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

Submitted by

taesiri

Gamma-World: Generative Multi-Agent World Modeling Beyond Two Players

nvidia

Submitted by

Jiafei1224

MolmoAct2: Action Reasoning Models for Real-world Deployment

allenai

Submitted by

imone

HRM-Text: Efficient Pretraining Beyond Scaling

sapientinc

Submitted by

zr-wang

CiteVQA: Benchmarking Evidence Attribution for Trustworthy Document Intelligence

opendatalab

Submitted by

taesiri

SkillOpt: Executive Strategy for Self-Evolving Agent Skills

MicrosoftResearch

Microsoft Research

Submitted by

StableKirito

Mean Mode Screaming: Mean--Variance Split Residuals for 1000-Layer Diffusion Transformers

·
1 authors

Submitted by

taesiri

Code as Agent Harness

·
42 authors

Submitted by

anchen1011

MinT: Managed Infrastructure for Training and Serving Millions of LLMs

mindlab-research

Submitted by

Violet24K

Heterogeneous Scientific Foundation Model Collaboration

illinois

University of Illinois at Urbana-Champaign

Submitted by

Cardlnal

DelTA: Discriminative Token Credit Assignment for Reinforcement Learning from Verifiable Rewards

·
3 authors

Submitted by

floyed

Anti-Self-Distillation for Reasoning RL via Pointwise Mutual Information

rednote-hilab

Submitted by

Paranioar

SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecture

sensenova

Submitted by

taesiri

AutoResearchClaw: Self-Reinforcing Autonomous Research with Human-AI Collaboration

·
35 authors

Submitted by

xiaochonglinghu

TransitLM: A Large-Scale Dataset and Benchmark for Map-Free Transit Route Generation

GD-ML

Submitted by

Ukpkmkkk

Perception or Prejudice: Can MLLMs Go Beyond First Impressions of Personality?

utokyo-ai

The University of Tokyo

Submitted by

ssz1111

From Context to Skills: Can Language Models Learn from Context Skillfully?

·
13 authors

Submitted by

yaful

Achieving Gold-Medal-Level Olympiad Reasoning via Simple and Unified Scaling

·
28 authors

Submitted by

DarthZhu

When Vision Speaks for Sound

ucdavis

Submitted by

shenqiorient

AgentDoG 1.5: A Lightweight and Scalable Alignment Framework for AI Agent Safety and Security

Submitted by

apocryphal

MemPrivacy: Privacy-Preserving Personalized Memory Management for Edge-Cloud Agents

MemTensor

Submitted by

xwm

Video2GUI: Synthesizing Large-Scale Interaction Trajectories for Generalized GUI Agent Pretraining

PekingUniversity

Peking University

Submitted by

taesiri

Qwen-VLA: Unifying Vision-Language-Action Modeling across Tasks, Environments, and Robot Embodiments

Qwen

Submitted by

taesiri

LocateAnything: Fast and High-Quality Vision-Language Grounding with Parallel Box Decoding

nvidia

5

Submitted by

LiamLian0727

PhysBrain 1.0 Technical Report

DeepCybo

Submitted by

RuofengYang

ARIS: Autonomous Research via Adversarial Multi-Agent Collaboration

SJTU

Shanghai Jiao Tong University

Submitted by

EilamSha

MulTaBench: Benchmarking Multimodal Tabular Learning with Text and Image

Technion

Technion Israel institute of technology

Submitted by

Nothing2Say

DVAO: Dynamic Variance-adaptive Advantage Optimization for Multi-reward Reinforcement Learning

·
6 authors

4

Submitted by

filicos

Mega-ASR: Towards In-the-wild^2 Speech Recognition via Scaling up Real-world Acoustic Simulation

NationalUniversityofSingapore

National University of Singapore

Submitted by

taesiri

δ-mem: Efficient Online Memory for Large Language Models

mindlab-research

Submitted by

hongyi-liu

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

IAAR-Shanghai

Memtensor Research Group

Submitted by

CoreloneH

Stream-R1: Reliability-Perplexity Aware Reward Distillation for Streaming Video Generation

FrameXAI

Submitted by

huiwon

RLDX-1 Technical Report

RLWRLD

Submitted by

ZhuofengLi

Beyond Semantic Similarity: Rethinking Retrieval for Agentic Search via Direct Corpus Interaction

TIGER-Lab

Submitted by

zhangkangning

MMSkills: Towards Multimodal Skills for General Visual Agents

ShanghaiJiaotongUniversity

Shanghai Jiaotong University 1(NOT OFFICIAL)

Submitted by

taesiri

Self-Distilled Agentic Reinforcement Learning

·
11 authors

Submitted by

lhjiang

Qwen-Image-2.0 Technical Report

Qwen

Submitted by

AaronHuangWei

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

nvidia

Submitted by

taesiri

Skill1: Unified Evolution of Skill-Augmented Agents via Reinforcement Learning

·
9 authors

Submitted by

Met4physics

Rethinking Cross-Layer Information Routing in Diffusion Transformers

RTP-LLM

6

Submitted by

Jinjing713

Lens: Rethinking Training Efficiency for Foundational Text-to-Image Models

microsoft

Submitted by

CoreloneH

Stream-T1: Test-Time Scaling for Streaming Video Generation

FrameXAI

Submitted by

zzzhr97

π-Bench: Evaluating Proactive Personal Assistant Agents in Long-Horizon Workflows

Simplified-Reasoning

Simplified Reasoning

Submitted by

csfufu

OpenSearch-VL: An Open Recipe for Frontier Multimodal Search Agents

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

Kaining

WBench: A Comprehensive Multi-turn Benchmark for Interactive Video World Model Evaluation

meituan-longcat

Submitted by

taesiri

AnyFlow: Any-Step Video Diffusion Model with On-Policy Flow Map Distillation

nvidia

Submitted by

Osilly

Flow-OPD: On-Policy Distillation for Flow Matching Models

·
11 authors

Submitted by

zykRichard

Full Attention Strikes Back: Transferring Full Attention into Sparse within Hundred Training Steps

RTP-LLM

1

Submitted by

lucianodelcorro

Active Learners as Efficient PRP Rerankers

UdeSA

Universidad de San Andrés

Submitted by

zhuhz22

Causal Forcing++: Scalable Few-Step Autoregressive Diffusion Distillation for Real-Time Interactive Video Generation

thu-ml

Tsinghua Machine Learning Group

3

Submitted by

taesiri

Agent Explorative Policy Optimization for Multimodal Agentic Reasoning

nvidia

4

Submitted by

xiaochonglinghu

Enhancing Train-Free Infinite-Frame Generation for Consistent Long Videos

AGI-LAB-HF

Submitted by

taesiri

Visual Generation in the New Era: An Evolution from Atomic Mapping to Agentic World Modeling

lmms-lab

Submitted by

HaoyiZhu

SANA-WM: Efficient Minute-Scale World Modeling with Hybrid Linear Diffusion Transformer

nvidia

Submitted by

ZhaoweiWang

Training Long-Context Vision-Language Models Effectively with Generalization Beyond 128K Context

ByteDance-Seed

3

Submitted by

Mithas-01

ProRL: Effective Reinforcement Learning for Proactive Recommendation via Rectified Policy Gradient Estimation

Fudan-University

Fudan University

Submitted by

taesiri

OpenComputer: Verifiable Software Worlds for Computer-Use Agents

yale-nlp

Submitted by

taesiri

UniVidX: A Unified Multimodal Framework for Versatile Video Generation via Diffusion Priors

·
11 authors

Submitted by

ZhenlongYuan

IndusAgent: Reinforcing Open-Vocabulary Industrial Anomaly Detection with Agentic Tools

ict-cas

Institute of Computing Technology, Chinese Academy of Sciences

1

Submitted by

taesiri

Continuous Latent Diffusion Language Model

ByteDance-Seed

Submitted by

anchen1011

Macaron-A2UI: A Model for Generative UI in Personal Agents

mindlab-research

5

Submitted by

Bang-UdeM-Mila

Foundation Protocol: A Coordination Layer for Agentic Society

University of Montreal

Submitted by

amphora

Soohak: A Mathematician-Curated Benchmark for Evaluating Research-level Math Capabilities of LLMs

EleutherAI

3

Submitted by

EddieYang428

EvalVerse: Pipeline-Aware and Expert-Calibrated Benchmarking for Professional Cinematic Video Generation

tencent

3

Submitted by

Yirany

MiniCPM-o 4.5: Towards Real-Time Full-Duplex Omni-Modal Interaction

openbmb

Submitted by

jinheon

OmniRetrieval: Unified Retrieval across Heterogeneous Knowledge Sources

kaist-ai

Submitted by

CoreloneH

Lance: Unified Multimodal Modeling by Multi-Task Synergy

bytedance-research

bytedance-research

Submitted by

ZhaoweiWang

MemLens: Benchmarking Multimodal Long-Term Memory in Large Vision-Language Models

nvidia

Submitted by

gaotang

RubricEM: Meta-RL with Rubric-guided Policy Decomposition beyond Verifiable Rewards

google

3

Submitted by

Paranioar

From Pixels to Words -- Towards Native One-Vision Models at Scale

·
21 authors

3

Submitted by

LMD0311

HERMES++: Toward a Unified Driving World Model for 3D Scene Understanding and Generation

H-EmbodVis

Submitted by

taesiri

OpenSeeker-v2: Pushing the Limits of Search Agents with Informative and High-Difficulty Trajectories

·
7 authors

Submitted by

lifuguan

SpatialBench: Is Your Spatial Foundation Model an All-Round Player?

ropedia-ai

Submitted by

taesiri

CollabVR: Collaborative Video Reasoning with Vision-Language and Video Generation Models

kaist-ai

Submitted by

ldkong

AI for Auto-Research: Roadmap & User Guide

·
20 authors

Submitted by

sinwang

World Action Models: The Next Frontier in Embodied AI

OpenMOSS-Team

Submitted by

ChengsongHuang

LLMs Improving LLMs: Agentic Discovery for Test-Time Scaling

google

Submitted by

yunqu

Listwise Policy Optimization: Group-based RLVR as Target-Projection on the LLM Response Simplex

Tencent-Hunyuan

Tencent Hunyuan

3

Submitted by

iieycx

Co-Evolving Policy Distillation

·
10 authors

3

Submitted by

DukeShen

FashionChameleon: Towards Real-Time and Interactive Human-Garment Video Customization

AGI-LAB-HF

Submitted by

seawolf2357

Darwin Family: MRI-Trust-Weighted Evolutionary Merging for Training-Free Scaling of Language-Model Reasoning

FINAL-Bench

Submitted by

Abyssaledge

MobileGym: A Verifiable and Highly Parallel Simulation Platform for Mobile GUI Agent Research

·
11 authors

Submitted by

Zhongzhu

OSCAR: Offline Spectral Covariance-Aware Rotation for 2-bit KV Cache Quantization

togethercomputer

Submitted by

patricebechard

Do Enterprise Systems Need Learned World Models? The Importance of Context to Infer Dynamics

ServiceNow-AI

1

Submitted by

DarkBluee

MemEye: A Visual-Centric Evaluation Framework for Multimodal Agent Memory

·
17 authors

Submitted by

hkuzxc

HyperEyes: Dual-Grained Efficiency-Aware Reinforcement Learning for Parallel Multimodal Search Agents

XiaohongshuAI

Submitted by

Xkev

Self-Improving Language Models with Bidirectional Evolutionary Search

Harvard

Harvard University

Submitted by

jamesliu1217

CollectionLoRA: Collecting 50 Effects in 1 LoRA via Multi-Teacher On-Policy Distillation

·
10 authors

Submitted by

lhjiang

Qwen-Image-VAE-2.0 Technical Report

Qwen

Submitted by

groundhogLLM

ACC: Compiling Agent Trajectories for Long-Context Training

ustc-community

University of Science and Technology of China

1

Submitted by

chanhee-luke

Why Far Looks Up: Probing Spatial Representation in Vision-Language Models

nvidia

Submitted by

Ningyu

SciAtlas: A Large-Scale Knowledge Graph for Automated Scientific Research

University College London CS

Submitted by

caiyuchen

Learning to Foresee: Unveiling the Unlocking Efficiency of On-Policy Distillation

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

DhavalPatel

MCP-Cosmos: World Model-Augmented Agents for Complex Task Execution in MCP Environments

ibm-research

2

Submitted by

taesiri

minWM: A Full-Stack Open-Source Framework for Real-Time Interactive Video World Models

·
12 authors

Submitted by

Suu

GoLongRL: Capability-Oriented Long Context Reinforcement Learning with Multitask Alignment

·
12 authors

Submitted by

Alllann

Leveraging Verifier-Based Reinforcement Learning in Image Editing

ByteDance-Seed

Submitted by

MindscapeRAG

MiA-Signature: Approximating Global Activation for Long-Context Understanding

tencent

Submitted by

yulunliu

YoCausal: How Far is Video Generation from World Model? A Causality Perspective

AlayaLab

Submitted by

Ziqi

PhysX-Omni: Unified Simulation-Ready Physical 3D Generation for Rigid, Deformable, and Articulated Objects

·
8 authors

Submitted by

weirayao

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

actava

Submitted by

Abyssaledge

DexJoCo: A Benchmark and Toolkit for Task-Oriented Dexterous Manipulation on MuJoCo

·
14 authors

Submitted by

LCZZZZ

Auditing Agent Harness Safety

ucsbai

Submitted by

taesiri

HumanNet: Scaling Human-centric Video Learning to One Million Hours

·
2 authors

Submitted by

jinyuan222

Process Rewards with Learned Reliability

HINT-lab

Huang's INTelligence lab

Submitted by

wyy-code

Geometry Conflict: Explaining and Controlling Forgetting in LLM Continual Post-Training

PolyUHK

The Hong Kong Polytechnic University