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 5

Submitted by

ArmanBehnam

RealCompanion: Benchmarking Human Understanding from Reasoning over Longitudinal Real-World Conversations

Quislab

1

Submitted by

liuyong9975

Does Learning Protein Folding Generalize to Broader Reasoning?

SJUT1

Shanghai JiaoTong University

Submitted by

bx6d

MotorMind: Scaffolding General Vision Language Models for Zero-Shot Robot Manipulation

UIUC-CS

University of Illinois at Urbana-Champaign

Submitted by

YINBO0927

FrameMorrow: Future-guided Frame Selection with Prospective Tokens for Long-Horizon Video Generation

NationalUniversityofSingapore

National University of Singapore

Submitted by

taesiri

Scaling Trajectories for Complex Tasks through Recursive Self-Rewrite

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

ZhaochongAn

World Action Modeling with Progressive Visual Planning

meta

Submitted by

shufanshen

On-Policy Parameter Update Direction Underlies Generalization in LLM Post-Training

UCAS

Submitted by

ZhaochongAn

Native Action-Prior Learning from Videos for World Action Models

meta

Submitted by

afaji

HyperBrowseComp: A Multilingual and Multimodal Stress Test for Web-Browsing Agents

MBZUAI

Mohamed Bin Zayed University of Artificial Intelligence

Submitted by

shkim0116

Pivot-SD: Efficient Self-Distillation for Masked Diffusion Language Models

·
6 authors

1

Submitted by

tztgo

PDE-JEPA: Predictive Representation Learning of Latent Dynamics Modeling for Parametric PDEs

Zhejiang University

Submitted by

GenuineWWD

SimuVerity: Benchmarking Agents for Engineering-Grade Simulink Model Generation

XianJiaotongUniversity

Xi'an Jiaotong University

Submitted by

yerim0210

Science or Slop?: Benchmarking and Mitigating Scientific Slop in AI-Generated Papers

SeoulNatlUniv

Seoul National University

Submitted by

fangzy

Latent-MOPD: Latent Multi-Teacher On-Policy Distillation

zillow

1

Submitted by

taesiri

EditHero: A Benchmark for Long-Horizon Part-Level 3D Editing and Vibe Modeling

·
12 authors

Submitted by

caiqizh

VeriHarness: Scaling Agentic Verification for Long-Horizon Tasks

google

Submitted by

yidacai

LexReward: A Taxonomy-Driven Reward Framework for Legal Language Models

TsinghuaNLP

Tsinghua NLP Group

1

Submitted by

KennethEnevoldsen

Multilingual GSM-Symbolic: What determines capability transfer across languages?

danish-foundation-models

Danish Foundation Models

Submitted by

gowitheflow

World Embedding Benchmark

World-Representation-Lab

World Representation Lab

Submitted by

yycfq1314

Spatial Memory Intelligence: Endowing World Models with Understanding-Driven Long-Term Memory

·
8 authors

Submitted by

Ahren09

Science Utopia? Closed-Loop LLM Simulation of Academic Research Ecosystems

GeorgiaTech

Georgia Institute of Technology

Submitted by

hyung22

Source Preference in the Wild: How LLM Agents Favor Items by Source, and How to Reduce It

SeoulNatlUniv

Seoul National University

Submitted by

Zeyue7

HelixWorld: A Real-time Interactive Audio-Visual World Model

NoizAI

Submitted by

dorienh

DEFINE: Exemplar-Guided Accent Control for Zero-Shot TTS

amaai-lab

Submitted by

LinghuiShen

ProAR: Learning Prospective Reasoning with Autoregressive Video Models

ucdavis

Submitted by

OliverSieberling

Triadic Linear Attention: Three-Dimensional Recurrent States for Long-Context Sequence Modeling

Massachusetts-Institute-of-Technology

Massachusetts Institute of Technology

Submitted by

xmz111

Beyond Future Prediction: Denoising as Generative Adaptation for Robot Control

·
7 authors

Submitted by

nexync

Language Models that Play Chess and Explain Their Moves

·
3 authors

Submitted by

chchenhui

FrugalEvo: Towards Cost-Aware LLM-Guided Program Evolution

·
8 authors

Submitted by

assafbk

Local Support Learning

MIT

Massachusetts Institute of Technology

Submitted by

dorienh

Diptych: Scoped, AI-Interpreted Comparison for Reference Listening in Music Production

amaai-lab

1

Submitted by

jaehong31

MetaRubric: Learning to Reward for Rubric-Based Reinforcement Learning

nanyang-technological-university-singapore

Nanyang Technological University Singapore

Submitted by

ruihong04

4DCodeBench: Benchmarking Agents on Inverse Graphics of Dynamic Scenes

4DCodeBench

Submitted by

markus583

Equal Ranking Quality, Different Decisions: Measuring and Reducing Order Dependence in LLM Scorers

thomsonreuters

Thomson Reuters

Submitted by

cjeen

Rollout-Marginal Distillation for Long-Horizon Autoregressive Video Generation

tencent

Submitted by

neikos00

LVMT: Video Mask Transformer for Long-term Video Segmentation

tue-mps

Mobile Perception Systems Lab

Submitted by

taesiri

Skill2Real: Agentic Skill Learning for Zero-Shot Sim-to-Real Robot Manipulation

·
8 authors

Submitted by

Marble666

WEFT: Scaling Tool-Use Post-Training for General-Purpose Agents

nex-agi

1

Submitted by

Elfsong

Collective Bias Mitigation via Model Routing and Collaboration

NUS-IDS

Institute of Data Science

Submitted by

Shiweiliuiiiiiii

Looping Beyond Twice: A Scalable Recipe for Looped Mixture-of-Experts

Intelligent-Systems

Max Planck Institute for Intelligent Systems

Submitted by

paulibo

Tail-Influence Sampling for CVaR Policy Evaluation

huawei-noah

HUAWEI Noah's Ark Lab

Submitted by

franciscoliu

Learning from Runtime Feedback through Failure-Bank Self-Evolution for Vision-Language-Action Models

notredame

University of Notre Dame

Submitted by

Samll

Efficient Reasoning Training Does Not Always Harm CoT Faithfulness and Monitorability

·
5 authors

Submitted by

Deema

Covert Assistance: Helpful LLM Agents Evade Oversight in Multi-Agent Systems

illinois

University of Illinois at Urbana-Champaign

1

Submitted by

ruyi101

MinkowskiPE: Minkowski Positional Encoding for Spatiotemporal Perception

·
7 authors

Submitted by

Xiiiiii0220

GTR: Gated Token Recurrence for Efficient Dense Prediction

Intellindust

Submitted by

Heyuan9

Where Does Retrieval-Based Open-Ended Evaluation Fail? Automatic Taxonomy Induction from Long-Form Medical Answer Factuality Verification

jhu-clsp

Center for Language and Speech Processing @ JHU

Submitted by

haonanqqq

Rethinking Token Reweighting for SFT: Suppress, Reverse, and Extrapolate Learned Features

Shanghai-AI-Laboratory

Shanghai AI Laboratory

Submitted by

zsqzz

From Gradients to Capabilities: Understanding Multi-Teacher On-Policy Distillation

UIUC-CS

University of Illinois at Urbana-Champaign

1

Submitted by

pritamdeka

From Retrieval to Typed Decisions: Calibrated System One Models from Biomedical Sentence Encoders

·
1 authors

Submitted by

taesiri

Octrees as an Explicit 3D Language

·
6 authors

Submitted by

maxr0ssi

Receiver-Conditioned Latent Communication gives 94% CacheBack

·
5 authors

Submitted by

wisepigeon

EyeRobot 2.0: Active Gaze for Precise Manipulation without Wrist Cameras

Berkeley

Submitted by

zjq0455

QuantWM: Temporally Consistent 2-Bit KV Cache Quantization for Video World Models

NationalUniversityofSingapore

National University of Singapore

1

Submitted by

yzha

WM-VLM: Probing Internal World Models for Interleaved Visual-Textual Reasoning

UCSanDiego

University of California at San Diego

Submitted by

Devin6666

Can Computation from Earlier Problems Help LLMs Solve New Ones?

RUC

Renmin University of China

1

Submitted by

zzzxyyyyxxx

Dream4ACT: A Shared Visual Action Interface for Multi-Embodiment Video-Action Modeling

·
9 authors

1

Submitted by

Zhongzhi1228

Self-Supervised Scaling of Terminal Environments for Scientific Domains

·
11 authors

Submitted by

ZihanC

KeyRec: Bounded Visual Memory for Streaming and Long-Video Understanding

google

1

Submitted by

tmaham

Strike a Chord! Modal Kinetic Typography

·
4 authors

Submitted by

lliutianc

Improving Atomic-Fact Recall via Focused Views in Unstructured Knowledge Editing

·
4 authors

Submitted by

jeriksson

The Extender: A Log-Structured Transformer

University of Illinois Chicago

1