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 5

Submitted by

Jiafei1224

MolmoAct2: Action Reasoning Models for Real-world Deployment

allenai

Submitted by

ssz1111

From Context to Skills: Can Language Models Learn from Context Skillfully?

·
13 authors

Submitted by

chamber111

Persistent Visual Memory: Sustaining Perception for Deep Generation in LVLMs

·
9 authors

Submitted by

stefan-it

Repetition over Diversity: High-Signal Data Filtering for Sample-Efficient German Language Modeling

Boldt

Submitted by

galyona

Hallucinations Undermine Trust; Metacognition is a Way Forward

google

Submitted by

Ningyu

OceanPile: A Large-Scale Multimodal Ocean Corpus for Foundation Models

Zhejiang University

Submitted by

Jinli4869

WindowsWorld: A Process-Centric Benchmark of Autonomous GUI Agents in Professional Cross-Application Environments

·
6 authors

Submitted by

taesiri

AcademiClaw: When Students Set Challenges for AI Agents

·
78 authors

Submitted by

Xrvitd

ComboStoc: Combinatorial Stochasticity for Diffusion Generative Models

HKUCDS

University of Hong Kong

Submitted by

RuoqiLiu

PhysicianBench: Evaluating LLM Agents in Real-World EHR Environments

StanfordUniversity

Stanford University

Submitted by

willhx

T^2PO: Uncertainty-Guided Exploration Control for Stable Multi-Turn Agentic Reinforcement Learning

·
10 authors

Submitted by

taesiri

Perceptual Flow Network for Visually Grounded Reasoning

·
11 authors

Submitted by

Newiz430

Hierarchical Abstract Tree for Cross-Document Retrieval-Augmented Generation

·
2 authors

Submitted by

taesiri

Generative Modeling with Orbit-Space Particle Flow Matching

·
6 authors

Submitted by

taesiri

Motion-Aware Caching for Efficient Autoregressive Video Generation

·
10 authors

Submitted by

RuizeHe

Linear-Time Global Visual Modeling without Explicit Attention

Tsinghua-LeapLab

Tsinghua-LeapLab

Submitted by

txdai

Counting as a minimal probe of language model reliability

StanfordUniversity

Stanford University

Submitted by

zsqzz

Agentic AI Systems Should Be Designed as Marginal Token Allocators

UIUC-CS

University of Illinois at Urbana-Champaign

Submitted by

taesiri

Code World Model Preparedness Report

·
24 authors

Submitted by

tu-trinh-scale

HiL-Bench (Human-in-Loop Benchmark): Do Agents Know When to Ask for Help?

ScaleAI

Submitted by

Francesco-pivi

BlenderRAG: High-Fidelity 3D Object Generation via Retrieval-Augmented Code Synthesis

University of Bologna

Submitted by

Kirscher

Assessing Pancreatic Ductal Adenocarcinoma Vascular Invasion: the PDACVI Benchmark

·
26 authors

2

Submitted by

lberti-equille

Prior-Aligned Data Cleaning for Tabular Foundation Models

·
1 authors

Submitted by

stefanos50

A Hybrid Approach for Closing the Sim2real Appearance Gap in Game Engine Synthetic Datasets

·
1 authors

Submitted by

zhihuang

Linking spatial biology and clinical histology via Haiku

zhihuanglab