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 12

Submitted by

manchery

S-Space: Exploring Spatial Workspace in Multimodal Language Models

MirroS-Lab

Submitted by

vanzl

DiffusionART: Efficient, Stable and Scalable Reinforcement of Diffusion Models

KlingTeam

KlingAI Research

Submitted by

Jiafei1224

Reflex: Visually Grounded Reactive Humanoid Control

University of Washington

Submitted by

wwh0411

SoftwareGym: Scaling Computer-Use Agent Training with Diverse Synthesized Software Environments

SJTU

Shanghai Jiao Tong University

Submitted by

JadeHuang

EmbodiedCLBench: Evaluating Continual Learning for Self-Evolving Embodied Agents

PKU-VaLuE-Lab

Submitted by

yuanfengustc

VisionWeave: Weaving Elastic Visual Representations as a Native Capability of MLLMs

Qwen

Submitted by

ptjiang

Here the World in Stereo: Learning Dynamic Spatial Correspondence for Immersive Joint Video-Audio Generation

·
10 authors

Submitted by

ChengyiYang

ProgramGym: Scaling Training Environments for Long-Horizon Program Reproduction Agents

ProgramGym

Submitted by

Reiss

Continually Post-training Small Models via Synthetic Self-distillation

LG-AI-Research

Submitted by

AviadDahan

SepGen: Multi-Stem Audio-Video Separation and Generation in a Single Model

·
6 authors

Submitted by

vyokky

Work2Skill: Capturing Human GUI Experience for Reusable Agent Skill Induction

nex-agi

Submitted by

Aoraku

Can AI Agents Learn Their Way to the Top? Evaluating Heuristic Learning in a Long-Running Game Agent Competition

THU-CST-SAST

Submitted by

donghyunli

SoloQ: Calibration-Free Quantization for Diffusion Language Models

UniversityofSouthernCalifornia

University of Southern California

Submitted by

arnodjiang

PolyOCR-Venus: Unified OCR Foundation Models for Text-Centric Visual Intelligence

inclusionAI

Submitted by

Lingaaaaaaa

RSI-Anything: Unlocking Recursive Self-Improvement with World Models

PhAILabs

Submitted by

Apostle723

Learning Representations to Guide Pixel Diffusion Transformers

alibabagroup

Submitted by

royeis

Contextual Correction of Parallel Draftsfor Speculative Decoding

·
8 authors

Submitted by

XuanlangDai

InstantFusion: A Shared Latent Interface for Heterogeneous Image Generators

Shanghai-AI-Laboratory

Shanghai AI Laboratory

Submitted by

LulaCola

Inducing Process Supervision from Outcome-Only Reinforcement Learning

RUC

Renmin University of China

Submitted by

RuijieZhu

Towards Physically Grounded Articulated 3D Assets: A Multi-Agent Pipeline and Benchmark

nanyang-technological-university-singapore

Nanyang Technological University Singapore

Submitted by

zhensuuu

SWE-Duel: Towards a Self-Scaling Arena for Coding Agents

SingaporeManagementUniversity

Singapore Management University

Submitted by

AlexMashiro

Judging in Latent Space: Efficient Generative Reward Modeling via Semantics-Preserving Compression

·
8 authors

Submitted by

YIGE24

MetaKernelBench: Measuring GPU Kernel Knowledge Transfer Beyond Code

encodelab

Westlake ENCODE Lab

Submitted by

YuxinXiao

Recurrent Harness for Agentic Continual Learning

·
7 authors

Submitted by

chenfeng1271

MixTok: Adaptive Visual Tokenization with Global 1D Tokens and Sparse 2D Refinement

Zhejiang University

Submitted by

VictorYXL

GraphDecisionBench: Benchmarking System One Models on Graph Tasks

·
3 authors

Submitted by

xypkent

DataVista: Diagnosing Multimodal LLMs on Data Video Understanding

·
7 authors

Submitted by

Vinyeh

From Given to Gathered Evidence: Agentic Learning for Longitudinal Medical Reasoning

CASE-Bench

Submitted by

natnitaract

AMLworld-Compact: Importance-Weighted Downsampling for Cost-Effective LLM Evaluation and Error Diagnosis

typhoon-ai

Submitted by

SingleBicycle

TouchScale: 500 Hours of Human Vision and Touch for Visual-Tactile Learning

2077AIDataFoundation

Submitted by

RaghavvGoel

Surfing the Activation Wave: Analyzing the Information and Geometry of LLM Hidden States for Speculative Decoding

Qualcomm-AI-Research

Qualcomm AI Research

Submitted by

GUALIN

Video-RSI: Recursive Self-Improvement of Video Understanding Agents via Harness Evolution

·
3 authors

Submitted by

Edaizi

Rethinking Relative Policy Differences Transfer of On-Policy Distillation

KnowledgeXLab

KnowledgeXLab@Shanghai AI Lab

Submitted by

Huan-WhoRegisteredMyName

Universal Textual Teaching for LLMs

·
2 authors

1

Submitted by

jiangnanhugo

PreSTO: Predictive Subtree Prefetching for Fast LLM Power Sampling

morganstanley

Submitted by

hlzhang109

When Do Biological Reasoning Models Use Their Biological Inputs?

·
6 authors

Submitted by

xferdie

The value of information and the emergence of communication in evolved e-puck robots

uv-mex

Universidad Veracruzana

Submitted by

HwanChang0106

Stateless Language Agents: Scaling Long-Horizon Automated Research

StanfordUniversity

Stanford University

Submitted by

YongganFu

When Do We Need On-Policy Distillation? Distilling on Offline Student Rollouts Is Often Better

nvidia

Submitted by

IgorSlinko

Coding-Agent Benchmarks Should Match Their Users' Task Flows

JetBrains-Research

JetBrains Research

Submitted by

TonyTeng

Distributionally Robust Mixture-of-Experts Training

newyorkuniversity

New York University

Submitted by

lmsdss

Less from More: Reinforcing Sparse Video Reasoning from Dense References

AmsUni

University of Amsterdam

Submitted by

aaronmac

The Life Cycle of a Massive Activation: Stochastic Birth, Weight-Decay-Driven Growth, and Competitive Consolidation

Aimpoint-Digital

Aimpoint Digital

Submitted by

zunhai

QuantMLA: Function-Aligned Dual-Path Quantization for Low-Bit MLA KV Caching

meituan-longcat

Submitted by

shantanuacharya

One Model to Loop Them All: Elastic Latent Recurrence for Language Models

nvidia

Submitted by

teemosliang

SAGE: Salient Factor Discovery and Generation with Visual Foundation Representations

·
8 authors

Submitted by

sqvareinch

Self-Preservation in Language Models

newyorkuniversity

New York University

Submitted by

MingyuanLei

MAMHOI: Factorizing Scene-Aware Human-Object Interaction through Affordances

nanyang-technological-university-singapore

Nanyang Technological University Singapore

Submitted by

NeutrinoLiu

Encore: Few-Shot Agentic Discovery of Manipulation Strategies

2077AIDataFoundation

Submitted by

atfortes

ThinkDrag: Semantic Drag-Based Image Editing with Visual Reasoning

mmlab-ntu

Submitted by

TianjinHuang

When to Rethink: Learning Multi-Perspective Self-Verification for Vision-Language Models

·
13 authors

Submitted by

jlee-larr

Dex-One2Many: Learning Dexterous Manipulation from a Single Human Demonstration

·
11 authors

1

Submitted by

limhyeonseok

Which Language Should a Skeleton Speak? Language Choices in Multilingual Reasoning

·
6 authors