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 22

Submitted by

Jungang

StreamEval: Evaluating Interactive Streaming Video Generation Models

·
1 authors

Submitted by

lhpku20010120

SpatialScout: Can VLM Agents Actively Explore and Map 3D Scenes?

PekingUniversity

Peking University

Submitted by

Caoza

DONIC: Generalist Dexterous Bimanual Tracking Control from Human Demonstrations

ropedia-ai

Submitted by

HELLORPG

Looking Ahead: Modeling Future Evolution in Autoregressive Video Generation

TencentARC

ARC Lab, Tencent

Submitted by

fuvty

Cache Engineering for Agents

nics-efc

Tsinghua-NICS-EFC

Submitted by

fyzzzzzz

Right Screen, Wrong Transition: World Models as Verifiers for GUI Agents

·
6 authors

Submitted by

Hunter26666666

Spectral-Spatial Interpretation

·
3 authors

Submitted by

ChilleD

Agent World Model in the Wild: Computer-Use Agent Environment Synthesis

MicrosoftResearch

Microsoft Research

Submitted by

QiushiSun

PrismaCoder: Pushing the Boundaries of Open Multimodal Code Intelligence through Decomposed Rubric Synthesis

hkunlp

NLP Group of The University of Hong Kong

Submitted by

iTheresaApocalypse

Scaling in Games: Continued Pre-Training For Embodied Agents In Diverse Virtual Worlds

THUSI-Lab

Tsinghua Spatial Intelligence Lab

Submitted by

CheeryLJH

Socratic Skill Learning: Interleaving Teaching and Practice for Transferable Agent Skills

NJU-LINK

Submitted by

dolev31

Retrieval Augmented Generation Compositionality

ibm

Submitted by

limingcv

AgenticRoboRL: Bringing Coding-Agent Post-Training to Agentic Robotics

University of Central Florida

Submitted by

PatrickHaller

PACE: Periodic and Channel-Aligned Encoding of Four-Bit Recurrent States

huberlin

Humboldt Universität zu Berlin

7

Submitted by

libertywing

TraceForge: Distilling with Your Own Thinking

Independent Research

Submitted by

crazyofapple

Memory-Centric Continual Learning for AI Agents

SLAIAITP

Submitted by

YanzheChen

Multimodal Embodied Agents: A Unified Capability-centric Perspective from Computer-Use to Robot-Use

showlab

Submitted by

Yanqing0327

ALPS: Visual State Encoders for World Models

UCSC-VLAA

Submitted by

Tianle

DeLoop: Repeated Actions as Self-Supervision for Long-Horizon Agents

google

Submitted by

JuliaKreutzerCohere

CulturalRiddles: A Multicultural Benchmark of Riddles

CohereLabsCommunity

Cohere Labs Community

Submitted by

yujianll

Building AI Research Teams for Autonomous Discovery of Linear Attention Architectures

UCSantaBarbara

University of California, Santa Barbara

Submitted by

Teaven

Harness Scaling for Generalizable Agentic Reinforcement Learning

RUC

Renmin University of China

Submitted by

Yesianrohn

Real In, Real Out: What If We Only Use Real Data for Scene Text Editing?

Fudan-University

Fudan University

Submitted by

Makayi

Sparsity Should Be ALLOTted by Output: Training-Free Allocation of Block-Sparse Attention for Video Diffusion Transformers

KlingTeam

KlingAI Research

3

Submitted by

MingZhong

Automating ML for Science: Can Frontier Agents Climb Scientific Hills in the Wild?

Anthropic

Submitted by

binghe-sjtu

CompactSplat: Spatially Adaptive Gaussian Distribution for Feedforward Scene Reconstruction

SJTU

Shanghai Jiao Tong University

Submitted by

krahets

RealMe: Immersive Volumetric Video from Minimal Captures

Submitted by

alan1027

Entertain, then Cancel: How Thinking Makes Reasoning Models Read Some and Or Literally

·
7 authors

Submitted by

yingmanji

SpatialBrain: Spatial Intelligence Growing from Experience

SJTU

Shanghai Jiao Tong University

2

Submitted by

taesiri

Qwen-Audio-3.1-ASR Technical Report

·
48 authors

Submitted by

TangJiakai5704

Think Harder Where It Matters: Item-Level Adaptive Recursion for Sequential Recommendation

·
5 authors

Submitted by

TimothyCzp

Evolving with Open Worlds: Towards Recursive Self-Improvement for Future Forecasting

·
15 authors

Submitted by

taesiri

Agentic World Action Models

·
18 authors

Submitted by

ghy0324

From VLM Planners to Robotic Agents: Self-Evolving Harnesses for History-Dependent Manipulation

·
12 authors

Submitted by

yananma

Which Agent's Skill Should Change? Group-Relative Credit Assignment for Evolving Multi-Agent Teams

Submitted by

nicholaslee

StyleStream 2.0: Fast and Controllable Streaming Voice Style Conversion

Berkeley

Submitted by

Purdy0228

From RL Papers to Framework Code: Inclusion and Review in Seven Open-Source Training Frameworks

·
1 authors

Submitted by

krennic

WorldStream: Streaming Interactive 3D Worlds with Hierarchical Planning and Persistent State

·
14 authors

Submitted by

ye11owz13

ACE-Brain-0.6: Scaling Spatial Grounding for Physical Foundation Models

ACERobotics

Submitted by

jsantillana

VX-One: Typed, Calibrated Decision Heads for Real-Time Perimeter Triage - An empirical study of unknown-attack honesty, cascade gray zones, and calibration under drift

·
1 authors

Submitted by

alan1027

Answering in the User's Variety: Calibrated Accommodation Pairs for Post-Training

·
7 authors

Submitted by

Aratako

RubatoCodec: Budgeted Adaptive Speech Tokenization with Streaming Decoding

third-intelligence

Third Intelligence, Inc.

Submitted by

Hanyu-Meng

Toward Part-Aware Choral Transcription with singing voice assignment

unswnkg

Submitted by

RobotLearningLab

Hidden Messages Between Diffusion Language Model Agents: Causal Evidence for Latent Communication

ITMO

ITMO University

Submitted by

nsivaku

ToM2Vec: Continuous Theory-of-Mind Tokens for Multi-Agent Coordination

UNC-ChapelHill

University of North Carolina at Chapel Hill

Submitted by

RobotLearningLab

DiffResearch: Accelerating Deep Research with Diffusion Language Models

ITMO

ITMO University

Submitted by

ventr1c

Language Models Learn to Improve Language Models

·
10 authors

Submitted by

DwanZhang

VerifierBench: Benchmarking the Generation of Executable Verifiers for LLM Agents

·
7 authors

Submitted by

saitejautpala

Looped Uniform Diffusion Language Models

·
4 authors

Submitted by

Ashitemaru

Quant-Arena: Benchmarking LLM Agents for Quantitative Trading Research

·
12 authors

1

Submitted by

zzhao0104

Exploring The Empirical Privacy Landscape of Synthetic Data

·
6 authors

2

Submitted by

Siyuan-Feng

Design Anything, Design at Scale

·
11 authors

Submitted by

Pd172944

Stratified Sparse Attention: Reducing the Sparse-to-Dense Gap at Extreme Sparsity

Skylight1

Submitted by

HaochengWang23

RoboCasa-Shift: Stage-Aware Robustness Evaluation of Vision-Language-Action Policies in Long-Horizon Household Manipulation

·
9 authors

Submitted by

yihongLiu

Bridged Self-Distillation for On-Policy Learning

TUM

Technical University of Munich

Submitted by

qmang

FrontierCS 2: A Benchmark for Computer Science R&D

·
61 authors

Submitted by

Roxanne-WANG

MMedAgent-Plus: Scaling Multimodal Medical Agents across Modalities, Tasks and Workflows

MMedAgent

Submitted by

grenadeComing

Same Simulation, Different Evidence: Where Retrieved Tool Output Costs LLM Agents Accuracy

ucriverside