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

May 18

Submitted by

zr-wang

CiteVQA: Benchmarking Evidence Attribution for Trustworthy Document Intelligence

opendatalab

Submitted by

LiamLian0727

PhysBrain 1.0 Technical Report

DeepCybo

Submitted by

zhangkangning

MMSkills: Towards Multimodal Skills for General Visual Agents

ShanghaiJiaotongUniversity

Shanghai Jiaotong University 1(NOT OFFICIAL)

Submitted by

DukeShen

FashionChameleon: Towards Real-Time and Interactive Human-Garment Video Customization

alibaba-inc

Submitted by

caiyuchen

Learning to Foresee: Unveiling the Unlocking Efficiency of On-Policy Distillation

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

Abyssaledge

DexJoCo: A Benchmark and Toolkit for Task-Oriented Dexterous Manipulation on MuJoCo

·
14 authors

Submitted by

hamzzi

Distilling Long-CoT Reasoning through Collaborative Step-wise Multi-Teacher Decoding

DISLab

Data Intelligence System Lab

Submitted by

yueyang2000

InsightTok: Improving Text and Face Fidelity in Discrete Tokenization for Autoregressive Image Generation

Tsinghua-LeapLab

Tsinghua-LeapLab

Submitted by

taesiri

Flash-GRPO: Efficient Alignment for Video Diffusion via One-Step Policy Optimization

·
12 authors

Submitted by

Nardien

Nudging Beyond the Comfort Zone: Efficient Strategy-Guided Exploration for RLVR

kaist-ai

Submitted by

INV-WZQ

ReactiveGWM: Steering NPC in Reactive Game World Models

·
7 authors

Submitted by

scyyc9

Hölder Policy Optimisation

·
11 authors

Submitted by

CheeryLJH

Solvita: Enhancing Large Language Models for Competitive Programming via Agentic Evolution

NJU-LINK

Submitted by

AniSundar18

From Plans to Pixels: Learning to Plan and Orchestrate for Open-Ended Image Editing

uw-madison

University of Wisconsin - Madison

Submitted by

Mercury7353

MetaAgent-X : Breaking the Ceiling of Automatic Multi-Agent Systems via End-to-End Reinforcement Learning

OregonStateUniversity

Oregon State University

Submitted by

Jungang

CM-EVS: Sparse Panoramic RGB-D-Pose Data for Complete Scene Coverage

·
16 authors

Submitted by

chengtan9907

PAGER: Bridging the Semantic-Execution Gap in Point-Precise Geometric GUI Control

opendatalab

Submitted by

aamixsh

Steered LLM Activations are Non-Surjective

JohnsHopkins

Johns Hopkins University

Submitted by

JonnyYu828

Unlocking Dense Metric Depth Estimation in VLMs

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

LCZZZZ

Auditing Agent Harness Safety

ucsbnlp

UC Santa Barbara NLP Group

Submitted by

DhavalPatel

DiagnosticIQ: A Benchmark for LLM-Based Industrial Maintenance Action Recommendation from Symbolic Rules

ibm

Submitted by

taesiri

Agentic Discovery of Neural Architectures: AIRA-Compose and AIRA-Design

·
8 authors

Submitted by

taesiri

Look Before You Leap: Autonomous Exploration for LLM Agents

·
9 authors

Submitted by

itsthanhtung

Efficient Image Synthesis with Sphere Latent Encoder

MBZUAI

Mohamed Bin Zayed University of Artificial Intelligence

1

Submitted by

Gigglingface

Sparse Autoencoders enable Robust and Interpretable Fine-tuning of CLIP models

·
4 authors

Submitted by

taesiri

FFAvatar: Few-Shot, Feed-Forward, and Generalizable Avatar Reconstruction

·
6 authors

Submitted by

satpalsr

MobileEgo Anywhere: Open Infrastructure for long horizon egocentric data on commodity hardware

fpvlabs

Submitted by

taesiri

WorldAct: Activating Monolithic 3D Worlds into Interactive-Ready Object-Centric Scenes

·
6 authors

Submitted by

Bowieee

Learning from Failures: Correction-Oriented Policy Optimization with Verifiable Rewards

·
9 authors

Submitted by

valentinsix

Learning POMDP World Models from Observations with Language-Model Priors

·
10 authors

Submitted by

n3il666

HodgeCover: Higher-Order Topological Coverage Drives Compression of Sparse Mixture-of-Experts

princetonu

Princeton University

1

Submitted by

blaz-r

ChangeFlow -- Latent Rectified Flow for Change Detection in Remote Sensing

·
4 authors

Submitted by

fxmeng

GQLA: Group-Query Latent Attention for Hardware-Adaptive Large Language Model Decoding

PekingUniversity

Peking University

Submitted by

QuanjianSong

OmniHumanoid: Streaming Cross-Embodiment Video Generation with Paired-Free Adaptation

·
5 authors

Submitted by

pratinavsetharya

Forgetting That Sticks: Quantization-Permanent Unlearning via Circuit Attribution

Lexsi

Submitted by

shanyangmie

Physics-R1: An Audited Olympiad Corpus and Recipe for Visual Physics Reasoning

·
1 authors

Submitted by

pedrocurvo

Follow the Mean: Reference-Guided Flow Matching

uva

University of Amsterdam

Submitted by

haopt

Raster2Seq: Polygon Sequence Generation for Floorplan Reconstruction

cornell

Cornell University

Submitted by

yyliu01

AuralSAM2: Enabling SAM2 Hear Through Pyramid Audio-Visual Feature Prompting

Oxford

University of Oxford

Submitted by

CoffeeGitta

Known By Their Actions: Fingerprinting LLM Browser Agents via UI Traces

·
5 authors

Submitted by

isaaccorley

No One Knows the State of the Art in Geospatial Foundation Models

·
9 authors

Submitted by

yjoonjang

MLAIRE: Multilingual Language-Aware Information Retrieval Evaluation Protocal

·
4 authors

Submitted by

Konstantine4096

Stress-Testing the Reasoning Competence of LLMs With Proofs Under Minimal Formalism

·
2 authors