silent objects. To address these challenges, we propose a novel sound source localization framework leveraging Multimodal Large Language Models (MLLMs) to generate detailed contextual information that explicitly distinguishes between sound-making foreground objects and silent background objects. To effectively integrate this detailed information, we introduce two novel loss functions: Object-aware Contrastive Alignment (OCA) loss and Object Region Isolation (ORI) loss. Extensive experimental results on MUSIC and VGGSound datasets demonstrate the effectiveness of our approach, significantly outperforming existing methods in both single-source and multi-source localization scenarios. Code and generated detailed contextual information are available at: https://github.com/VisualAIKHU/OA-SSL. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 12:08:07 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 02:05:51 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Um",
"Sung Jin",
""
],
[
"Kim",
"Dongjin",
""
],
[
"Lee",
"Sangmin",
""
],
[
"Kim",
"Jung Uk",
""
]
] |
2506.18560 | Jiexin Zhang | Jiexin Zhang, Shu Xu, Chunguo Li, Yongming Huang, and Luxi Yang | Efficient Beam Selection for ISAC in Cell-Free Massive MIMO via Digital Twin-Assisted Deep Reinforcement Learning | Submitted to IEEE Transactions on Wireless Communications | null | null | null | cs.ET cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Beamforming enhances signal strength and quality by focusing energy in specific directions. This capability is particularly crucial in cell-free integrated sensing and communication (ISAC) systems, where multiple distributed access points (APs) collaborate to provide both communication and sensing services. In this work, we first derive the distribution of joint target detection probabilities across multiple receiving APs under false alarm rate constraints, and then formulate the beam selection procedure as a Markov decision process (MDP). We establish a deep reinforcement learning (DRL) framework, in which reward shaping and sinusoidal embedding are introduced to facilitate agent learning. To eliminate the high costs and associated risks of real-time agent-environment interactions, we further propose a novel digital twin (DT)-assisted offline DRL approach. Different from traditional online DRL, a conditional generative adversarial network (cGAN)-based DT module, operating as a replica of the real world, is meticulously designed to generate virtual state-action transition pairs and enrich data diversity, enabling offline adjustment of the agent's policy. Additionally, we address the out-of-distribution issue by incorporating an extra penalty term into the loss function design. The convergency of agent-DT interaction and the upper bound of the Q-error function are theoretically derived. Numerical results demonstrate the remarkable performance of our proposed approach, which significantly reduces online interaction overhead while maintaining effective beam selection across diverse conditions including strict false alarm control, low signal-to-noise ratios, and high target velocities. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 12:17:57 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Zhang",
"Jiexin",
""
],
[
"Xu",
"Shu",
""
],
[
"Li",
"Chunguo",
""
],
[
"Huang",
"Yongming",
""
],
[
"Yang",
"Luxi",
""
]
] |
2506.18561 | Claudio Bonati | Matteo Bacci and Claudio Bonati | Stability of universal properties against perturbations of the Markov Chain Monte Carlo algorithm | 11 pages, 16 pdf figures | null | null | null | cond-mat.stat-mech hep-lat hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We numerically investigate the stability of universal properties at continuous phase transitions against perturbations of the Markov Chain Monte Carlo algorithm used to simulate the system. We consider the three dimensional XY model as test bed, and both local (single site Metropolis) and global (single cluster) updates, introducing deterministic truncation-like perturbations and stochastic perturbations in the acceptance probabilities. In (almost) all the cases we find a remarkable stability of the universal properties, even against large perturbations of the Markov Chain Monte Carlo algorithm, with critical exponents and scaling curves consistent with those of the standard XY model within statistical uncertainties. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 12:18:31 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Bacci",
"Matteo",
""
],
[
"Bonati",
"Claudio",
""
]
] |
2506.18570 | Apostolos Argyris | Apostolos Argyris | Temporal pattern generation with tunable repetition rate using semiconductor laser periodic dynamics | null | null | null | null | physics.optics | http://creativecommons.org/licenses/by/4.0/ | In this work, a long-cavity semiconductor laser subject to optical feedback is exploited to generate repetitive temporal patterns with enhanced intra-pattern sample diversity. Stable limit cycle dynamics characterized by multiple frequency harmonics are experimentally demonstrated by considering a 2 m external cavity with appropriate optical feedback conditions. This simple configuration can enable stable and continuous generation of periodic waveforms with a tunable repetition rate ranging from 3.28 to 4.21 GHz. Distinct, highly consistent temporal patterns are obtained by applying high-pass filtering to adjust the relative power of the harmonics. The resulting system operates as a physical waveform generator capable of producing diverse and repeatable signal patterns, making it well-suited for use in the masking stage of time delay reservoir computing architectures. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 12:23:30 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 08:18:43 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Argyris",
"Apostolos",
""
]
] |
2506.18591 | Mauricio Byrd Victorica | Mauricio Byrd Victorica, Gy\"orgy D\'an, Henrik Sandberg | SpaNN: Detecting Multiple Adversarial Patches on CNNs by Spanning Saliency Thresholds | 2025 IEEE Conference on Secure and Trustworthy Machine Learning (SaTML2025) | null | 10.1109/SaTML64287.2025.00032 | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | State-of-the-art convolutional neural network models for object detection and image classification are vulnerable to physically realizable adversarial perturbations, such as patch attacks. Existing defenses have focused, implicitly or explicitly, on single-patch attacks, leaving their sensitivity to the number of patches as an open question or rendering them computationally infeasible or inefficient against attacks consisting of multiple patches in the worst cases. In this work, we propose SpaNN, an attack detector whose computational complexity is independent of the expected number of adversarial patches. The key novelty of the proposed detector is that it builds an ensemble of binarized feature maps by applying a set of saliency thresholds to the neural activations of the first convolutional layer of the victim model. It then performs clustering on the ensemble and uses the cluster features as the input to a classifier for attack detection. Contrary to existing detectors, SpaNN does not rely on a fixed saliency threshold for identifying adversarial regions, which makes it robust against white box adversarial attacks. We evaluate SpaNN on four widely used data sets for object detection and classification, and our results show that SpaNN outperforms state-of-the-art defenses by up to 11 and 27 percentage points in the case of object detection and the case of image classification, respectively. Our code is available at https://github.com/gerkbyrd/SpaNN. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 12:51:10 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Victorica",
"Mauricio Byrd",
""
],
[
"Dán",
"György",
""
],
[
"Sandberg",
"Henrik",
""
]
] |
2506.18592 | Rui Chen | Rui Chen, Andrew Tang, Jayita Dutta, Virat Tara, Julian Ye, Zhuoran Fang, Arka Majumdar | NEO-PGA: Nonvolatile electro-optically programmable gate array | null | null | null | null | physics.optics physics.app-ph | http://creativecommons.org/licenses/by/4.0/ | Programmable photonic integrated circuits (PICs) offer a unique opportunity to create a flexible platform, akin to electronic field programmable gate array (FPGA). These photonic PGAs can implement versatile functionalities for applications ranging from optical interconnects to microwave photonics. However, state-of-the-art programmable photonics relies predominantly on volatile thermo-optic tuning, which suffers from high static power consumption, large footprints, and thermal crosstalk. All these dramatically limit the gate density and pose a fundamental limit to the scalability. Chalcogenide-based phase-change materials (PCMs) offer a superior alternative due to their nonvolatility and substantial optical contrast, though challenges such as optical loss, and bit precision severely limited their application in large-scale PICs. Here, we demonstrate precise, multi-bit, low-loss tuning of the emerging PCM Sb2Se3 using a closed-loop, "program-and-verify" method. Electrically reconfigurable PCM-integrated silicon photonic gates are implemented on a 300mm silicon photonic platform, using circulating and forward Mach-Zehnder interferometer (MZI) meshes. In the circulating mesh, we realize broadband optical switching fabrics and high-Q coupled resonators with unprecedented local control of coupling rates, which further enable exploration of coupled-cavity systems. The forward mesh supports self-configurable MZIs that sort two orthogonal beams to different ports. These results showcase a new type of scalable photonic PGA enabled by PCMs, offering a pathway toward general-purpose, on-chip programmable photonic systems. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 12:52:31 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Chen",
"Rui",
""
],
[
"Tang",
"Andrew",
""
],
[
"Dutta",
"Jayita",
""
],
[
"Tara",
"Virat",
""
],
[
"Ye",
"Julian",
""
],
[
"Fang",
"Zhuoran",
""
],
[
"Majumdar",
"Arka",
""
]
] |
2506.18616 | Etienne Marion | Etienne Marion (ENS de Lyon) | A Formalization of the Ionescu-Tulcea Theorem in Mathlib | null | null | null | null | math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We describe the formalization of the Ionescu-Tulcea theorem, showing the existence of a probability measure on the space of trajectories of a Markov chain, in the proof assistant Lean using the integrated library Mathlib. We first present a mathematical proof before exposing the difficulties which arise when trying to formalize it, and how they were overcome. We then build on this work to formalize the construction of the product of an arbitrary family of probability measures. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 13:24:06 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 08:38:36 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Marion",
"Etienne",
"",
"ENS de Lyon"
]
] |
2506.18631 | Chenxing Wei | Chenxing Wei, Jiarui Yu, Ying Tiffany He, Hande Dong, Yao Shu, Fei Yu | ReDit: Reward Dithering for Improved LLM Policy Optimization | 10 pages, 15 figures | null | null | null | cs.LG cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | DeepSeek-R1 has successfully enhanced Large Language Model (LLM) reasoning capabilities through its rule-based reward system. While it's a ''perfect'' reward system that effectively mitigates reward hacking, such reward functions are often discrete. Our experimental observations suggest that discrete rewards can lead to gradient anomaly, unstable optimization, and slow convergence. To address this issue, we propose ReDit (Reward Dithering), a method that dithers the discrete reward signal by adding simple random noise. With this perturbed reward, exploratory gradients are continuously provided throughout the learning process, enabling smoother gradient updates and accelerating convergence. The injected noise also introduces stochasticity into flat reward regions, encouraging the model to explore novel policies and escape local optima. Experiments across diverse tasks demonstrate the effectiveness and efficiency of ReDit. On average, ReDit achieves performance comparable to vanilla GRPO with only approximately 10% the training steps, and furthermore, still exhibits a 4% performance improvement over vanilla GRPO when trained for a similar duration. Visualizations confirm significant mitigation of gradient issues with ReDit. Moreover, theoretical analyses are provided to further validate these advantages. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 13:36:24 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 07:07:57 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Wei",
"Chenxing",
""
],
[
"Yu",
"Jiarui",
""
],
[
"He",
"Ying Tiffany",
""
],
[
"Dong",
"Hande",
""
],
[
"Shu",
"Yao",
""
],
[
"Yu",
"Fei",
""
]
] |
2506.18633 | Yating Wu | Yating Wu, Haozhe Sun, Bo Xiong, Yalong Yv, Jiale Zhang, Zhaojie Zheng, Wei Ma, and Tao Chu | Lossless, Non-Volatile Post-Fabrication Trimming of PICs via On-Chip High-Temperature Annealing of Undercut Waveguides | null | null | null | null | physics.optics physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Limited by equipment precision, manufacturing deviations in waveguide width, etch depth, and layer thickness inevitably occur in photonic integrated circuits (PICs). These variations cause initial phase errors, compromising the reliability of phase-sensitive devices such as Mach-Zehnder Interferometers (MZI) and microring resonators. To overcome this, we report a nonvolatile, near-lossless post-trimming method utilizing sufficient high-temperature thermal treatment for undercut waveguides, reported here for the first time to the best of our knowledge. This CMOS-compatible approach requires no additional processes or equipment, enables simple electrical heating for trimming, and retains long-term stability after high-temperature removal, ensuring high energy efficiency. Transmission electron microscopy indicates that high-temperature thermal treatment induces irreversible lattice expansion in silicon waveguides, leading to a reduction in the real refractive index and enabling compensation for process errors. Experimental results using MZIs confirm a permanent refractive index reduction of 0.0173 and high-resolution tuning up to 5.25 bits, effective across a broadband spectrum and stable for over 218 days after final trimming. Furthermore, 15 MZIs on a single wafer are precisely calibrated to BAR, CROSS, or orthogonal states, demonstrating the method universality. This practical and scalable technique enables reliable post-fabrication trimming for next-generation low-cost, energy-efficient PIC applications such as optical switches and optical computing. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 13:36:52 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 10:08:27 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Wu",
"Yating",
""
],
[
"Sun",
"Haozhe",
""
],
[
"Xiong",
"Bo",
""
],
[
"Yv",
"Yalong",
""
],
[
"Zhang",
"Jiale",
""
],
[
"Zheng",
"Zhaojie",
""
],
[
"Ma",
"Wei",
""
],
[
"Chu",
"Tao",
""
]
] |
2506.18654 | Jozsef Cserti | J\'ozsef Cserti and Gyula D\'avid | General theory of perturbation of infinite resistor networks | 16 figures, 31 pages | null | null | null | math-ph cond-mat.stat-mech math.MP | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The effective resistance between any two nodes in a perturbed resistor network is determined by removing multiple bonds from an infinite resistor lattice. We have developed an efficient method for calculating the Green operator of the Laplacian for such perturbed networks, which is directly related to the two-point resistance. Unlike the recursive techniques that remove bonds one at a time, our approach handles all bond modifications simultaneously. To demonstrate the versatility of our method, several analytical and numerical examples are presented. In addition, we computed bond current distributions to gain deeper insight into the nature of resistor perturbations. We emphasize that our method has a broad range of applications, including condensed matter physics describing the quantum mechanical effects of impurities in crystal lattices, recently emerging topoelectronics, the study of vibrations in spring networks, and problems involving random walks. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 13:55:24 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Cserti",
"József",
""
],
[
"Dávid",
"Gyula",
""
]
] |
2506.18660 | Xinyi Lin | Xinyi Lin, Peizheng Li, Adnan Aijaz | RL-Driven Semantic Compression Model Selection and Resource Allocation in Semantic Communication Systems | Accepted by PIMRC 2025 | null | null | null | cs.NI eess.SP | http://creativecommons.org/licenses/by/4.0/ | Semantic communication (SemCom) is an emerging paradigm that leverages semantic-level understanding to improve communication efficiency, particularly in resource-constrained scenarios. However, existing SemCom systems often overlook diverse computational and communication capabilities and requirements among different users. Motivated by the need to adaptively balance semantic accuracy, latency, and energy consumption, this paper presents a reinforcement learning (RL)-driven framework for semantic compression model (SCM) selection and resource allocation in multi-user SemCom systems. To address the challenges of balancing image reconstruction quality and communication performance, a system-level optimization metric called Rate-Distortion Efficiency (RDE) has been defined. The framework considers multiple SCMs with varying complexity and resource requirements. A proximal policy optimization (PPO)-based RL approach is developed to dynamically select SCMs and allocate bandwidth and power under non-convex constraints. Simulations demonstrate that the proposed method outperforms several baseline strategies. This paper also discusses the generalization ability, computational complexity, scalability, and practical implications of the framework for real-world SemCom systems. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 14:02:01 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Lin",
"Xinyi",
""
],
[
"Li",
"Peizheng",
""
],
[
"Aijaz",
"Adnan",
""
]
] |
2506.18710 | Robin Ince | Maxime Leli\`evre, Amy Waldock, Meng Liu, Natalia Vald\'es Aspillaga, Alasdair Mackintosh, Mar\'ia Jos\'e Ogando Portela, Jared Lee, Paul Atherton, Robin A. A. Ince, Oliver G. B. Garrod | Benchmarking the Pedagogical Knowledge of Large Language Models | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Benchmarks like Massive Multitask Language Understanding (MMLU) have played a pivotal role in evaluating AI's knowledge and abilities across diverse domains. However, existing benchmarks predominantly focus on content knowledge, leaving a critical gap in assessing models' understanding of pedagogy - the method and practice of teaching. This paper introduces The Pedagogy Benchmark, a novel dataset designed to evaluate large language models on their Cross-Domain Pedagogical Knowledge (CDPK) and Special Education Needs and Disability (SEND) pedagogical knowledge. These benchmarks are built on a carefully curated set of questions sourced from professional development exams for teachers, which cover a range of pedagogical subdomains such as teaching strategies and assessment methods. Here we outline the methodology and development of these benchmarks. We report results for 97 models, with accuracies spanning a range from 28% to 89% on the pedagogical knowledge questions. We consider the relationship between cost and accuracy and chart the progression of the Pareto value frontier over time. We provide online leaderboards at https://rebrand.ly/pedagogy which are updated with new models and allow interactive exploration and filtering based on various model properties, such as cost per token and open-vs-closed weights, as well as looking at performance in different subjects. LLMs and generative AI have tremendous potential to influence education and help to address the global learning crisis. Education-focused benchmarks are crucial to measure models' capacities to understand pedagogical concepts, respond appropriately to learners' needs, and support effective teaching practices across diverse contexts. They are needed for informing the responsible and evidence-based deployment of LLMs and LLM-based tools in educational settings, and for guiding both development and policy decisions. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 14:49:01 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 12:36:22 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Lelièvre",
"Maxime",
""
],
[
"Waldock",
"Amy",
""
],
[
"Liu",
"Meng",
""
],
[
"Aspillaga",
"Natalia Valdés",
""
],
[
"Mackintosh",
"Alasdair",
""
],
[
"Portela",
"María José Ogando",
""
],
[
"Lee",
"Jared",
""
],
[
"Atherton",
"Paul",
""
],
[
"Ince",
"Robin A. A.",
""
],
[
"Garrod",
"Oliver G. B.",
""
]
] |
2506.18719 | Jesus Carlos Pimentel-Garcia | Jesus Carlos Pimentel-Garcia , Luis Antonio Jimenez-Ponce | The Full Nonlinear Vortex Tube-Vorton Method: the post-stall condition | 31 pages | null | null | null | physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | The original concepts behind The Full Nonlinear Vortex Tube-Vorton Method (FTVM) have been applied to the study of massively separated fluid flow past a thin body. In the pre-stall condition, the Kutta-Zhukovski (KJ) force calculation was successfully implemented. However, for the post-stall condition, this approach is not capable of capturing the form drag component. Thus, an alternative force calculation method is implemented. The results obtained demonstrate that the fluid motion involving complex wake dynamics can be accurately approximated by a low-order, low-discretization numerical method based purely on vorticity and velocity concepts. This method avoids the need to discuss and speculate about the role of pressure on fluid motion, even in the turbulent regime. Consequently, it simplifies the understanding of complex low-speed aerodynamics by returning to the first principles. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 14:56:11 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Pimentel-Garcia",
"Jesus Carlos",
""
],
[
"Jimenez-Ponce",
"Luis Antonio",
""
]
] |
2506.18729 | Fang-Duo Tsai | Fang-Duo Tsai, Shih-Lun Wu, Weijaw Lee, Sheng-Ping Yang, Bo-Rui Chen, Hao-Chung Cheng, Yi-Hsuan Yang | MuseControlLite: Multifunctional Music Generation with Lightweight Conditioners | Accepted by the 42nd International Conference on Machine Learning (ICML 2025) | null | null | null | cs.SD cs.AI eess.AS | http://creativecommons.org/licenses/by/4.0/ | We propose MuseControlLite, a lightweight mechanism designed to fine-tune text-to-music generation models for precise conditioning using various time-varying musical attributes and reference audio signals. The key finding is that positional embeddings, which have been seldom used by text-to-music generation models in the conditioner for text conditions, are critical when the condition of interest is a function of time. Using melody control as an example, our experiments show that simply adding rotary positional embeddings to the decoupled cross-attention layers increases control accuracy from 56.6% to 61.1%, while requiring 6.75 times fewer trainable parameters than state-of-the-art fine-tuning mechanisms, using the same pre-trained diffusion Transformer model of Stable Audio Open. We evaluate various forms of musical attribute control, audio inpainting, and audio outpainting, demonstrating improved controllability over MusicGen-Large and Stable Audio Open ControlNet at a significantly lower fine-tuning cost, with only 85M trainble parameters. Source code, model checkpoints, and demo examples are available at: https://musecontrollite.github.io/web/. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 15:08:03 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 15:53:56 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Tsai",
"Fang-Duo",
""
],
[
"Wu",
"Shih-Lun",
""
],
[
"Lee",
"Weijaw",
""
],
[
"Yang",
"Sheng-Ping",
""
],
[
"Chen",
"Bo-Rui",
""
],
[
"Cheng",
"Hao-Chung",
""
],
[
"Yang",
"Yi-Hsuan",
""
]
] |
2506.18744 | Ben Letham | Qing Feng, Samuel Dalton, Benjamin Letham, Maximilian Balandat, Eytan Bakshy | Experimenting, Fast and Slow: Bayesian Optimization of Long-term Outcomes with Online Experiments | null | null | 10.1145/3690624.3709419 | null | cs.LG stat.ML | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Online experiments in internet systems, also known as A/B tests, are used for a wide range of system tuning problems, such as optimizing recommender system ranking policies and learning adaptive streaming controllers. Decision-makers generally wish to optimize for long-term treatment effects of the system changes, which often requires running experiments for a long time as short-term measurements can be misleading due to non-stationarity in treatment effects over time. The sequential experimentation strategies--which typically involve several iterations--can be prohibitively long in such cases. We describe a novel approach that combines fast experiments (e.g., biased experiments run only for a few hours or days) and/or offline proxies (e.g., off-policy evaluation) with long-running, slow experiments to perform sequential, Bayesian optimization over large action spaces in a short amount of time. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 15:18:54 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Feng",
"Qing",
""
],
[
"Dalton",
"Samuel",
""
],
[
"Letham",
"Benjamin",
""
],
[
"Balandat",
"Maximilian",
""
],
[
"Bakshy",
"Eytan",
""
]
] |
2506.18772 | Hassan Al Khatib | Hassan S. Al Khatib, Subash Neupane, Sudip Mittal, Shahram Rahimi, Nina Marhamati, Sean Bozorgzad | Patient Journey Ontology: Representing Medical Encounters for Enhanced Patient-Centric Applications | null | null | null | null | cs.DB cs.CY | http://creativecommons.org/licenses/by/4.0/ | The healthcare industry is moving towards a patient-centric paradigm that requires advanced methods for managing and representing patient data. This paper presents a Patient Journey Ontology (PJO), a framework that aims to capture the entirety of a patient's healthcare encounters. Utilizing ontologies, the PJO integrates different patient data sources like medical histories, diagnoses, treatment pathways, and outcomes; it enables semantic interoperability and enhances clinical reasoning. By capturing temporal, sequential, and causal relationships between medical encounters, the PJO supports predictive analytics, enabling earlier interventions and optimized treatment plans. The ontology's structure, including its main classes, subclasses, properties, and relationships, as detailed in the paper, demonstrates its ability to provide a holistic view of patient care. Quantitative and qualitative evaluations by Subject Matter Experts (SMEs) demonstrate strong capabilities in patient history retrieval, symptom tracking, and provider interaction representation, while identifying opportunities for enhanced diagnosis-symptom linking. These evaluations reveal the PJO's reliability and practical applicability, demonstrating its potential to enhance patient outcomes and healthcare efficiency. This work contributes to the ongoing efforts of knowledge representation in healthcare, offering a reliable tool for personalized medicine, patient journey analysis and advancing the capabilities of Generative AI in healthcare applications. | [
{
"version": "v1",
"created": "Tue, 04 Mar 2025 13:05:47 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Khatib",
"Hassan S. Al",
""
],
[
"Neupane",
"Subash",
""
],
[
"Mittal",
"Sudip",
""
],
[
"Rahimi",
"Shahram",
""
],
[
"Marhamati",
"Nina",
""
],
[
"Bozorgzad",
"Sean",
""
]
] |
2506.18807 | Pietro Bonazzi | Pietro Bonazzi, Nicola Farronato, Stefan Zihlmann, Haotong Qin, Michele Magno | PicoSAM2: Low-Latency Segmentation In-Sensor for Edge Vision Applications | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Real-time, on-device segmentation is critical for latency-sensitive and privacy-aware applications like smart glasses and IoT devices. We introduce PicoSAM2, a lightweight (1.3M parameters, 336M MACs) promptable segmentation model optimized for edge and in-sensor execution, including the Sony IMX500. It builds on a depthwise separable U-Net, with knowledge distillation and fixed-point prompt encoding to learn from the Segment Anything Model 2 (SAM2). On COCO and LVIS, it achieves 51.9% and 44.9% mIoU, respectively. The quantized model (1.22MB) runs at 14.3 ms on the IMX500-achieving 86 MACs/cycle, making it the only model meeting both memory and compute constraints for in-sensor deployment. Distillation boosts LVIS performance by +3.5% mIoU and +5.1% mAP. These results demonstrate that efficient, promptable segmentation is feasible directly on-camera, enabling privacy-preserving vision without cloud or host processing. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 16:16:02 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 09:56:22 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Bonazzi",
"Pietro",
""
],
[
"Farronato",
"Nicola",
""
],
[
"Zihlmann",
"Stefan",
""
],
[
"Qin",
"Haotong",
""
],
[
"Magno",
"Michele",
""
]
] |
2506.18808 | Caren Marzban | Caren Marzban, Yikun Zhang, Nicholas Bond, Michael Richman | A Practical Introduction to Regression-based Causal Inference in Meteorology (I): All confounders measured | null | null | null | null | stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Whether a variable is the cause of another, or simply associated with it, is often an important scientific question. Causal Inference is the name associated with the body of techniques for addressing that question in a statistical setting. Although assessing causality is relatively straightforward in the presence of temporal information, outside of that setting - the situation considered here - it is more difficult to assess causal effects. The development of the field of causal inference has involved concepts from a wide range of topics, thereby limiting its adoption across some fields, including meteorology. However, at its core, the requisite knowledge for causal inference involves little more than basic probability theory and regression, topics familiar to most meteorologists. By focusing on these core areas, this and a companion article provide a steppingstone for the meteorology community into the field of (non-temporal) causal inference. Although some theoretical foundations are presented, the main goal is the application of a specific method, called matching, to a problem in meteorology. The data for the application are in public domain, and R code is provided as well, forming an easy path for meteorology students and researchers to enter the field. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 16:19:02 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 14:44:52 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Marzban",
"Caren",
""
],
[
"Zhang",
"Yikun",
""
],
[
"Bond",
"Nicholas",
""
],
[
"Richman",
"Michael",
""
]
] |
2506.18810 | Siao Tang | Siao Tang, Xinyin Ma, Gongfan Fang, Xinchao Wang | ConciseHint: Boosting Efficient Reasoning via Continuous Concise Hints during Generation | Codes are available at https://github.com/tsa18/ConciseHint | null | null | null | cs.AI cs.CL cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent advancements in large reasoning models (LRMs) like DeepSeek-R1 and OpenAI o1 series have achieved notable performance enhancements on complex reasoning tasks by scaling up the generation length by Chain-of-Thought (CoT). However, an emerging issue is their inclination to produce excessively verbose reasoning processes, leading to the inefficiency problem. Existing literature on improving efficiency mainly adheres to the before-reasoning paradigms such as prompting and reasoning or fine-tuning and reasoning, but ignores the promising direction of directly encouraging the model to speak concisely by intervening during the generation of reasoning. In order to fill the blank, we propose a framework dubbed ConciseHint, which continuously encourages the reasoning model to speak concisely by injecting the textual hint (manually designed or trained on the concise data) during the token generation of the reasoning process. Besides, ConciseHint is adaptive to the complexity of the query by adaptively adjusting the hint intensity, which ensures it will not undermine model performance. Experiments on the state-of-the-art LRMs, including DeepSeek-R1 and Qwen-3 series, demonstrate that our method can effectively produce concise reasoning processes while maintaining performance well. For instance, we achieve a reduction ratio of 65\% for the reasoning length on GSM8K benchmark with Qwen-3 4B with nearly no accuracy loss. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 16:20:44 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 13:08:33 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Tang",
"Siao",
""
],
[
"Ma",
"Xinyin",
""
],
[
"Fang",
"Gongfan",
""
],
[
"Wang",
"Xinchao",
""
]
] |
2506.18838 | Tarik Aougab | Tarik Aougab, Matt Clay, Tawfiq Hamed | Subgraph Entropy | 16 pages, 4 figures | null | null | null | math.GT math.DS | http://creativecommons.org/licenses/by/4.0/ | Given $r \geq 3$, we prove that there exists $\lambda >0$ depending only on $r$ so that if $G$ is a metric graph of rank $r$ with metric entropy $1$, then there exists a proper subgraph $H$ of $G$ with metric entropy at least $\lambda$. This answers a question of the second two authors together with Rieck. We interpret this as a graph theoretic version of the Bers Lemma from hyperbolic geometry, and explain some connections to the pressure metric on the Culler-Vogtmann Outer Space. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 16:57:40 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 15:44:16 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Aougab",
"Tarik",
""
],
[
"Clay",
"Matt",
""
],
[
"Hamed",
"Tawfiq",
""
]
] |
2506.18855 | Aviral Prakash | Aviral Prakash, Ben S. Southworth, Marc L. Klasky | ECLEIRS: Exact conservation law embedded identification of reduced states for parameterized partial differential equations from sparse and noisy data | null | null | null | null | physics.comp-ph physics.data-an | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multi-query applications such as parameter estimation, uncertainty quantification and design optimization for parameterized PDE systems are expensive due to the high computational cost of high-fidelity simulations. Reduced/Latent state dynamics approaches for parameterized PDEs offer a viable method where high-fidelity data and machine learning techniques are used to reduce the system's dimensionality and estimate the dynamics of low-dimensional reduced states. These reduced state dynamics approaches rely on high-quality data and struggle with highly sparse spatiotemporal noisy measurements typically obtained from experiments. Furthermore, there is no guarantee that these models satisfy governing physical conservation laws, especially for parameters that are not a part of the model learning process. In this article, we propose a reduced state dynamics approach, which we refer to as ECLEIRS, that satisfies conservation laws exactly even for parameters unseen in the model training process. ECLEIRS is demonstrated for two applications: 1) obtaining clean solution signals from sparse and noisy measurements of parametric systems, and 2) predicting dynamics for unseen system parameters. We compare ECLEIRS with other reduced state dynamics approaches, those that do not enforce any physical constraints and those with physics-informed loss functions, for three shock-propagation problems: 1-D advection, 1-D Burgers and 2-D Euler equations. The numerical experiments conducted in this study demonstrate that ECLEIRS provides the most accurate prediction of dynamics for unseen parameters even in the presence of highly sparse and noisy data. We also demonstrate that ECLEIRS yields solutions and fluxes that satisfy the governing conservation law up to machine precision for unseen parameters, while the other methods yield much higher errors and do not satisfy conservation laws. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 17:18:11 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Prakash",
"Aviral",
""
],
[
"Southworth",
"Ben S.",
""
],
[
"Klasky",
"Marc L.",
""
]
] |
2506.18882 | Houyuan Chen | Hong Li, Houyuan Chen, Chongjie Ye, Zhaoxi Chen, Bohan Li, Shaocong Xu, Xianda Guo, Xuhui Liu, Yikai Wang, Baochang Zhang, Satoshi Ikehata, Boxin Shi, Anyi Rao, Hao Zhao | Light of Normals: Unified Feature Representation for Universal Photometric Stereo | Home: https://houyuanchen111.github.io/lino.github.io Github: https://github.com/houyuanchen111/LINO_UniPS HuggingFace Demo: https://huggingface.co/spaces/houyuanchen/lino | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Universal photometric stereo (PS) aims to recover high-quality surface normals from objects under arbitrary lighting conditions without relying on specific illumination models. Despite recent advances such as SDM-UniPS and Uni MS-PS, two fundamental challenges persist: 1) the deep coupling between varying illumination and surface normal features, where ambiguity in observed intensity makes it difficult to determine whether brightness variations stem from lighting changes or surface orientation; and 2) the preservation of high-frequency geometric details in complex surfaces, where intricate geometries create self-shadowing, inter-reflections, and subtle normal variations that conventional feature processing operations struggle to capture accurately. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 17:53:11 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 15:34:59 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Li",
"Hong",
""
],
[
"Chen",
"Houyuan",
""
],
[
"Ye",
"Chongjie",
""
],
[
"Chen",
"Zhaoxi",
""
],
[
"Li",
"Bohan",
""
],
[
"Xu",
"Shaocong",
""
],
[
"Guo",
"Xianda",
""
],
[
"Liu",
"Xuhui",
""
],
[
"Wang",
"Yikai",
""
],
[
"Zhang",
"Baochang",
""
],
[
"Ikehata",
"Satoshi",
""
],
[
"Shi",
"Boxin",
""
],
[
"Rao",
"Anyi",
""
],
[
"Zhao",
"Hao",
""
]
] |
2506.18889 | Eleni Koronaki | G.P. Gakis, G. Loacham\'in-Suntaxi, I.G. Aviziotis, E.D. Koronaki, P.P. Filippatos, G. Chatzigiannakis, S.P.A. Bordas, D. Davazoglou, A.G. Boudouvis | CFD Modelling and Sensitivity-Guided Design of Silicon Filament CVD Reactors | null | null | null | null | physics.app-ph physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | Filament-based chemical vapor deposition (CVD) for silicon (Si) coatings is often treated as a straightforward adaptation of planar deposition, just with a cylindrical substrate. But this overlooks a fundamental shift in how transport phenomena and reaction kinetics interact. In filament CVD, the filament is not just a substrate; it is the dominant heat source and flow disruptor all in one. In this work, we ask: What really governs Si film growth on filaments? Using a validated three-dimensional computational fluid dynamics (CFD) model, we show that filament geometry, thermal gradients, and flow-induced buoyancy do not merely affect uniformity -- they define the very regimes (reaction-limited, transition, diffusion-limited) in which deposition occurs. Our model, validated against three independent experimental studies (R-squared = 0.969), reveals how temperature, flow rate, and reactor layout fundamentally shift the growth dynamics. We show that reducing filament diameter from 2 mm to 500 microns does not just increase local surface area -- it can triple growth rates (up to approximately 39 microns per minute), but only in carefully tuned high-temperature regimes. Similarly, multi-filament setups do not simply scale deposition, they reshape flow fields and compound thermal asymmetries, sometimes improving uniformity, other times worsening it. These effects are not evident without modeling the interplay between transport and kinetics, something experiments alone cannot resolve. To bridge physical understanding with design actionability, we apply global sensitivity analysis via Polynomial Chaos Expansion (PCE) and Sobol indices. These reveal how in reaction-limited regimes, temperature dominates; in diffusion-limited regimes, reactant transport takes over. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 17:57:26 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Gakis",
"G. P.",
""
],
[
"Loachamín-Suntaxi",
"G.",
""
],
[
"Aviziotis",
"I. G.",
""
],
[
"Koronaki",
"E. D.",
""
],
[
"Filippatos",
"P. P.",
""
],
[
"Chatzigiannakis",
"G.",
""
],
[
"Bordas",
"S. P. A.",
""
],
[
"Davazoglou",
"D.",
""
],
[
"Boudouvis",
"A. G.",
""
]
] |
2506.18902 | Han Xiao | Michael G\"unther, Saba Sturua, Mohammad Kalim Akram, Isabelle Mohr, Andrei Ungureanu, Bo Wang, Sedigheh Eslami, Scott Martens, Maximilian Werk, Nan Wang, Han Xiao | jina-embeddings-v4: Universal Embeddings for Multimodal Multilingual Retrieval | 22 pages, 1-10 main, 14-22 experimental results, benchmark tables | null | null | null | cs.AI cs.CL cs.IR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We introduce jina-embeddings-v4, a 3.8 billion parameter multimodal embedding model that unifies text and image representations through a novel architecture supporting both single-vector and multi-vector embeddings in the late interaction style. The model incorporates task-specific Low-Rank Adaptation (LoRA) adapters to optimize performance across diverse retrieval scenarios, including query-document retrieval, semantic text similarity, and code search. Comprehensive evaluations demonstrate that jina-embeddings-v4 achieves state-of-the-art performance on both single-modal and cross-modal retrieval tasks, with particular strength in processing visually rich content such as tables, charts, diagrams, and mixed-media formats. To facilitate evaluation of this capability, we also introduce Jina-VDR, a novel benchmark specifically designed for visually rich image retrieval. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 17:59:55 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Jun 2025 15:52:37 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Günther",
"Michael",
""
],
[
"Sturua",
"Saba",
""
],
[
"Akram",
"Mohammad Kalim",
""
],
[
"Mohr",
"Isabelle",
""
],
[
"Ungureanu",
"Andrei",
""
],
[
"Wang",
"Bo",
""
],
[
"Eslami",
"Sedigheh",
""
],
[
"Martens",
"Scott",
""
],
[
"Werk",
"Maximilian",
""
],
[
"Wang",
"Nan",
""
],
[
"Xiao",
"Han",
""
]
] |
2506.18907 | Saeed Hashemi Sababe | Saeed Hashemi Sababe, Ismail Nikoufar | $L^p$-Theory and Noncommutative Geometry in Quantum Harmonic Analysis | null | null | null | null | math.FA math.OA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quantum harmonic analysis extends classical harmonic analysis by integrating quantum mechanical observables, replacing functions with operators and classical convolution structures with their noncommutative counterparts. This paper explores four interrelated developments in this field: (i) a noncommutative $L^p$-theory tailored for quantum harmonic analysis, (ii) the extension of quantum harmonic analysis beyond Euclidean spaces to include Lie groups and homogeneous spaces, (iii) its deep connections with Connes' noncommutative geometry, and (iv) the role of spectral synthesis and approximation properties in quantum settings. We establish novel results concerning the structure and spectral properties of quantum Segal algebras, analyze their functional-analytic aspects, and discuss their implications in quantum physics and operator theory. Our findings provide a unified framework for quantum harmonic analysis, laying the foundation for further advancements in noncommutative analysis and mathematical physics. | [
{
"version": "v1",
"created": "Tue, 08 Apr 2025 20:15:31 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Sababe",
"Saeed Hashemi",
""
],
[
"Nikoufar",
"Ismail",
""
]
] |
2506.18908 | Hao Yu | Hao Yu | Proof of admissible weights | null | null | null | null | math.FA | http://creativecommons.org/licenses/by/4.0/ | Admissable weight is an important tool for studying spectral invariance in operator algebra. Common admissable weights include polynomial weights and sub exponential weights. This article mainly provides a proof that polynomial weights are permissible weights. | [
{
"version": "v1",
"created": "Sat, 26 Apr 2025 14:53:59 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Yu",
"Hao",
""
]
] |
2506.18909 | Marko Kosti\'c | Marko Kostic | Multidimensional vector-valued Laplace transform and applications | null | null | null | null | math.FA | http://creativecommons.org/licenses/by/4.0/ | In this paper, we introduce and analyze multidimensional vector-valued Laplace transform of functions with values in sequentially complete locally convex spaces. A great number of our results seem to be new even for the functions with values in Banach spaces. We provide several new applications of multidimensional vector-valued Laplace transform to the abstract Volterra integro-differential inclusions with multiple variables, as well. | [
{
"version": "v1",
"created": "Tue, 29 Apr 2025 13:57:47 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Kostic",
"Marko",
""
]
] |
2506.18910 | Di Zhang | Di Zhang and Ligang Liu | Asymptotic analysis and design of linear elastic shell lattice metamaterials | null | null | null | null | math.AP cs.NA math.NA | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We present an asymptotic analysis of shell lattice metamaterials based on Ciarlet's shell theory, introducing a new metric--asymptotic directional stiffness (ADS)--to quantify how the geometry of the middle surface governs the effective stiffness. We prove a convergence theorem that rigorously characterizes ADS and establishes its upper bound, along with necessary and sufficient condition for achieving it. As a key result, our theory provides the first rigorous explanation for the high bulk modulus observed in Triply Periodic Minimal Surfaces (TPMS)-based shell lattices. To optimize ADS on general periodic surfaces, we propose a triangular-mesh-based discretization and shape optimization framework. Numerical experiments validate the theoretical findings and demonstrate the effectiveness of the optimization under various design objectives. Our implementation is available at https://github.com/lavenklau/minisurf. | [
{
"version": "v1",
"created": "Mon, 19 May 2025 13:53:35 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Zhang",
"Di",
""
],
[
"Liu",
"Ligang",
""
]
] |
2506.18912 | Rahul Shukla | Digvijay Singh, Rahul Shukla, Karunesh Kumar Singh | On wavelet-based sampling Kantorovich operators and their study in multi-resolution analysis | 18 pages | null | null | null | math.FA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, wavelet-based filtering operators are constructed by introducing a basic function $D(t_1, t_2, t_3)$ using a general wavelet transform. The cardinal orthogonal scaling functions (COSF) provide an idea to derive the standard sampling theorem in multiresolution spaces which motivates us to study wavelet approximation analysis. With the help of modulus of continuity, we establish a fundamental theorem of approximation. Moreover, we unfold some other aspects in the form of an upper bound of the estimation taken between the operators and functions with various conditions. In that order, a rate of convergence corresponding to the wavelet-based filtering operators is derived, by which we are able to draw some important interferences regarding the error near the sharp edges and smooth areas of the function. Eventually, some examples are demonstrated and empirically proven to justify the fact about the rate of convergence. Besides that, some derivation of inequalities with justifications through examples and important remarks emphasizes the depth and significance of our work. | [
{
"version": "v1",
"created": "Thu, 29 May 2025 19:59:54 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Singh",
"Digvijay",
""
],
[
"Shukla",
"Rahul",
""
],
[
"Singh",
"Karunesh Kumar",
""
]
] |
2506.18913 | K Mahesh Krishna | K. Mahesh Krishna | p-adic Ghobber-Jaming Uncertainty Principle | 11 Pages, 0 Figures | null | null | null | math.FA cs.IT math-ph math.IT math.MP math.NT math.OC | http://creativecommons.org/licenses/by/4.0/ | Let $\{\tau_j\}_{j=1}^n$ and $\{\omega_k\}_{k=1}^n$ be two orthonormal bases for a finite dimensional p-adic Hilbert space $\mathcal{X}$. Let $M,N\subseteq \{1, \dots, n\}$ be such that \begin{align*} \displaystyle \max_{j \in M, k \in N}|\langle \tau_j, \omega_k \rangle|<1, \end{align*} where $o(M)$ is the cardinality of $M$. Then for all $x \in \mathcal{X}$, we show that \begin{align} (1) \quad \quad \quad \quad \|x\|\leq \left(\frac{1}{1-\displaystyle \max_{j \in M, k \in N}|\langle \tau_j, \omega_k \rangle|}\right)\max\left\{\displaystyle \max_{j \in M^c}|\langle x, \tau_j\rangle |, \displaystyle \max_{k \in N^c}|\langle x, \omega_k\rangle |\right\}. \end{align}
We call Inequality (1) as \textbf{p-adic Ghobber-Jaming Uncertainty Principle}. Inequality (1) is the p-adic version of uncertainty principle obtained by Ghobber and Jaming \textit{[Linear Algebra Appl., 2011]}. We also derive analogues of Inequality (1) for non-Archimedean Banach spaces. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 03:08:01 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Krishna",
"K. Mahesh",
""
]
] |
2506.18914 | Anton Alexa | Anton Alexa | Essential Self-Adjointness of the Geometric Deformation Operator on a Compact Interval | 5 pages | null | null | null | math.SP math.FA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We define a second-order differential operator $\hat{C}$ on the Hilbert space $L^2([-v_c, v_c])$, constructed from a smooth deformation function $C(v)$. The operator is considered on the Sobolev domain $H^2([-v_c, v_c]) \cap H^1_0([-v_c, v_c])$ with Dirichlet boundary conditions. We prove that $\hat{C}$ is essentially self-adjoint by verifying its symmetry and computing von Neumann deficiency indices, which vanish. All steps are carried out explicitly. This result ensures the mathematical consistency of the operator and enables future spectral analysis on compact intervals. | [
{
"version": "v1",
"created": "Sun, 08 Jun 2025 01:58:04 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Alexa",
"Anton",
""
]
] |
2506.18915 | Siyang Song | Siyang Song, Yupeng Huo, Shiqing Tang, Jiaee Cheong, Rui Gao, Michel Valstar, Hatice Gunes | Automatic Depression Assessment using Machine Learning: A Comprehensive Survey | null | null | null | null | q-bio.NC cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Depression is a common mental illness across current human society. Traditional depression assessment relying on inventories and interviews with psychologists frequently suffer from subjective diagnosis results, slow and expensive diagnosis process as well as lack of human resources. Since there is a solid evidence that depression is reflected by various human internal brain activities and external expressive behaviours, early traditional machine learning (ML) and advanced deep learning (DL) models have been widely explored for human behaviour-based automatic depression assessment (ADA) since 2012. However, recent ADA surveys typically only focus on a limited number of human behaviour modalities. Despite being used as a theoretical basis for developing ADA approaches, existing ADA surveys lack a comprehensive review and summary of multi-modal depression-related human behaviours. To bridge this gap, this paper specifically summarises depression-related human behaviours across a range of modalities (e.g. the human brain, verbal language and non-verbal audio/facial/body behaviours). We focus on conducting an up-to-date and comprehensive survey of ML-based ADA approaches for learning depression cues from these behaviours as well as discussing and comparing their distinctive features and limitations. In addition, we also review existing ADA competitions and datasets, identify and discuss the main challenges and opportunities to provide further research directions for future ADA researchers. | [
{
"version": "v1",
"created": "Mon, 09 Jun 2025 14:40:16 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Song",
"Siyang",
""
],
[
"Huo",
"Yupeng",
""
],
[
"Tang",
"Shiqing",
""
],
[
"Cheong",
"Jiaee",
""
],
[
"Gao",
"Rui",
""
],
[
"Valstar",
"Michel",
""
],
[
"Gunes",
"Hatice",
""
]
] |
2506.18916 | Ganesh Parab | Ganesh Parab, Zishan Ahmad, and Dagnachew Birru | HI-SQL: Optimizing Text-to-SQL Systems through Dynamic Hint Integration | null | null | null | null | cs.LG cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Text-to-SQL generation bridges the gap between natural language and databases, enabling users to query data without requiring SQL expertise. While large language models (LLMs) have significantly advanced the field, challenges remain in handling complex queries that involve multi-table joins, nested conditions, and intricate operations. Existing methods often rely on multi-step pipelines that incur high computational costs, increase latency, and are prone to error propagation. To address these limitations, we propose HI-SQL, a pipeline that incorporates a novel hint generation mechanism utilizing historical query logs to guide SQL generation. By analyzing prior queries, our method generates contextual hints that focus on handling the complexities of multi-table and nested operations. These hints are seamlessly integrated into the SQL generation process, eliminating the need for costly multi-step approaches and reducing reliance on human-crafted prompts. Experimental evaluations on multiple benchmark datasets demonstrate that our approach significantly improves query accuracy of LLM-generated queries while ensuring efficiency in terms of LLM calls and latency, offering a robust and practical solution for enhancing Text-to-SQL systems. | [
{
"version": "v1",
"created": "Wed, 11 Jun 2025 12:07:55 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Parab",
"Ganesh",
""
],
[
"Ahmad",
"Zishan",
""
],
[
"Birru",
"Dagnachew",
""
]
] |
2506.18917 | Louis du Plessis | Louis Du Plessis, Christo Venter, Alice K. Harding, Zorawar Wadiasingh, Constantinos Kalapotharakos | Towards Modelling AR Sco: Calibration -- Reproducing High-Energy Pulsar Emission and Testing Convergence to Aristotelian Electrodynamics | 24 pages, 21 figures Follow up to methods paper | null | 10.1093/mnras/staf958 | null | astro-ph.HE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, kinetic simulations have been crucial to further our understanding of pulsar electrodynamics. Yet, due to the large-scale separation between the gyro-period vs the stellar rotation period, resolving the particle gyration has been computationally unfeasible for realistic pulsar parameters. The main aim of this work is comparing our gyro-phase-resolved model with a gyro-centric pulsar model, where our model solves the general equations of motion with included radiation reaction using a higher-order numerical solver with adaptive time steps. Specifically, we aim to (i) reproduce a pulsar's high-energy emission maps, namely one with $10\%$ the surface $B$-field strength of Vela, and spectra produced by an independent gyro-centric pulsar emission model; (ii) test convergence of these results to the radiation-reaction limit of Aristotelian Electrodynamics. (iii) Additionally, we identify the effect that a large $E_{\parallel}$-field has on the trajectories and radiation calculations. We find that we can reproduce the curvature radiation emission maps and spectra well, using $10\%$ field strengths of the Vela pulsar and injecting our particles at a higher altitude in the magnetosphere. Using sufficiently large $E_{\parallel}$-fields, our numeric results converge to the analytic radiation-reaction limit trajectories. Additionally, we illustrate the importance of accounting for the $\mathbf{E}\times \mathbf{B}$-drift in the particle trajectories and radiation calculations, validating the Harding and collaborators' model approach. Lastly, we found that our model deals very well with the high-radiation-reaction and high-field regimes present in pulsars. | [
{
"version": "v1",
"created": "Thu, 12 Jun 2025 08:38:43 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Plessis",
"Louis Du",
""
],
[
"Venter",
"Christo",
""
],
[
"Harding",
"Alice K.",
""
],
[
"Wadiasingh",
"Zorawar",
""
],
[
"Kalapotharakos",
"Constantinos",
""
]
] |
2506.18918 | Daniel Tubbenhauer | Daniel Tubbenhauer | Quantum topology without topology | 265 pages, many figures, comments welcome | null | null | null | math.QA math.CT math.RT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | These lecture notes cover 13 sessions and are presented as an e-print, intended to evolve over time.
Quantum invariants do more than distinguish topological objects; they build bridges between topology, algebra, number theory and quantum physics helping to transfer ideas, and stimulating mutual development. They also possess deep and intriguing connections to representation theory, particularly through representations of quantum groups.
These lecture notes aim to illustrate how categorical algebra provides a framework for studying both algebra and topology. Specifically, they demonstrate how quantum invariants emerge naturally from a mostly categorical perspective. | [
{
"version": "v1",
"created": "Fri, 13 Jun 2025 07:56:18 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Tubbenhauer",
"Daniel",
""
]
] |
2506.18919 | Hexiang Gu | Hexiang Gu, Qifan Yu, Saihui Hou, Zhiqin Fang, Huijia Wu, Zhaofeng He | MemeMind: A Large-Scale Multimodal Dataset with Chain-of-Thought Reasoning for Harmful Meme Detection | null | null | null | null | cs.CL cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rapid development of social media has intensified the spread of harmful content. Harmful memes, which integrate both images and text, pose significant challenges for automated detection due to their implicit semantics and complex multimodal interactions. Although existing research has made progress in detection accuracy and interpretability, the lack of a systematic, large-scale, diverse, and highly explainable dataset continues to hinder further advancement in this field. To address this gap, we introduce MemeMind, a novel dataset featuring scientifically rigorous standards, large scale, diversity, bilingual support (Chinese and English), and detailed Chain-of-Thought (CoT) annotations. MemeMind fills critical gaps in current datasets by offering comprehensive labeling and explicit reasoning traces, thereby providing a solid foundation for enhancing harmful meme detection. In addition, we propose an innovative detection framework, MemeGuard, which effectively integrates multimodal information with reasoning process modeling, significantly improving models' ability to understand and identify harmful memes. Extensive experiments conducted on the MemeMind dataset demonstrate that MemeGuard consistently outperforms existing state-of-the-art methods in harmful meme detection tasks. | [
{
"version": "v1",
"created": "Sun, 15 Jun 2025 13:45:30 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Gu",
"Hexiang",
""
],
[
"Yu",
"Qifan",
""
],
[
"Hou",
"Saihui",
""
],
[
"Fang",
"Zhiqin",
""
],
[
"Wu",
"Huijia",
""
],
[
"He",
"Zhaofeng",
""
]
] |
2506.18920 | Michael Williams | Michael Williams | Signal Use and Emergent Cooperation | 167 pages, 19 figures, PhD dissertation, UCLA, 2006 | null | null | null | cs.AI cs.LG cs.MA cs.NE cs.SI | http://creativecommons.org/licenses/by/4.0/ | In this work, we investigate how autonomous agents, organized into tribes, learn to use communication signals to coordinate their activities and enhance their collective efficiency. Using the NEC-DAC (Neurally Encoded Culture - Distributed Autonomous Communicators) system, where each agent is equipped with its own neural network for decision-making, we demonstrate how these agents develop a shared behavioral system -- akin to a culture -- through learning and signalling. Our research focuses on the self-organization of culture within these tribes of agents and how varying communication strategies impact their fitness and cooperation. By analyzing different social structures, such as authority hierarchies, we show that the culture of cooperation significantly influences the tribe's performance. Furthermore, we explore how signals not only facilitate the emergence of culture but also enable its transmission across generations of agents. Additionally, we examine the benefits of coordinating behavior and signaling within individual agents' neural networks. | [
{
"version": "v1",
"created": "Mon, 16 Jun 2025 20:24:30 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Williams",
"Michael",
""
]
] |
2506.18922 | Yiran Zhou | Yiran Zhou, Yingyu Wang, Shoudong Huang and Liang Zhao | Correspondence-Free Multiview Point Cloud Registration via Depth-Guided Joint Optimisation | 8 pages, accepted for publication in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2025) | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multiview point cloud registration is a fundamental task for constructing globally consistent 3D models. Existing approaches typically rely on feature extraction and data association across multiple point clouds; however, these processes are challenging to obtain global optimal solution in complex environments. In this paper, we introduce a novel correspondence-free multiview point cloud registration method. Specifically, we represent the global map as a depth map and leverage raw depth information to formulate a non-linear least squares optimisation that jointly estimates poses of point clouds and the global map. Unlike traditional feature-based bundle adjustment methods, which rely on explicit feature extraction and data association, our method bypasses these challenges by associating multi-frame point clouds with a global depth map through their corresponding poses. This data association is implicitly incorporated and dynamically refined during the optimisation process. Extensive evaluations on real-world datasets demonstrate that our method outperforms state-of-the-art approaches in accuracy, particularly in challenging environments where feature extraction and data association are difficult. | [
{
"version": "v1",
"created": "Wed, 18 Jun 2025 05:14:12 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Zhou",
"Yiran",
""
],
[
"Wang",
"Yingyu",
""
],
[
"Huang",
"Shoudong",
""
],
[
"Zhao",
"Liang",
""
]
] |
2506.18923 | Pengyu Nie | Yifan Zong, Yuntian Deng, Pengyu Nie | Mix-of-Language-Experts Architecture for Multilingual Programming | Accepted at LLM4Code @ ICSE 2025 | null | null | null | cs.PL cs.CL cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) have demonstrated impressive capabilities in aiding developers with tasks like code comprehension, generation, and translation. Supporting multilingual programming -- i.e., coding tasks across multiple programming languages -- typically requires either (1) finetuning a single LLM across all programming languages, which is cost-efficient but sacrifices language-specific specialization and performance, or (2) finetuning separate LLMs for each programming language, which allows for specialization but is computationally expensive and storage-intensive due to the duplication of parameters. This paper introduces MoLE (Mix-of-Language-Experts), a novel architecture that balances efficiency and specialization for multilingual programming. MoLE is composed of a base model, a shared LoRA (low-rank adaptation) module, and a collection of language-specific LoRA modules. These modules are jointly optimized during the finetuning process, enabling effective knowledge sharing and specialization across programming languages. During inference, MoLE automatically routes to the language-specific LoRA module corresponding to the programming language of the code token being generated. Our experiments demonstrate that MoLE achieves greater parameter efficiency compared to training separate language-specific LoRAs, while outperforming a single shared LLM finetuned for all programming languages in terms of accuracy. | [
{
"version": "v1",
"created": "Wed, 18 Jun 2025 06:20:51 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Zong",
"Yifan",
""
],
[
"Deng",
"Yuntian",
""
],
[
"Nie",
"Pengyu",
""
]
] |
2506.18924 | Nonso Nnamoko | Ammar K Al Mhdawi and Nonso Nnamoko and Safanah Mudheher Raafat and M.K.S. Al-Mhdawi and Amjad J Humaidi | Connecting Vision and Emissions: A Behavioural AI Approach to Carbon Estimation in Road Design | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | We present an enhanced YOLOv8 real time vehicle detection and classification framework, for estimating carbon emissions in urban environments. The system enhances YOLOv8 architecture to detect, segment, and track vehicles from live traffic video streams. Once a vehicle is localized, a dedicated deep learning-based identification module is employed to recognize license plates and classify vehicle types. Since YOLOv8 lacks the built-in capacity for fine grained recognition tasks such as reading license plates or determining vehicle attributes beyond class labels, our framework incorporates a hybrid pipeline where each detected vehicle is tracked and its bounding box is cropped and passed to a deep Optical Character Recognition (OCR) module. This OCR system, composed of multiple convolutional neural network (CNN) layers, is trained specifically for character-level detection and license plate decoding under varied conditions such as motion blur, occlusion, and diverse font styles. Additionally, the recognized plate information is validated using a real time API that cross references with an external vehicle registration database to ensure accurate classification and emission estimation. This multi-stage approach enables precise, automated calculation of per vehicle carbon emissions. Extensive evaluation was conducted using a diverse vehicle dataset enriched with segmentation masks and annotated license plates. The YOLOv8 detector achieved a mean Average Precision (mAP@0.5) of approximately 71% for bounding boxes and 70% for segmentation masks. Character level OCR accuracy reached up to 99% with the best performing CNN model. These results affirm the feasibility of combining real time object detection with deep OCR for practical deployment in smart transportation systems, offering a scalable solution for automated, vehicle specific carbon emission monitoring. | [
{
"version": "v1",
"created": "Wed, 18 Jun 2025 11:50:24 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Mhdawi",
"Ammar K Al",
""
],
[
"Nnamoko",
"Nonso",
""
],
[
"Raafat",
"Safanah Mudheher",
""
],
[
"Al-Mhdawi",
"M. K. S.",
""
],
[
"Humaidi",
"Amjad J",
""
]
] |
2506.18925 | Tahereh Zarrat Ehsan | Tahereh Zarrat Ehsan, Michael Tangermann, Ya\u{g}mur G\"u\c{c}l\"ut\"urk, Bastiaan R. Bloem, Luc J. W. Evers | Interpretable and Granular Video-Based Quantification of Motor Characteristics from the Finger Tapping Test in Parkinson Disease | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurately quantifying motor characteristics in Parkinson disease (PD) is crucial for monitoring disease progression and optimizing treatment strategies. The finger-tapping test is a standard motor assessment. Clinicians visually evaluate a patient's tapping performance and assign an overall severity score based on tapping amplitude, speed, and irregularity. However, this subjective evaluation is prone to inter- and intra-rater variability, and does not offer insights into individual motor characteristics captured during this test. This paper introduces a granular computer vision-based method for quantifying PD motor characteristics from video recordings. Four sets of clinically relevant features are proposed to characterize hypokinesia, bradykinesia, sequence effect, and hesitation-halts. We evaluate our approach on video recordings and clinical evaluations of 74 PD patients from the Personalized Parkinson Project. Principal component analysis with varimax rotation shows that the video-based features corresponded to the four deficits. Additionally, video-based analysis has allowed us to identify further granular distinctions within sequence effect and hesitation-halts deficits. In the following, we have used these features to train machine learning classifiers to estimate the Movement Disorder Society Unified Parkinson Disease Rating Scale (MDS-UPDRS) finger-tapping score. Compared to state-of-the-art approaches, our method achieves a higher accuracy in MDS-UPDRS score prediction, while still providing an interpretable quantification of individual finger-tapping motor characteristics. In summary, the proposed framework provides a practical solution for the objective assessment of PD motor characteristics, that can potentially be applied in both clinical and remote settings. Future work is needed to assess its responsiveness to symptomatic treatment and disease progression. | [
{
"version": "v1",
"created": "Thu, 19 Jun 2025 12:49:06 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Ehsan",
"Tahereh Zarrat",
""
],
[
"Tangermann",
"Michael",
""
],
[
"Güçlütürk",
"Yağmur",
""
],
[
"Bloem",
"Bastiaan R.",
""
],
[
"Evers",
"Luc J. W.",
""
]
] |
2506.18926 | Fuzel Shaik Mr | Fuzel Shaik, Getnet Demil, Mourad Oussalah | AI-based Approach in Early Warning Systems: Focus on Emergency Communication Ecosystem and Citizen Participation in Nordic Countries | null | null | null | null | cs.CY cs.AI | http://creativecommons.org/licenses/by/4.0/ | Climate change and natural disasters are recognized as worldwide challenges requiring complex and efficient ecosystems to deal with social, economic, and environmental effects. This chapter advocates a holistic approach, distinguishing preparedness, emergency responses, and postcrisis phases. The role of the Early Warning System (EWS), Risk modeling and mitigation measures are particularly emphasized. The chapter reviews the various Artificial Intelligence (AI)-enabler technologies that can be leveraged at each phase, focusing on the INFORM risk framework and EWSs. Emergency communication and psychological risk perception have been emphasized in emergency response times. Finally, a set of case studies from Nordic countries has been highlighted. | [
{
"version": "v1",
"created": "Fri, 20 Jun 2025 12:32:16 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Shaik",
"Fuzel",
""
],
[
"Demil",
"Getnet",
""
],
[
"Oussalah",
"Mourad",
""
]
] |
2506.18928 | Lingyu Yang | Lingyu Yang (1) ((1) Shanghai Jiao Tong University) | Do LLMs Know When to Flip a Coin? Strategic Randomization through Reasoning and Experience | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Strategic randomization is a key principle in game theory, yet it remains underexplored in large language models (LLMs). Prior work often conflates the cognitive decision to randomize with the mechanical generation of randomness, leading to incomplete evaluations. To address this, we propose a novel zero-sum game inspired by the Tian Ji Horse Race, where the Nash equilibrium corresponds to a maximal entropy strategy. The game's complexity masks this property from untrained humans and underdeveloped LLMs. We evaluate five LLMs across prompt styles -- framed, neutral, and hinted -- using competitive multi-tournament gameplay with system-provided random choices, isolating the decision to randomize. Results show that weaker models remain deterministic regardless of prompts, while stronger models exhibit increased randomization under explicit hints. When facing weaker models, strong LLMs adopt deterministic strategies to exploit biases, but converge toward equilibrium play when facing peers. Through win/loss outcomes and Bayes factor analysis, we demonstrate meaningful variation in LLMs' strategic reasoning capabilities, highlighting opportunities for improvement in abstract reasoning and adaptive learning. We make our implementation publicly available at https://github.com/ocelopus/llm-when-to-throw-coin to ensure full reproducibility. | [
{
"version": "v1",
"created": "Sat, 21 Jun 2025 05:26:33 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Yang",
"Lingyu",
"",
"Shanghai Jiao Tong University"
]
] |
2506.18930 | Da Chen | Chong Di, Shuwang Zhou, Da Chen, Jean-Marie Mirebeau, Minglei Shu, Laurent D. Cohen | Reinforcement Learning-Based Dynamic Grouping for Tubular Structure Tracking | null | null | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The computation of minimal paths for the applications in tracking tubular structures such as blood vessels and roads is challenged by complex morphologies and environmental variations. Existing approaches can be roughly categorized into two research lines: the point-wise based models and the segment-wise based models. Although segment-wise approaches have obtained promising results in many scenarios, they often suffer from computational inefficiency and heavily rely on a prescribed prior to fit the target elongated shapes. We propose a novel framework that casts segment-wise tracking as a Markov Decision Process (MDP), enabling a reinforcement learning approach. Our method leverages Q-Learning to dynamically explore a graph of segments, computing edge weights on-demand and adaptively expanding the search space. This strategy avoids the high cost of a pre-computed graph and proves robust to incomplete initial information. Experimental reuslts on typical tubular structure datasets demonstrate that our method significantly outperforms state-of-the-art point-wise and segment-wise approaches. The proposed method effectively handles complex topologies and maintains global path coherence without depending on extensive prior structural knowledge. | [
{
"version": "v1",
"created": "Sat, 21 Jun 2025 11:00:17 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Di",
"Chong",
""
],
[
"Zhou",
"Shuwang",
""
],
[
"Chen",
"Da",
""
],
[
"Mirebeau",
"Jean-Marie",
""
],
[
"Shu",
"Minglei",
""
],
[
"Cohen",
"Laurent D.",
""
]
] |
2506.18931 | Shuang Ao | Shuang Ao, Yi Dong, Jinwei Hu, Sarvapali Ramchurn | Safe Pruning LoRA: Robust Distance-Guided Pruning for Safety Alignment in Adaptation of LLMs | 13 pages, 3 figures | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fine-tuning Large Language Models (LLMs) with Low-Rank Adaptation (LoRA) enhances adaptability while reducing computational costs. However, fine-tuning can compromise safety alignment, even with benign data, increasing susceptibility to harmful outputs. Existing safety alignment methods struggle to capture complex parameter shifts, leading to suboptimal safety-utility trade-offs. To address this issue, we propose Safe Pruning LoRA (SPLoRA), a novel pruning-based approach that selectively removes LoRA layers that weaken safety alignment, improving safety while preserving performance. At its core, we introduce Empirical-DIEM (E-DIEM), a dimension-insensitive similarity metric that effectively detects safety misalignment in LoRA-adapted models. We conduct extensive experiments on LLMs fine-tuned with mixed of benign and malicious data, and purely benign datasets, evaluating SPLoRA across utility, safety, and reliability metrics. Results demonstrate that SPLoRA outperforms state-of-the-art safety alignment techniques, significantly reducing safety risks while maintaining or improving model performance and reliability. Additionally, SPLoRA reduces inference overhead, making it a scalable and efficient solution for deploying safer and more reliable LLMs. The code is available at https://github.com/AoShuang92/SPLoRA. | [
{
"version": "v1",
"created": "Sat, 21 Jun 2025 14:59:54 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Ao",
"Shuang",
""
],
[
"Dong",
"Yi",
""
],
[
"Hu",
"Jinwei",
""
],
[
"Ramchurn",
"Sarvapali",
""
]
] |
2506.18932 | Zhiqiang Lin | Zhiqiang Lin, Huan Sun, Ness Shroff | AI Safety vs. AI Security: Demystifying the Distinction and Boundaries | null | null | null | null | cs.CY cs.AI cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Artificial Intelligence (AI) is rapidly being integrated into critical systems across various domains, from healthcare to autonomous vehicles. While its integration brings immense benefits, it also introduces significant risks, including those arising from AI misuse. Within the discourse on managing these risks, the terms "AI Safety" and "AI Security" are often used, sometimes interchangeably, resulting in conceptual confusion. This paper aims to demystify the distinction and delineate the precise research boundaries between AI Safety and AI Security. We provide rigorous definitions, outline their respective research focuses, and explore their interdependency, including how security breaches can precipitate safety failures and vice versa. Using clear analogies from message transmission and building construction, we illustrate these distinctions. Clarifying these boundaries is crucial for guiding precise research directions, fostering effective cross-disciplinary collaboration, enhancing policy effectiveness, and ultimately, promoting the deployment of trustworthy AI systems. | [
{
"version": "v1",
"created": "Sat, 21 Jun 2025 18:36:03 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Lin",
"Zhiqiang",
""
],
[
"Sun",
"Huan",
""
],
[
"Shroff",
"Ness",
""
]
] |
2506.18934 | John Mashford PhD | John Mashford | Computation of the masses of the elementary particles | Correction of published paper. 33 pages, 7 figures | AIP Advances 14.015007 (2024) | 10.1063/5.0177079 | null | math-ph math.MP | http://creativecommons.org/licenses/by/4.0/ | An approach to gauge theory in the context of locally conformally flat space-time is described. It is discussed how there are a number of natural principal bundles associated with any given locally conformally flat space-time $X$. The simplest of these principal bundles is the bundle $P_X(G)$ with structure group $G=U(2,2)$. An 11 dimensional bundle $Q$ with structure group a certain 7 dimensional group $K$ is constructed by a method involving a reduction of structure group for the bundle $P_X(G)$. It is shown how the gauge groups $U(1)$, $SU(2)$ and $SU(3)$ can be derived from the geometry of locally conformally flat space-time. Fock spaces of multiparticle states for the fields of the standard model are constructed in the context of bundles with these groups as structure groups. Scattering and other particle interaction processes are defined in terms of linear maps between multiparticle state spaces. A technique for computing analytically and/or computationally the masses of the elementary particles is described. This method involves the computation of a certain quantity called the integral mass spectrum for a given family of particles and then the masses of the particles in the family are determined to be the locations of the peaks of the integral mass spectrum. The method is applied successfully in the electroweak sector to the cases of the charged leptons, $\mu$ and $\tau$, and the Z$^0$ particle. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 01:13:36 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Mashford",
"John",
""
]
] |
2506.18935 | Shri Lal Raghudev Ram Singh | Shri Lal Raghudev Ram Singh | Which Consciousness Can Be Artificialized? Local Percept-Perceiver Phenomenon for the Existence of Machine Consciousness | Paper accepted for the 18th Annual AGI Conference, AGI-2025, Reykjavik, Iceland, August 10-13, 2025 | null | null | null | q-bio.NC cs.AI | http://creativecommons.org/licenses/by/4.0/ | This paper presents a novel paradigm of the local percept-perceiver phenomenon to formalize certain observations in neuroscientific theories of consciousness. Using this model, a set-theoretic formalism is developed for artificial systems, and the existence of machine consciousness is proved by invoking Zermelo-Fraenkel set theory. The article argues for the possibility of a reductionist form of epistemic consciousness within machines. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 01:53:14 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Singh",
"Shri Lal Raghudev Ram",
""
]
] |
2506.18936 | Masanobu Yahiro Prof. | Shingo Tagami, Takaykui Myo, Masanobu Yahiro | Neutron skin thickness for $^{nat}$Pb, $^{159}$Tb and effects of deformation on matter radii for Mg, Na, Ar isotopes | null | null | null | null | nucl-ex nucl-th | http://creativecommons.org/licenses/by/4.0/ | Data on reaction cross section $\sigma_{\rm R}$ of proton scattering are available for $^{nat}$Pb (natural lead) and $^{159}$Tb. In addition, data on $\sigma_{\rm R}$ and/or interaction ones $\sigma_{\rm I}$ are available for Mg, Na, Ar isotopes. Our aim is to determine neutron skin thickness $r_{\rm skin}$ for $^{nat}$Pb, $^{159}$Tb and to investigate effects of deformation for Mg, Na, Ar isotopes. We use the chiral (Kyushu) $g$-matrix folding model for lower energies and the folding model based on the Love-Franey (LF) $t$-matrix for higher energies. For $^{159}$Tb, our skin value is $r_{\rm skin}=0.273 \pm 0.073$~fm. As for $^{24}$Mg, our matter radius $r_{\rm m}(\sigma_{\rm R})=3.03 \pm 0.08$~fm determined from $\sigma_{\rm R}$ includes effects of deformation, whereas the corresponding $r_{\rm m}(\sigma_{\rm I})=2.79 \pm 0.15$~fm does not. Effects of deformation are seen for Mg isotopes, $^{21, 23}$Na, $^{40}$Ar. Our result $r_{\rm skin}=0.271 \pm 0.008$~fm for $^{nat}$Pb agrees with $r_{\rm skin}^{208}({\rm PREX\text{-}II}) = 0.283\pm 0.071$~fm for $^{208}$Pb, where we assume $A=208$ for $^{nat}$Pb. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 07:18:06 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Tagami",
"Shingo",
""
],
[
"Myo",
"Takaykui",
""
],
[
"Yahiro",
"Masanobu",
""
]
] |
2506.18937 | Guina Flores | Guina Flores Montalico | Experiential marketing strategy and tourism demand in the contribution of the positioning of the floating islands Los Uros, Puno | 21 pages, in Spanish language | null | null | null | econ.GN q-fin.EC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Experiential focused on creating memorable and meaningful experiences for consumers, has emerged as a key strategy in promoting tourist destinations. particularly in destinations seeking to highlight their unique cultural characteristics. In this study, the influence of experiential marketing on the tourism demand of the floating islands Los Uros, Puno, an emblematic destination in Peru, is analyzed. The objective of the research was to evaluate how experiential marketing strategies impact tourism demand in this destination, focusing on the sensorial, affective and thought experiences that tourists perceive. The methodology adopted a quantitative approach, with a non experimental cross-sectional design, descriptive scope and a correlational analysis. The study population consisted of 158 tourists. To collect the data, a questionnaire based on a 5-point Likert scale was used, with a total of 21 items. Among the results, a significant positive correlation was found (r = 0.809, p < 0.001) between experiential marketing and tourism demand. Sensory (r = 0.780, p < 0.001), affective (r = 0.772, p < 0.001) and thought (r = 0.797, p < 0.001) were described as determining factors in tourists' decisions when visiting Los Uros. In addition, the multiple regression analysis indicated that experiential marketing explains 83% of the variability in tourism demand. In conclusion, experiential marketing strategies have a significant impact on tourism demand. It is recommended that marketing managers of Los Uros Island implement tactics that promote memorable experiences, prioritizing strategies that involve experiences that stimulate the thinking and emotions of tourists. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 08:54:51 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Montalico",
"Guina Flores",
""
]
] |
2506.18938 | Yanke Wang Dr. | Yanke Wang and Yu Hin Ng and Haobo Liang and Ching-Wei Chang and Hao Chen | Bird's-eye view safety monitoring for the construction top under the tower crane | null | null | null | null | cs.CV cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The tower crane is involving more automated and intelligent operation procedure, and importantly, the application of automation technologies to the safety issues is imperative ahead of the utilization of any other advances. Among diverse risk management tasks on site, it is essential to protect the human workers on the workspace between the tower crane and constructed building top area (construction top) from the bird's-eye view, especially with Modular Integrated Construction (MiC) lifted. Also, the camera and Light Detection And Ranging (LiDAR) can capture abundant 3D information on site, which is however yet made the best use. Considering the safety protection for humans and tower cranes, we present an AI-based fully automated safety monitoring system for tower crane lifting from the bird's-eye view, surveilling to shield the human workers on the construction top and avoid cranes' collision by alarming the crane operator. The system achieved a 3D data fusion for localization of humans and MiCs by integrating the captured information from camera and LiDAR. The state-of-the-art methods were explored and implemented into our proposed software pipeline coupled with the hardware and display systems. Furthermore, we conducted an analysis of the components in the pipeline to verify the accuracy and effectiveness of the involved methods. The display and visualization on the real site proved that our system can serve as a valuable safety monitoring toolkit on site. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 12:52:20 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Wang",
"Yanke",
""
],
[
"Ng",
"Yu Hin",
""
],
[
"Liang",
"Haobo",
""
],
[
"Chang",
"Ching-Wei",
""
],
[
"Chen",
"Hao",
""
]
] |
2506.18939 | Rui An | Rui An, Yifeng Zhang, Ziran Liang, Wenqi Fan, Yuxuan Liang, Xuequn Shang, Qing Li | Damba-ST: Domain-Adaptive Mamba for Efficient Urban Spatio-Temporal Prediction | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Training urban spatio-temporal foundation models that generalize well across diverse regions and cities is critical for deploying urban services in unseen or data-scarce regions. Recent studies have typically focused on fusing cross-domain spatio-temporal data to train unified Transformer-based models. However, these models suffer from quadratic computational complexity and high memory overhead, limiting their scalability and practical deployment. Inspired by the efficiency of Mamba, a state space model with linear time complexity, we explore its potential for efficient urban spatio-temporal prediction. However, directly applying Mamba as a spatio-temporal backbone leads to negative transfer and severe performance degradation. This is primarily due to spatio-temporal heterogeneity and the recursive mechanism of Mamba's hidden state updates, which limit cross-domain generalization. To overcome these challenges, we propose Damba-ST, a novel domain-adaptive Mamba-based model for efficient urban spatio-temporal prediction. Damba-ST retains Mamba's linear complexity advantage while significantly enhancing its adaptability to heterogeneous domains. Specifically, we introduce two core innovations: (1) a domain-adaptive state space model that partitions the latent representation space into a shared subspace for learning cross-domain commonalities and independent, domain-specific subspaces for capturing intra-domain discriminative features; (2) three distinct Domain Adapters, which serve as domain-aware proxies to bridge disparate domain distributions and facilitate the alignment of cross-domain commonalities. Extensive experiments demonstrate the generalization and efficiency of Damba-ST. It achieves state-of-the-art performance on prediction tasks and demonstrates strong zero-shot generalization, enabling seamless deployment in new urban environments without extensive retraining or fine-tuning. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 15:40:01 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"An",
"Rui",
""
],
[
"Zhang",
"Yifeng",
""
],
[
"Liang",
"Ziran",
""
],
[
"Fan",
"Wenqi",
""
],
[
"Liang",
"Yuxuan",
""
],
[
"Shang",
"Xuequn",
""
],
[
"Li",
"Qing",
""
]
] |
2506.18940 | Zhenke Liu | Zhenke Liu, Jien Li, Ziqi Zhang | eccDNAMamba: A Pre-Trained Model for Ultra-Long eccDNA Sequence Analysis | Accepted by ICML 2025 Generative AI and Biology (GenBio) Workshop | null | null | null | q-bio.GN cs.AI | http://creativecommons.org/licenses/by/4.0/ | Extrachromosomal circular DNA (eccDNA) plays key regulatory roles and contributes to oncogene overexpression in cancer through high-copy amplification and long-range interactions. Despite advances in modeling, no pre-trained models currently support full-length circular eccDNA for downstream analysis. Existing genomic models are either limited to single-nucleotide resolution or hindered by the inefficiency of the quadratic attention mechanism. Here, we introduce eccDNAMamba, the first bidirectional state-space encoder tailored for circular DNA sequences. It combines forward and reverse passes for full-context representation learning with linear-time complexity, and preserves circular structure through a novel augmentation strategy. Tested on two real-world datasets, eccDNAMamba achieves strong classification performance and scales to sequences up to 200 Kbp, offering a robust and efficient framework for modeling circular genomes. Our codes are available at https://github.com/zzq1zh/GenAI-Lab. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 17:50:57 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Liu",
"Zhenke",
""
],
[
"Li",
"Jien",
""
],
[
"Zhang",
"Ziqi",
""
]
] |
2506.18941 | Neha Rani | Neha Rani, Sharan Majumder, Ishan Bhardwaj, Pedro Guillermo Feijoo Garcia | Can AI support student engagement in classroom activities in higher education? | null | null | null | null | cs.CY cs.AI cs.ET cs.HC | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Lucrative career prospects and creative opportunities often attract students to enroll in computer science majors and pursue advanced studies in the field. Consequently, there has been a significant surge in enrollment in computer science courses, resulting in large class sizes that can range from hundreds to even thousands of students. A common challenge in such large classrooms is the lack of engagement between students and both the instructor and the learning material. However, with advancements in technology and improvements in large language models (LLMs), there is a considerable opportunity to utilize LLM-based AI models, such as conversational artificial intelligence (CAI), to enhance student engagement with learning content in large classes. To explore the potential of CAI to support engagement, especially with learning content, we designed an activity in a software Engineering course (with a large class size) where students used CAI for an in-class activity. We conducted a within-subject investigation in a large classroom at a US university where we compared student engagement during an in-class activity that used CAI tool vs. one without CAI tool. The CAI tool we used was ChatGPT due to its widespread popularity and familiarity. Our results indicate that CAI (ChatGPT) has the potential to support engagement with learning content during in-class activities, especially in large class sizes. We further discuss the implications of our findings. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 19:30:47 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Rani",
"Neha",
""
],
[
"Majumder",
"Sharan",
""
],
[
"Bhardwaj",
"Ishan",
""
],
[
"Garcia",
"Pedro Guillermo Feijoo",
""
]
] |
2506.18942 | Simon Hatzesberger | Simon Hatzesberger and Iris Nonneman | Advanced Applications of Generative AI in Actuarial Science: Case Studies Beyond ChatGPT | null | null | null | null | cs.CY q-fin.RM | http://creativecommons.org/licenses/by/4.0/ | This article demonstrates the transformative impact of Generative AI (GenAI) on actuarial science, illustrated by four implemented case studies. It begins with a historical overview of AI, tracing its evolution from early neural networks to modern GenAI technologies. The first case study shows how Large Language Models (LLMs) improve claims cost prediction by deriving significant features from unstructured textual data, significantly reducing prediction errors in the underlying machine learning task. In the second case study, we explore the automation of market comparisons using the GenAI concept of Retrieval-Augmented Generation to identify and process relevant information from documents. A third case study highlights the capabilities of fine-tuned vision-enabled LLMs in classifying car damage types and extracting contextual information. The fourth case study presents a multi-agent system that autonomously analyzes data from a given dataset and generates a corresponding report detailing the key findings. In addition to these case studies, we outline further potential applications of GenAI in the insurance industry, such as the automation of claims processing and fraud detection, and the verification of document compliance with internal or external policies. Finally, we discuss challenges and considerations associated with the use of GenAI, covering regulatory issues, ethical concerns, and technical limitations, among others. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 19:36:03 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Hatzesberger",
"Simon",
""
],
[
"Nonneman",
"Iris",
""
]
] |
2506.18943 | Andrew Kiruluta | Andrew Kiruluta, Priscilla Burity | From Pixels and Words to Waves: A Unified Framework for Spectral Dictionary vLLMs | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Vision-language models (VLMs) unify computer vision and natural language processing in a single architecture capable of interpreting and describing images. Most state-of-the-art systems rely on two computationally intensive components: convolutions in the vision encoder and quadratic self-attention for multimodal fusion. This work removes both by introducing a spectral dictionary token mixer, which represents each image patch or wordpiece as a sparse combination of learnable frequency atoms. Our 1.1B-parameter prototype, SDict-VLM, achieves BLEU-4 of 39.2, CIDEr of 127.5, and SPICE of 27.0 on MS-COCO captioning, along with 50.3 percent accuracy on VQAv2. These results close approximately 85 percent of the performance gap to BLIP-2 while using 60 percent fewer parameters, 2.3 times less peak GPU memory, and 2.2 times faster inference than PaLI-3. To our knowledge, this is the first VLM to eliminate both convolutions and self-attention while matching mid-scale transformer baselines. In addition to its O(L log L) complexity, the shared frequency dictionary enables transparent cross-modal alignment and offers a tunable trade-off between accuracy and compute, paving the way for efficient and interpretable VLMs. | [
{
"version": "v1",
"created": "Sun, 22 Jun 2025 23:48:02 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Kiruluta",
"Andrew",
""
],
[
"Burity",
"Priscilla",
""
]
] |
2506.18944 | Xue-Ping Wang | Yun-Mao Zhang, Xue-ping Wang | Characterizations of monotone right continuous functions which generate associative functions | 16. arXiv admin note: substantial text overlap with arXiv:2409.02941 | null | null | null | math.FA | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Associativity of a two-place function $T: [0,1]^2\rightarrow [0,1]$ defined by $T(x,y)=f^{(-1)}(T^*(f(x),f(y)))$ where $T^*:[0,1]^2\rightarrow[0,1]$ is an associative function with neutral element in $[0,1]$, $f: [0,1]\rightarrow [0,1]$ is a monotone right continuous function and $f^{(-1)}:[0,1]\rightarrow[0,1]$ is the pseudo-inverse of $f$ depends only on properties of the range of $f$. The necessary and sufficient conditions for the $T$ to be associative are presented by applying the properties of the monotone right continuous function $f$. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 00:34:04 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Zhang",
"Yun-Mao",
""
],
[
"Wang",
"Xue-ping",
""
]
] |
2506.18945 | Zihan Wang | Zihan Wang, Rui Pan, Jiarui Yao, Robert Csordas, Linjie Li, Lu Yin, Jiajun Wu, Tong Zhang, Manling Li, Shiwei Liu | Chain-of-Experts: Unlocking the Communication Power of Mixture-of-Experts Models | null | null | null | null | cs.LG cs.CL | http://creativecommons.org/licenses/by/4.0/ | We propose Chain-of-Experts (CoE), a new Mixture-of-Experts (MoE) architecture that introduces sequential expert communication within each layer. Unlike traditional MoE models, where experts operate independently in parallel, CoE processes tokens iteratively across a chain of experts inside a layer. To support dynamic expert selection across iterations, CoE employs a dedicated router at each iteration step within a layer. This design allows tokens to re-evaluate and select different experts during each iteration, rather than being statically assigned. As a result, CoE introduces a flexible routing mechanism that increases the diversity of expert combinations and enriches the model's representational capacity. CoE demonstrates improved performance under fixed compute: on math reasoning tasks, it reduces validation loss from 1.20 to 1.12 compared to a standard MoE. Beyond performance, CoE offers a new scaling axis: depth through expert iteration, which complements conventional width/depth scaling. For example, using 2x iterations matches the performance of 3x expert selections (in width), while reducing memory usage by 17.6-42% relative to other scaling strategies. Our analysis reveals that CoE's benefits stem from its iterative residual structure and enhanced expert specialization empowered by iterative routing, which together unlock more expressive representations. Code is available at https://github.com/ZihanWang314/coe. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 02:15:43 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Wang",
"Zihan",
""
],
[
"Pan",
"Rui",
""
],
[
"Yao",
"Jiarui",
""
],
[
"Csordas",
"Robert",
""
],
[
"Li",
"Linjie",
""
],
[
"Yin",
"Lu",
""
],
[
"Wu",
"Jiajun",
""
],
[
"Zhang",
"Tong",
""
],
[
"Li",
"Manling",
""
],
[
"Liu",
"Shiwei",
""
]
] |
2506.18946 | Zhe Dong | Zhe Dong, Yuzhe Sun, Tianzhu Liu, Yanfeng Gu | DiffRIS: Enhancing Referring Remote Sensing Image Segmentation with Pre-trained Text-to-Image Diffusion Models | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Referring remote sensing image segmentation (RRSIS) enables the precise delineation of regions within remote sensing imagery through natural language descriptions, serving critical applications in disaster response, urban development, and environmental monitoring. Despite recent advances, current approaches face significant challenges in processing aerial imagery due to complex object characteristics including scale variations, diverse orientations, and semantic ambiguities inherent to the overhead perspective. To address these limitations, we propose DiffRIS, a novel framework that harnesses the semantic understanding capabilities of pre-trained text-to-image diffusion models for enhanced cross-modal alignment in RRSIS tasks. Our framework introduces two key innovations: a context perception adapter (CP-adapter) that dynamically refines linguistic features through global context modeling and object-aware reasoning, and a progressive cross-modal reasoning decoder (PCMRD) that iteratively aligns textual descriptions with visual regions for precise segmentation. The CP-adapter bridges the domain gap between general vision-language understanding and remote sensing applications, while PCMRD enables fine-grained semantic alignment through multi-scale feature interaction. Comprehensive experiments on three benchmark datasets-RRSIS-D, RefSegRS, and RISBench-demonstrate that DiffRIS consistently outperforms existing methods across all standard metrics, establishing a new state-of-the-art for RRSIS tasks. The significant performance improvements validate the effectiveness of leveraging pre-trained diffusion models for remote sensing applications through our proposed adaptive framework. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 02:38:56 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Dong",
"Zhe",
""
],
[
"Sun",
"Yuzhe",
""
],
[
"Liu",
"Tianzhu",
""
],
[
"Gu",
"Yanfeng",
""
]
] |
2506.18947 | Alper Karakas | Alper Deniz Karakas | The Persistent Effects of Peru's Mining MITA: Double Machine Learning Approach | null | null | null | null | econ.EM | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This study examines the long-term economic impact of the colonial Mita system in Peru, building on Melissa Dell's foundational work on the enduring effects of forced labor institutions. The Mita, imposed by the Spanish colonial authorities from 1573 to 1812, required indigenous communities within a designated boundary to supply labor to mines, primarily near Potosi. Dell's original regression discontinuity design (RDD) analysis, leveraging the Mita boundary to estimate the Mita's legacy on modern economic outcomes, indicates that regions subjected to the Mita exhibit lower household consumption levels and higher rates of child stunting. In this paper, I replicate Dell's results and extend this analysis. I apply Double Machine Learning (DML) methods--the Partially Linear Regression (PLR) model and the Interactive Regression Model (IRM)--to further investigate the Mita's effects. DML allows for the inclusion of high-dimensional covariates and enables more flexible, non-linear modeling of treatment effects, potentially capturing complex relationships that a polynomial-based approach may overlook. While the PLR model provides some additional flexibility, the IRM model allows for fully heterogeneous treatment effects, offering a nuanced perspective on the Mita's impact across regions and district characteristics. My findings suggest that the Mita's economic legacy is more substantial and spatially heterogeneous than originally estimated. The IRM results reveal that proximity to Potosi and other district-specific factors intensify the Mita's adverse impact, suggesting a deeper persistence of regional economic inequality. These findings underscore that machine learning addresses the realistic non-linearity present in complex, real-world systems. By modeling hypothetical counterfactuals more accurately, DML enhances my ability to estimate the true causal impact of historical interventions. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 04:11:41 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Karakas",
"Alper Deniz",
""
]
] |
2506.18948 | Dakang Cen | Dakang Cen, Zhiyuan Li, Wenlong Zhang | Numerical analysis of scattered point measurement-based regularization for backward problems for fractional wave equations | 27 pages. arXiv admin note: text overlap with arXiv:2506.17575 | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by/4.0/ | In this work, our aim is to reconstruct the unknown initial value from terminal data. We develop a numerical framework on nonuniform time grids for fractional wave equations under the lower regularity assumptions. Then, we introduce a regularization method that effectively handles scattered point measurements contaminated with stochastic noise. The optimal error estimates of stochastic convergence not only balance discretization errors, the noise, and the number of observation points, but also propose an a priori choice of regularization parameters. Finally, several numerical experiments are presented to demonstrate the efficiency and accuracy of the algorithm. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 05:37:39 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Cen",
"Dakang",
""
],
[
"Li",
"Zhiyuan",
""
],
[
"Zhang",
"Wenlong",
""
]
] |
2506.18949 | Mrinnoy Gohain Dr. | Kalyan Malakar, Mrinnoy M Gohain and Kalyan Bhuyan | Gravitational Baryogenesis Constraints on Nojiri-Odintsov $f(R)$ Gravity | 20 pages; 8 figures and 6 tables | null | null | null | gr-qc | http://creativecommons.org/licenses/by/4.0/ | This paper focuses on exploring the imbalance between matter and antimatter via the gravitational baryogenesis mechanism within the framework of Nojiri-Odintsov $f(R)$ gravity models in a spatially flat FLRW universe, where $R$ is the Ricci curvature scalar. This mechanism is based on \(\mathcal{CP}\)-violation generated via introducing an interaction between baryonic matter current (\(J^\mu\)) and the derivative of curvature scalar (\(\partial_\mu R\)), which finally results in the baryon asymmetry. We examine three distinct $f(R)$ models: (i) \(f(R)=\frac{-\alpha}{R^m}+\frac{R}{2k^2}+\beta R^2\), (ii) \(f(R)=\alpha R^{\gamma} +\beta R^{\delta}\) and (iii) \(f(R)=\frac{\alpha R^{2m}-\beta R^m}{1+\gamma R^m} \). We demonstrate that even during the matter-dominated epoch of the universe, the $f(R)$ gravity models under consideration can yield a non-vanishing matter-antimatter asymmetry. We constrain the model parameters of each model within the scenario of gravitational baryogenesis to produce the observed value of the baryon-to-entropy ratio, providing a continuous set of acceptable physical values for those parameters. Our results show highly consistent agreement to the baryon-to-entropy ratio with the observational bound from the cosmological data. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 07:22:13 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Malakar",
"Kalyan",
""
],
[
"Gohain",
"Mrinnoy M",
""
],
[
"Bhuyan",
"Kalyan",
""
]
] |
2506.18950 | Maoyuan Li | Maoyuan Li, Sihong Li, Guancheng Shen, Yun Zhang, Huamin Zhou | Online high-precision prediction method for injection molding product weight by integrating time series/non-time series mixed features and feature attention mechanism | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | To address the challenges of untimely detection and online monitoring lag in injection molding quality anomalies, this study proposes a mixed feature attention-artificial neural network (MFA-ANN) model for high-precision online prediction of product weight. By integrating mechanism-based with data-driven analysis, the proposed architecture decouples time series data (e.g., melt flow dynamics, thermal profiles) from non-time series data (e.g., mold features, pressure settings), enabling hierarchical feature extraction. A self-attention mechanism is strategically embedded during cross-domain feature fusion to dynamically calibrate inter-modality feature weights, thereby emphasizing critical determinants of weight variability. The results demonstrate that the MFA-ANN model achieves a RMSE of 0.0281 with 0.5 g weight fluctuation tolerance, outperforming conventional benchmarks: a 25.1% accuracy improvement over non-time series ANN models, 23.0% over LSTM networks, 25.7% over SVR, and 15.6% over RF models, respectively. Ablation studies quantitatively validate the synergistic enhancement derived from the integration of mixed feature modeling (contributing 22.4%) and the attention mechanism (contributing 11.2%), significantly enhancing the model's adaptability to varying working conditions and its resistance to noise. Moreover, critical sensitivity analyses further reveal that data resolution significantly impacts prediction reliability, low-fidelity sensor inputs degrade performance by 23.8% RMSE compared to high-precision measurements. Overall, this study provides an efficient and reliable solution for the intelligent quality control of injection molding processes. | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 08:40:50 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Li",
"Maoyuan",
""
],
[
"Li",
"Sihong",
""
],
[
"Shen",
"Guancheng",
""
],
[
"Zhang",
"Yun",
""
],
[
"Zhou",
"Huamin",
""
]
] |
2506.18951 | Jinyang Li | Jinyang Li, Xiaolong Li, Ge Qu, Per Jacobsson, Bowen Qin, Binyuan Hui, Shuzheng Si, Nan Huo, Xiaohan Xu, Yue Zhang, Ziwei Tang, Yuanshuai Li, Florensia Widjaja, Xintong Zhu, Feige Zhou, Yongfeng Huang, Yannis Papakonstantinou, Fatma Ozcan, Chenhao Ma, Reynold Cheng | SWE-SQL: Illuminating LLM Pathways to Solve User SQL Issues in Real-World Applications | 26 pages, 9 figures | null | null | null | cs.DB cs.AI | http://creativecommons.org/licenses/by-sa/4.0/ | Resolution of complex SQL issues persists as a significant bottleneck in real-world database applications. Current Large Language Models (LLMs), while adept at text-to-SQL translation, have not been rigorously evaluated on the more challenging task of debugging SQL issues. To address this gap, we introduce BIRD-CRITIC, a new SQL issue debugging benchmark comprising 530 PostgreSQL tasks (BIRD-CRITIC-PG) and 570 multi-dialect tasks (BIRD-CRITIC-Multi), distilled from authentic user issues and replayed within new environments to facilitate rigorous evaluation. Baseline evaluations underscore the task's complexity, with the leading reasoning model O3-Mini achieving only 38.87% success rate on BIRD-CRITIC-PG and 33.33% on BIRD-CRITIC-Multi. Meanwhile, advancing open-source models for database tasks is crucial for empowering local development while safeguarding data privacy. Therefore, we present Six-Gym (Sql-fIX-Gym), a training environment for elevating open-source model capabilities for SQL issue debugging. This environment leverages SQL-Rewind strategy, which automatically generates executable issue-solution datasets by reverse-engineering issues from verified SQLs. However, popular trajectory-based fine-tuning methods do not explore substantial supervisory signals. We further propose f-Plan Boosting, which extracts high-level debugging plans from SQL solutions, enabling teacher LLMs to produce 73.7% more successful trajectories for training. We integrate these components into an open-source agent, Bird-Fixer. Based on Qwen-2.5-Coder-14B, Bird-Fixer achieves 38.11% success rate on BIRD-CRITIC-PG and 29.65% on BIRD-CRITIC-Multi, surpassing leading proprietary models such as Claude-3.7-Sonnet and GPT-4.1, marking a significant step toward democratizing sophisticated SQL-debugging capabilities. The leaderboard and source code are available: https://bird-critic.github.io/ | [
{
"version": "v1",
"created": "Mon, 23 Jun 2025 09:41:37 GMT"
}
] | 2025-06-25T00:00:00 | [
[
"Li",
"Jinyang",
""
],
[
"Li",
"Xiaolong",
""
],
[
"Qu",
"Ge",
""
],
[
"Jacobsson",
"Per",
""
],
[
"Qin",
"Bowen",
""
],
[
"Hui",
"Binyuan",
""
],
[
"Si",
"Shuzheng",
""
],
[
"Huo",
"Nan",
""
],
[
"Xu",
"Xiaohan",
""
],
[
"Zhang",
"Yue",
""
],
[
"Tang",
"Ziwei",
""
],
[
"Li",
"Yuanshuai",
""
],
[
"Widjaja",
"Florensia",
""
],
[
"Zhu",
"Xintong",
""
],
[
"Zhou",
"Feige",
""
],
[
"Huang",
"Yongfeng",
""
],
[
"Papakonstantinou",
"Yannis",
""
],
[
"Ozcan",
"Fatma",
""
],
[
"Ma",
"Chenhao",
""
],
[
"Cheng",
"Reynold",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.