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

Oct 14

Submitted by

LIQIIIII

Jeva: System One Models from First Principles

NationalUniversityofSingapore

National University of Singapore

Submitted by

dongguanting

Agent-Reflex: Teaching Language Agents to Act Reflectively

RUC-NLPIR

NLPIR Lab @ RUC

Submitted by

ZhuYuhan

VideoJev: System One Model for Video Understanding and Interaction

MCG-NJU

Multimedia Computing Group-Nanjing University

Submitted by

vvangfaye

Urbex: Agentic Spatial Grounding for City-Scale 3D Scenes

alibabagroup

Submitted by

haksorus

GreenWAM: A World Foundation Model for Robotics

SberRoboticsCenter

Sber Robotics Center

Submitted by

gaoxin492

AesCode: Aesthetic Code Generation with Decoupled Cross-Modal Rewards

microsoft

Submitted by

ruiyaox

Harness^2: Recursive Agent Harnessing for an Open World

google

Submitted by

astrid01052

TPPO: Tracking-Value PPO For Language Model Reasoning

·
9 authors

Submitted by

gordonhu

Tev: Scaling Long-Horizon Agentic Environments with Dense Progress Rewards

·
6 authors

Submitted by

YangXiao-nlp

Vela: User Behavior as Supervision Sparks LLM Intelligence

AllSpark-Research

AllSpark Research

Submitted by

ZhenlongYuan

X-Ragent: Tool-Augmented Reinforcement Learning for Evidence-Grounded Chest X-Ray Reasoning

CarnegieMellonU

Carnegie Mellon University

Submitted by

fxmeng

DoSA: Dots Sparse Attention

dots-studio

Submitted by

Benchanything

WorkRover: Scaling Long-Horizon Cowork Agent Training Via Recursive Environment Evolution

nex-agi

Submitted by

HelenPeng

ExeWorldBench: Benchmarking Code-Based Executable 3D World Modeling

ustc-community

University of Science and Technology of China

1

Submitted by

MinghaoFu

LeWAM: End-to-End World and Action Modeling with JEPAs

Submitted by

jwhe

Scoring Higher, Answering Worse: Mitigating Reward Hacking in Rubric-Based RL via Protocol-Level Rubrics

ByteDance

Submitted by

NinaKarine

Multics: Generative Multi-Player World Model with Hierarchical Communication

TencentARC

ARC Lab, Tencent

Submitted by

affjljoo3581

UltraBERT: Beyond the Limit of BERT Pre-Training

·
3 authors

1

Submitted by

Zhengbo-Zhang

EviScene: Beyond "Is It AI?" for Agentic Image Factuality Assessment

·
13 authors

Submitted by

vkvermaa

SALD: Self-Referenced Advantage Learning for Diffusion Models

IITK

Indian Institute of Technology, Kanpur

Submitted by

daviddink

APEX: Asynchronous Prefix Denoising for Extreme Compression

Lightricks

Submitted by

Nova2001

AReA: Adaptive Receptive-Field Attention for Efficient High-Resolution Diffusion Transformers

uoft-cs

University of Toronto Computer Science

Submitted by

txxnrd

Learning How to Remember: Memory Modality Selection for GUI Agents

Korea University

Submitted by

GSerussi

VEDJE: Video-Efficient Discriminative Joint Encoder for Scalable Video-Text Retrieval

Insight-bgu

Submitted by

wutianyi86

Last Secure Code Benchmark: Is the Code You Vibe Secure?

Qwen

Submitted by

ruchangyao

FlashVSR+: Efficient Video Super-Resolution via Hybrid Attention and Reward-Yield Distribution Matching

MMLab-CUHK

Submitted by

jiazhengli

Self-Distillation with Verbal Regularization

KingsCollegeLondon

King's College London

Submitted by

LulaCola

DeltaTTG: Training Multi-Turn Agents with a Token-to-Go Progress Meter

NanyangTechnologicalUniversity

Nanyang Technological University

Submitted by

SongzeLi

Data-Centric Recursive Self-Improvement for Navigation via Policy-Synchronized Language-Action Dual Bootstrapping

OpenGVLab

Submitted by

yulunliu

Cepheid: Decoupling Scale from Structure for Cross-Family Motion Completion

NYCU

National Yang Ming Chiao Tung University

Submitted by

itaowe

A Decade in a Day: Synthesizing Agent-Native Interfaces from Human-Oriented Software Evolution

tencent

Submitted by

KiyotakaWang

WAVE-0: Bridging VLA Policies and WAMs with Lightweight Latent Future Prediction

·
7 authors

Submitted by

hhua2

VIEWCRAFT: Evolving What Multimodal Agents See and How They Read It

·
10 authors

Submitted by

YunsongZhou

EgoExplorer: A Vision-Motion-Touch Observatory for Human Loco-Manipulation In Situ

InternRobotics

Intern Robotics

Submitted by

UncleWang233

CineRecon: Benchmarking Frontier Models for Agentic Cinematic Reconstruction

hkuhk

The University of Hong Kong

Submitted by

animawang

ClawLoop: Self-Improving Agents via Harness-Native Training Loop

meituan-longcat

Submitted by

Yeobin

GeoFix: Geometry Fixing for 4D Consistent Generation

·
10 authors

1

Submitted by

KaituoFeng

ReTerm: Recovery Training for Self-Improving Terminal Agents

·
6 authors

Submitted by

marinero4972

Open-o3-Video-v2: Seek, Ground, and Reason with Verifiable Spatio-Temporal Evidence

PKU

Peking University

Submitted by

qingzhengwang

Fish Audio Drama 3 Technical Report

fishaudio

1

Submitted by

SlowWorker

JevGPT: Thinking Fast and Slow in Autonomous Research

·
12 authors

Submitted by

yasNing

LifeSearch: Benchmarking Deep Research Agents for Real-World Life and Productivity

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

vinesmsuic

PhyProbe: Rethinking Physical Consistency Evaluation in Generated Videos

nvidia

Submitted by

jasonren

Elephant in the Fridge: Constant-Memory Frame Packing for Long Video Understanding

UNC-ChapelHill

University of North Carolina at Chapel Hill

Submitted by

Skywalker0410

LongR: Unleashing Long-Context Reasoning via Reinforcement Learning with Dense Utility Rewards

PKU

Peking University

1

Submitted by

hyulee

EMHO: EMbodied Agent Harness Optimization via Experience Traces

·
6 authors

Submitted by

hchoi256

Latent World Action Model for GUI Agents

kaist-ai

Submitted by

Dubhe-zmc

WorldRecon: Generative Multi-View Mesh Reconstruction for Large-Scale 3D World

Zhejiang University

Submitted by

AniSundar18

Pristine Models

uw-madison

University of Wisconsin - Madison

Submitted by

simonycl

AutoEnvScaling: Automating the Data Flywheel with Terminal Agents

microsoft

Submitted by

xishushu

UniForge3D: Joint 3D Reconstruction and Generation in a Shared Anchor-Guided Space

PekingUniversity

Peking University

Submitted by

alphadl

What Really Works for Universal Multimodal Embeddings? A Controlled Study of Contrastive Training Strategies

alibaba-multimodal-industrial-ai

1688 multimodal & industrial AI

Submitted by

dxlong2000

SWE-MISBEHAVIOR: Benchmarking and Mitigating Boundary Violations in Software Coding Agents

NationalUniversityofSingapore

National University of Singapore

Submitted by

haoyuw

Rubrics from the Task, Verdicts from Evidence: Pointwise Rubric Rewards with Tool-Augmented Judges

Submitted by

changyeon

ACCRUE: Continual Off-Policy Reinforcement Learning for Robot Foundation Models

nvidia

1

Submitted by

Iman998

HexaBee-4B: Six-Language Translation with Multilingual Answering and Bidirectional Supervision

·
3 authors

Submitted by

pcczc15

PsyBer-Agent: A Psychology-Driven User Agent for Recommender Systems

UoLiverpool

University of Liverpool

Submitted by

timcuhk

CIViC: Evidence Gates for Audio-Conditioned Visual Prediction on Proxy Tasks

NRC-CNRC

National Research Council Canada

Submitted by

TimothyCzp

Spectral-Driven Adaptive-Rank Optimizer for Large Language Model Reinforcement Learning with Verifiable Rewards

RUC-AIBOX

Submitted by

jaehong31

Agent Sociodrama: Social-Dynamics-Guided Persona Recasting for LLM Agent Teams

nanyang-technological-university-singapore

Nanyang Technological University Singapore

Submitted by

yigitturali

PANINI-DR: Controlled Deep Research through Generative Semantic Workspaces

·
6 authors

Submitted by

Karl28

RSIPO: Recursive Self-Improving Policy Optimization for Agentic Reinforcement Learning via Skill-Policy Co-Evolution

CUHK-CSE

The Chinese University of Hong Kong

Submitted by

mack-williams

LightVAE: Towards Compact and Efficient Video Autoencoders

·
8 authors

Submitted by

timcuhk

Controlled Acquisition and Abstention in Three-Channel Score Conflicts

NRC-CNRC

National Research Council Canada

Submitted by

yubol

Reading Right, Acting Wrong: Diagnosing Decision Failures of System One Models in Agent Loops

CarnegieMellonU

Carnegie Mellon University

Submitted by

ZiyangYe

WorldCast: Distributed Multiplayer World Models

CUHKSZ

Chinese University of Hong Kong, Shenzhen

Submitted by

ShiyuHuang

AUGUstus: Autonomous Driving VLA Model with Unified Generation and Understanding

XPENG-AI

Submitted by

leoluo25933

DoubtLess: Training-Free Contrastive Decoding for Efficient Reasoning

·
2 authors

Submitted by

augustinLib

Web$hell: A Unified Interface for Agentic Web Access

yonsei-dli

Data & Language Intelligence Lab

1

Submitted by

charrywhite

Diffusion Editing with Soft Mask: Pixel Level Redo of Image and Video with Adjustable Strength

Submitted by

dzy97

Beyond the Verdict: Agentic Code Judging with Verifiable Feedback

·
7 authors

Submitted by

ruiyangw

Who to Trust in Multi-Robot Systems: Joint Robot and Object Hypothesis Trust Inference via Neural-Symbolic Cross-Validation

DukeUniversity

Duke University

Submitted by

LeoZhang1999

Scaling Verifiable Environments for Long-horizon Work Agents

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

JackBAI

Enabling Token-wise Importance Sampling in Off-Policy RLVR with Proximal KL Regularization

UIUC-CS

University of Illinois at Urbana-Champaign

Submitted by

ashutosh1919

Can an Agent Learn to Predict Its Own Failure? Outcome-Grounded Confidence Calibration in LLM Agents

SAP

1

Submitted by

namin0202

Omni-Modal LLMs Know More Than They Say

naver-ai

Submitted by

iTheresaApocalypse

Video-Arena: A Scalable Agentic Video Benchmark for Spatiotemporal Capability

·
10 authors

Submitted by

sewon

Scaling Repeated-Data Pretraining: Batch Size, Regularization, and GPU-Time Tradeoffs

·
1 authors

Submitted by

oodgnas

Learning Transferable World-Action Models from Task-Paired Human Videos

naver-ai

Submitted by

hongst

MixLID: Token-Level Language Identification in Mixed-Language Text

nlpai-lab

NLP & AI - Korea University

Submitted by

Lostgreen

EventLens: Event-Structure Reinforcement Learning for Video Understanding

·
9 authors

1

Submitted by

duskybomb

SHELDON: Reasoning through Dual Search over Hypothesis and Observation Spaces

StanfordUniversity

Stanford University

Submitted by

jianhuizhang

4LC: 4D Latent Cache for Efficient Conditioning in Dynamic World Modeling

ZGCA

Zhongguancun Academy

Submitted by

MLukauskas

Probable but Not Valid: A Psychometric Audit of Decision Models as Synthetic Survey Respondents

·
2 authors

Submitted by

ParadiseYu

CURIO: Curiosity-Driven Test-Time Learning for Open-Ended Discovery

UIUC-CS

University of Illinois at Urbana-Champaign

Submitted by

Can111

ShellElastic: Scalable Elastic Pretraining with Simultaneous Maximal Updates and Hyperparameter Transfer

·
8 authors

Submitted by

lm2445

Judges Follow the Reference: The Derivation Gap in Verifying Tool-Call Arguments

·
5 authors

Submitted by

jarmstrong001

Privacy-Preserving Intent Fulfilment and Assurance for 6G RAN

EricssonAB

1

Submitted by

Obeisswenger

Image Tokenizers Should Be Trained Between Images

CGTuebingen

Submitted by

wdlctc

Words to World: Benchmarking LLMs in Interactive Physical Environments

·
12 authors

Submitted by

jirufengyu

Let the Mind Conceive, Let the Hand Achieve: Following Multimodal Instructions for Long-Horizon Robotic Manipulation

CUOHK

Chinese University of Hong Kong

Submitted by

wyy1112

Beyond Prompt Count: How Data Shapes Transfer in On-Policy Distillation

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

Jacklu0831

TacitVLA: Reasoning Vision-Language-Action Models Without Reasoning Latency

nvidia

1