new

Get trending papers in your email inbox once a day!

Get trending papers in your email inbox!

Daily Papers

byAK and the research community

May 19

Submitted by

hongyi-liu

SkillsVote: Lifecycle Governance of Agent Skills from Collection, Recommendation to Evolution

IAAR-Shanghai

Memtensor Research Group

Submitted by

AaronHuangWei

LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generation

nvidia

Submitted by

CoreloneH

Lance: Unified Multimodal Modeling by Multi-Task Synergy

bytedance-research

bytedance-research

Submitted by

ldkong

AI for Auto-Research: Roadmap & User Guide

·
20 authors

Submitted by

B3rrYang

Code-as-Room: Generating 3D Rooms from Top-Down View Images via Agentic Code Synthesis

Submitted by

kkakkkka

KVPO: ODE-Native GRPO for Autoregressive Video Alignment via KV Semantic Exploration

THU1911

Tsinghua University

Submitted by

taesiri

Code as Agent Harness

·
42 authors

Submitted by

CheeryLJH

OProver: A Unified Framework for Agentic Formal Theorem Proving

m-a-p

Multimodal Art Projection

Submitted by

weirayao

CHI-Bench: Can AI Agents Automate End-to-End, Long-Horizon, Policy-Rich Healthcare Workflows?

actava

Submitted by

XingtaiHF

Post-Trained MoE Can Skip Half Experts via Self-Distillation

THU1911

Tsinghua University

Submitted by

jjihwannn

LiteFrame: Efficient Vision Encoders Unlock Frame Scaling in Video LLMs

deepmind

Submitted by

ZhishanQ

Stop When Reasoning Converges: Semantic-Preserving Early Exit for Reasoning Models

University of Illinois Chicago

Submitted by

monster119120

Measuring Maximum Activations in Open Large Language Models

baidu

Submitted by

yfdeng10

StableVLA: Towards Robust Vision-Language-Action Models without Extra Data

PekingUniversity

Peking University

Submitted by

monster119120

EndPrompt: Efficient Long-Context Extension via Terminal Anchoring

baidu

Submitted by

youuor7r

Where Should Diffusion Enter a Language Model? Geometry-Guided Hidden-State Replacement

HOLILAB

1

Submitted by

yizecheng

Model-Adaptive Tool Necessity Reveals the Knowing-Doing Gap in LLM Tool Use

UMaryland

University of Maryland

Submitted by

jiwonsong

CompactAttention: Accelerating Chunked Prefill with Block-Union KV Selection

SNU-VLSI

Seoul National University VLSI Lab

Submitted by

iforgott

From Runnable to Shippable: Multi-Agent Test-Driven Development for Generating Full-Stack Web Applications from Requirements

CUHK-CSE

The Chinese University of Hong Kong

Submitted by

emozilla

Targeted Neuron Modulation via Contrastive Pair Search

NousResearch

Submitted by

taesiri

AtlasVA: Self-Evolving Visual Skill Memory for Teacher-Free VLM Agents

·
6 authors

Submitted by

KOHbDS

FINESSE-Bench: A Hierarchical Benchmark Suite for Financial Domain Knowledge and Technical Analysis in Large Language Models

·
7 authors

Submitted by

milkkarten

Agent Bazaar: Enabling Economic Alignment in Multi-Agent Marketplaces

princetonu

Princeton University

2

Submitted by

Automationyw

NGM: A Plug-and-Play Training-Free Memory Module for LLMs

nanjinguniv

Nanjing University

Submitted by

taesiri

Actionable World Representation

·
7 authors

Submitted by

ItsMaxNorm

SafeDiffusion-R1: Online Reward Steering for Safe Diffusion Post-Training

MBZUAI

Mohamed Bin Zayed University of Artificial Intelligence

Submitted by

QingchuanMa

A2RBench: An Automatic Paradigm for Formally Verifiable Abstract Reasoning Benchmark Generation

MAC-AutoML

Submitted by

Automationyw

TOBench: A Task-Oriented Omni-Modal Benchmark for Real-World Tool-Using Agents

AI-Safeguard

Submitted by

haizhongzheng

AstraFlow: Dataflow-Oriented Reinforcement Learning for Agentic LLMs

Carnegie Mellon University

Submitted by

SivanSX

Evaluating Cognitive Age Alignment in Interactive AI Agents

PediaMedAI

Submitted by

hhua2

MementoGUI: Learning Agentic Multimodal Memory Control for Long-Horizon GUI Agents

·
6 authors

Submitted by

taesiri

Incantation: Natural Language as the Action Interface for Multi-Entity Video World Models

·
14 authors

Submitted by

ligongh

SNLP: Layer-Parallel Inference via Structured Newton Corrections

RedHatAI

Submitted by

juezhi

E-PMQ: Expert-Guided Post-Merge Quantization with Merged-Weight Anchoring

PolyUHK

The Hong Kong Polytechnic University

Submitted by

Jerry999

MixSD: Mixed Contextual Self-Distillation for Knowledge Injection

CarnegieMellonU

Carnegie Mellon University

Submitted by

pcr2120

Geometric Phase Transition Enables Extreme Hippocampal Memory Capacity

·
1 authors

Submitted by

gaotiexinqu

VideoSeeker: Incentivizing Instance-level Video Understanding via Native Agentic Tool Invocation

·
14 authors

Submitted by

danielqing99

Auditing Multimodal LLM Raters: Central Tendency Bias in Clinical Ordinal Scoring

ufl

University of Florida

Submitted by

mchraba

Monitoring the Internal Monologue: Probe Trajectories Reveal Reasoning Dynamics

·
5 authors

1

Submitted by

taesiri

AgentKernelArena: Generalization-Aware Benchmarking of GPU Kernel Optimization Agents

·
14 authors

Submitted by

arkimjh

GRASP: Learning to Ground Social Reasoning in Multi-Person Non-Verbal Interactions

UIUC-CS

University of Illinois at Urbana-Champaign

1

Submitted by

zarazetlin

TopoPrimer: The Missing Topological Context in Forecasting Models

apple

1

Submitted by

you2who

AR-VLA: True Autoregressive Action Expert for Vision-Language-Action Models

INSAIT-Institute

Institute for Computer Science, Artificial intelligence and Technology