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

Apr 7

Submitted by

taesiri

OpenWorldLib: A Unified Codebase and Definition of Advanced World Models

PekingUniversity

Peking University

Submitted by

wanderkid

MinerU2.5-Pro: Pushing the Limits of Data-Centric Document Parsing at Scale

opendatalab

Submitted by

chanyoungkim

LIBERO-Para: A Diagnostic Benchmark and Metrics for Paraphrase Robustness in VLA Models

HAI-Lab

Human-centered AI Laboratory

Submitted by

AaronHuangWei

TriAttention: Efficient Long Reasoning with Trigonometric KV Compression

nvidia

Submitted by

hongyyyyy

Adam's Law: Textual Frequency Law on Large Language Models

FaceMind

Submitted by

jinpeng0528

AURA: Always-On Understanding and Real-Time Assistance via Video Streams

·
12 authors

Submitted by

taesiri

FileGram: Grounding Agent Personalization in File-System Behavioral Traces

·
9 authors

Submitted by

EasonXiao-888

SpatialEdit: Benchmarking Fine-Grained Image Spatial Editing

jdopensource

Submitted by

richardxp888

ClawArena: Benchmarking AI Agents in Evolving Information Environments

UNC-ChapelHill

University of North Carolina at Chapel Hill

Submitted by

Yukirsh

LightThinker++: From Reasoning Compression to Memory Management

Zhejiang University

Submitted by

zd21

Unifying Group-Relative and Self-Distillation Policy Optimization via Sample Routing

·
9 authors

Submitted by

gallilmaimon

Self-Execution Simulation Improves Coding Models

·
7 authors

1

Submitted by

231sm

SkillX: Automatically Constructing Skill Knowledge Bases for Agents

Zhejiang University

Submitted by

taesiri

Vero: An Open RL Recipe for General Visual Reasoning

·
6 authors

Submitted by

PahaII

Your Agent, Their Asset: A Real-World Safety Analysis of OpenClaw

UCSC-VLAA

Submitted by

Shenzhi

Can LLMs Learn to Reason Robustly under Noisy Supervision?

·
10 authors

Submitted by

CUDAOUTOFMEMORY

PLUME: Latent Reasoning Based Universal Multimodal Embedding

CASIA

Chinese Academic of Science Institute of Automation

Submitted by

zzzlou

Learning to Learn-at-Test-Time: Language Agents with Learnable Adaptation Policies

NationalUniversityofSingapore

National University of Singapore

Submitted by

taesiri

Memory Intelligence Agent

·
9 authors

Submitted by

lhmd

Less Detail, Better Answers: Degradation-Driven Prompting for VQA

ziplab

Submitted by

MartaYang007

ONE-SHOT: Compositional Human-Environment Video Synthesis via Spatial-Decoupled Motion Injection and Hybrid Context Integration

NationalUniversityofSingapore

National University of Singapore

1

Submitted by

CUDAOUTOFMEMORY

CLEAR: Unlocking Generative Potential for Degraded Image Understanding in Unified Multimodal Models

·
9 authors

Submitted by

KumaPower

AvatarPointillist: AutoRegressive 4D Gaussian Avatarization

antgroup

Submitted by

zyhang1998

Synthetic Sandbox for Training Machine Learning Engineering Agents

Submitted by

siri-helixar

HDP: A Lightweight Cryptographic Protocol for Human Delegation Provenance in Agentic AI Systems

HelixarAI

Submitted by

pcr2120

The Geometric Alignment Tax: Tokenization vs. Continuous Geometry in Scientific Foundation Models

·
1 authors

Submitted by

JiahaoChen1

SciLT: Long-Tailed Classification in Scientific Image Domains

·
2 authors

1

Submitted by

Elfsong

Paper Espresso: From Paper Overload to Research Insight

NationalUniversityofSingapore

National University of Singapore

1

Submitted by

Nicolas-BZRD

BidirLM: From Text to Omnimodal Bidirectional Encoders by Adapting and Composing Causal LLMs

BidirLM

BidirLM (Decoder To Encoder)

1

Submitted by

huangrm

Mimic Intent, Not Just Trajectories

sii-research

Shanghai Innovation Institute

Submitted by

dinobby

Cog-DRIFT: Exploration on Adaptively Reformulated Instances Enables Learning from Hard Reasoning Problems

·
7 authors

Submitted by

cl-dataset

POEMetric: The Last Stanza of Humanity

·
3 authors

Submitted by

TheDetective

Do Audio-Visual Large Language Models Really See and Hear?

gamma-lab-umd

Submitted by

arkanemystic

Type-Checked Compliance: Deterministic Guardrails for Agentic Financial Systems Using Lean 4 Theorem Proving

·
2 authors

Submitted by

Liancheng2

Locally Confident, Globally Stuck: The Quality-Exploration Dilemma in Diffusion Language Models

University of Illinois Chicago

Submitted by

jasonwu1017

Scaling Teams or Scaling Time? Memory Enabled Lifelong Learning in LLM Multi-Agent Systems

emory

Emory University

Submitted by

TomekKaszynski

Emergent Compositional Communication for Latent World Properties

·
1 authors