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

Aug 23-29

Submitted by

oriuta

Apodex 1.1: Scaling Agentic Intelligence for Complex Work

apodex

Submitted by

taesiri

VBVR-Pro: A Scalable and Verifiable Suite for Native Visual Reasoning

Video-Reason

Submitted by

zhifeixie

VoiceMem: Streaming Dual-Brain Memory for Real-Time Interaction

nanyang-technological-university-singapore

Nanyang Technological University Singapore

Submitted by

HKU-Liangcai

FrontierChallenge: Evaluating Scientific Workflow Completion

apodex

2

Submitted by

hexuan21

VGI-Bench: Probing Visual Intelligence in Video Generation Models

UIUC-CS

University of Illinois at Urbana-Champaign

Submitted by

jometeorie

UrbanGround: From Local Perception to Spatial Agency in a Real-Scale City

SJTU

Shanghai Jiao Tong University

Submitted by

lyhisme

Annotations as Rollouts: Efficient and Scalable Reinforcement Learning for Video MLLMs

·
7 authors

Submitted by

LZXzju

TTPO: Test-Time Policy Optimization

·
11 authors

Submitted by

sayakpaul

PAWBench: How Far Are We from Probabilistically Aligned World Modeling?

·
14 authors

Submitted by

zxshamson

What Makes Good Agentic Data? An ACE Lens on Data Generation for LLM Agents

·
14 authors

2

Submitted by

JUNJIE99

WeMM-Embedding: WeChat Multi-Modal Embedding Technical Report

tencent

Submitted by

WeiChow

On-Policy Self-Distillation in Diffusion Models

ByteDance-Seed

Submitted by

taesiri

Self-OPD: On-Policy Distillation for Flow Matching Models without Teacher

·
11 authors

Submitted by

pshlego

AutoSaddler: Automatic Harness Optimization with Durable Updates from Agent Execution Traces

microsoft

Submitted by

eric-xiang

Graph Engineering in the Era of LLM Agents: From Individual Intelligence to System Intelligence

·
35 authors

Submitted by

Franklinzhang

EchoWM: Open and Enterable Omnimodal World Models

jdopensource

JD.com Open Source

3

Submitted by

IntJudge

Agentic Game Development as a Verifiable Trajectory Data Engine for Scaling World Models

NationalUniversityofSingapore

National University of Singapore

Submitted by

milkkarten

Prime Agent: A Self-Improving RLM Harness

PrimeIntellect

Prime Intellect

Submitted by

Leon1207

TLive-Omni: An Omni-Modal Understanding Model for E-Commerce Live Streaming

Submitted by

CuiLong7

Unlocking the Potential of Image Editing via Concept Scaling and Dense Supervision

inclusionAI

Submitted by

IffYuan

WarpSAC: Towards the Pinnacle of Scalable Off-policy RL by Rethinking Exploration and Exploitation

TianjinUniversity

Tianjin University

Submitted by

Yunncheng

GameWAM: A World Action Model for Video Games

tencent

Submitted by

taesiri

Code World Model: Coding Agent as World Brain

·
3 authors

Submitted by

YerbaPage

ParaTempo: Efficient Parallel Reasoning via Temporal Confidence

SJTU

Shanghai Jiao Tong University

Submitted by

supermodelteam

GigaBrain-0.7: Scaling Embodied Foundation Models to Emergent Capabilities with a Three-System Architecture

open-gigaai

Submitted by

taesiri

InfinityEdit: Infinite Video Editing with a Lightweight Edit-Ignition Adapter

·
12 authors

Submitted by

zsytony

Beyond Correctness: Benchmarking and Aligning Response Behaviors in Hybrid-Thinking MLLMs

tencent

Submitted by

YangXiao-nlp

PILOT in the Loop: Live Self-Improvement for Long-Horizon Agents

AllSpark-Research

AllSpark Research

4

Submitted by

taesiri

WikiSkill: Compiling Agent Experience into Persistent Knowledge for Skill Evolution

google

Submitted by

csjiaya

CyberFactory: Scaling Cyber Security Capabilities with Instances from the Wild

IQuestLab

Submitted by

lhmd

Block3D: Efficient Text-to-3D Generation via Block-Wise Diffusion

zju

Zhejiang University

Submitted by

lana-ny

Let's Scale Step by Step: Compute-Efficient Hyperparameter Transfer for Large-Scale Mixture-of-Experts

kakaocorp

3

Submitted by

Lingaaaaaaa

Recursive Experiential-Working Memory Evolution for Long-Horizon Agent Harnesses

princetonu

Princeton University

Submitted by

Hiiamein

D^3-MOPD: Adaptive Dynamic Domain ScheDuling for Efficient Multi-Teacher Distillation

·
10 authors

2

Submitted by

weigao266

MobilePA-Bench: Benchmarking Mobile Planner Agents on Complex Real-World Tasks

Tongyi-MAI

Submitted by

David-Dou

Towards a Densing Law for User Representation Learning at Billion-Scale Capacity

antgroup

Submitted by

1e12Leon

RISE: Adaptive Imagination for World Action Models

COWARobot

Submitted by

BradyFU

OmniAssistBench: Assistant-style Interaction Benchmark for Omni-LLMs

nanjinguniv

Nanjing University

Submitted by

yuuhan

Training Agents to Evolve with Their Harness: TaoLive Digital Avatar Agent Technical Report

2

Submitted by

HomieZ

Zero-WAM: In-Context World-Action Modeling from Human Videos for Open-Ended Task Generalization

Robbyant-Research

Robbyant Research

Submitted by

LZXzju

Agent-G^2: Gaussian Guidance for Agentic Reinforcement Learning

Zhejiang University

Submitted by

wileewang

ReWorld: An Interactive World Model with Long-Horizon Memory

AlibabaTongyiLab

Submitted by

c7w

Open-MOPD: Diagnosing and Fixing Capability Imbalance in Multi-Teacher On-Policy Distillation

BytedTsinghua-SIA

BytedTsinghua-SIA

Submitted by

wanng

AutoResearch: Insight In, Hallucination Out

EvoMapAI

Submitted by

tangyinhao

Is Next-Chunk Reasoning RL Really Better than SFT? Revisiting Training Strategies under no-CoT Data

·
11 authors

Submitted by

zz1358m

Understanding Evolution Strategies for LLM Reasoning: Broader Reasoning Coverage than GRPO

·
10 authors

3

Submitted by

happinessqq

StreamPI: Streaming Multimodal Temporal Modeling for Vision-Language-Action Models

·
10 authors

Submitted by

QPHutu

Best Practice Critic Optimization

NUS-DSA3101

National University of Singapore

Submitted by

zpy777

Video-IFBench: Evaluating Instruction Following of Multimodal LLMs in Video Understanding Scenarios

·
15 authors

Submitted by

taesiri

On-policy Distillation with Verifiable Reward

·
8 authors

Submitted by

xinyu1607

Beyond the Stability-Exploration Dilemma: Environmental Regularization for LLM Policy Optimization

alibabagroup

Submitted by

mis2

ARC: Fair Relative Advantage Comparison in Open-Ended Real-World Interaction

ant-intl

ant-international

Submitted by

taesiri

GameXpert-Bench: How Far Are Coding Agents from Expert Game Development?

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

LoYoT

Procedura: Agentic 3D Modeling with Procedural Control

·
8 authors

Submitted by

taesiri

A Programming Paradigm for Spatiotemporal Composability

deepseek-ai

Submitted by

taesiri

LAION-BVD: A 10-Million-Hour Open Video Dataset for Multimodal Pre-training

laion

Submitted by

shulin16

V-Rubrics: Visual Faithfulness via Rubric-Based Reinforcement Learning

nanyang-technological-university-singapore

Nanyang Technological University Singapore

Submitted by

taesiri

Meta^n: Recursive Self-Improvement through Emergent Depth

·
4 authors

Submitted by

CLoud5-real

LongRCA Bench: Diagnosing Responsible Roles and Root Causes in Long-Horizon Agent Failures

·
17 authors

2

Submitted by

teawhite

WorldMind: Decoupled Game World Model for State-Aware NPC Behavior

·
4 authors

Submitted by

DeyangKong

Every Coin Has Two Sides: On the Dual Nature of Generalization in On-Policy Distillation of Large Language Models

IQuestLab

2

Submitted by

hongdy22

SWE Refactor Bench: Can Coding Agents Complete a Long-Horizon, Whole-Repository Stack Migration?

·
10 authors

Submitted by

proy

The Handoff Tax: Continuing Non-Native Trajectories in LLM Agents

·
4 authors

Submitted by

txiong23

Rubrics as Visual-Repair Context for Self-Evolving UI-to-Code Generation

·
13 authors

Submitted by

AntonioTN

Quantization-Aware Healing: A Practical Recipe for Recovering Compressed, 4-Bit LLMs

MultiverseComputingCAI

Multiverse Computing

8

Submitted by

yjb6

UniSpace: Unified Visual Representation and Scalable Multimodal Modeling

meituan-longcat

Submitted by

guo9

Are Android GUI Agents Robust Against Runtime Anomalies? AnTrap: Evaluating Agents in Dynamic Adversarial Environments

·
8 authors

Submitted by

Minbyul

AgentMercury: Your Agent Can Synthesize Verifiable Environments for Business Scenarios at scale

·
2 authors

Submitted by

yinghuihe

CritICL: Inference-Time Weak-to-Strong Generalization from Small Language Model Failure Modes

·
7 authors

Submitted by

taesiri

Magpie: Real-Time World Renderer for Interactive Games

·
9 authors

Submitted by

Amr-Hegazy

Gated Recurrent Transformers: Expressive Depth through Recurrent Modulation

TUM

Technical University of Munich

Submitted by

alsu-sagirova

MARS: Multi-Specialist LLM Relay System for Competitive Programming

·
3 authors

Submitted by

shuaishuaicdp

Game2World Engine: Unlocking In-the-Wild Gameplay Videos for World Model Training

GameWorldData

Submitted by

AtsuMiyai

Task-CoEvolve: Efficient Harness Optimization via Adaptive Validation Task Selection

hal-utokyo

Submitted by

EnjunDu

EviRank: Structured Relevance Evidence for Multimodal Image Re-ranking

·
11 authors

2

Submitted by

HaojunChen

From Seeing to Acting: Smart Glasses as First-Person Intelligence Platforms

·
3 authors

Submitted by

bardofcodes

Luce: Relightable Gaussians for 3D Asset Generation

apple

Submitted by

abhi1nandy2

CaRGo-T: Causal Reasoning Graph-of-Thought improves Multimodal Humor Comprehension

·
8 authors

Submitted by

MrBean2024

MA-VLA: Multi-Arm Vision-Language-Action Model for Collaboration and Compositional Generalization

dalian-university-of-technology

DaLian University of Technology

Submitted by

hongyu-nvidia

Hydra-0: Action Flow for Generalist World Modeling and Control

nvidia

2

Submitted by

PardisTaghavi

Partition the Support, Reconstruct the Residual: Training-Free Sparse Attention for Video Generation and World Models

TexasAMUniversity

Texas A&M University

4

Submitted by

Abrar2652

GUI-Primitives: Diagnosing Spatial Reasoning Failures in Vision-Language GUI Grounding

usc-isi

USC Information Sciences Institute

Submitted by

xzyhku

Long-Horizon Audio-Visual Generation for Persistent Stories and Interactive Worlds

·
16 authors

Submitted by

HY-Wan

Beyond Imitation: Filtering On-Policy Distillation by Reasoning Progress

·
5 authors

Submitted by

cseeychen

Human-Centric Intelligence in the Era of Foundation Models: A Survey

·
16 authors

Submitted by

Unseen1980

Daedalus-150M: A Convolution-Attention Hybrid Designed for CPU Inference

·
1 authors

Submitted by

taesiri

Llama-Mobile: Efficient 2.7-Bit Quantization of VLMs

·
3 authors

Submitted by

seonglae

AgentRoom: Concurrent Multi-Agent Coding in a CRDT-Backed Shared Workspace

·
2 authors

3

Submitted by

wanng

LongWoF-Bench: Evaluating EvoMap Genes for Verifiable Long-Workflow Tasks

EvoMapAI

2

Submitted by

nielsr

One Success Isn't Reliability: Thinkingbox, a Sandbox and Benchmark for Agents in Stateful Business Workflows

microsoft

Submitted by

FireCRT

Thinking on Shots: Consistent Multi-Shot Video Editing with Agentic Reasoning

NankaiUniversity

Nankai University

Submitted by

taesiri

Super Star: Towards Streaming Real-time Interactive Agents for Digital Humans

·
7 authors

Submitted by

BanmaXM

When "Must" Becomes "Maybe": Constraint Weakening in LLM Agent Workflows

·
6 authors

2

Submitted by

taesiri

CAFE: Self-Improving Search Agents Need Co-Evolving Feedback

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

KingZ23

Better Retrieval, Worse Robustness:How Multi-hop RAG Amplifies Upstream ASR Errors

·
1 authors

2

Submitted by

TangJiakai5704

Towards Faithful Simulation of Human Shopping Behavior

·
12 authors

2

Submitted by

zhijie3

TacForcing: Streaming Action Generation with Execution-Time Tactile Feedback

·
11 authors

Submitted by

qxxxxxxxxxxx

What Does an Evaluation License? A Commit-Bound Census of Claim Replay in Inspect Evals

·
2 authors

3

Submitted by

vinthony

EditaLive! Unified Character Video Editing for Live Streaming

GVCLab

GVC Lab at Great Bay University

Submitted by

wchengad

Aphanta: Diagnosing Task-Aligned Image-Edited Intermediates for Multimodal Reasoning

Fudan-University

Fudan University

2

Submitted by

zhiyuan218

CaSKG: Counterfactual-Causal Skill Graphs for Scalable Agent Skill Retrieval

antgroup

Submitted by

bobbycxy

Skill Issue: Are Skills Language-Invariant in LLMs?

The-CoLab

Submitted by

Muennighoff

Prefix Sliding for efficient test-time scaling

·
18 authors

Submitted by

Wolfda95

A Modular Agent for Reliable and Auditable Spatial Relation Verification in CT Scans

·
6 authors

Submitted by

cuijiaxing

Stream4D: 4D-Consistency for Streaming Autoregressive Diffusion Video Models

UCLA

University of California, Los Angeles