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 8

Submitted by

taesiri

Skill1: Unified Evolution of Skill-Augmented Agents via Reinforcement Learning

·
9 authors

Submitted by

MindscapeRAG

MiA-Signature: Approximating Global Activation for Long-Context Understanding

tencent

Submitted by

Canyu

MARBLE: Multi-Aspect Reward Balance for Diffusion RL

Zhejiang University

Submitted by

linjhong

When to Trust Imagination: Adaptive Action Execution for World Action Models

·
7 authors

Submitted by

taesiri

Continuous Latent Diffusion Language Model

·
11 authors

Submitted by

bond005

RaguTeam at SemEval-2026 Task 8: Meno and Friends in a Judge-Orchestrated LLM Ensemble for Faithful Multi-Turn Response Generation

NSU

Novosibirsk State University

Submitted by

byliutao

Continuous-Time Distribution Matching for Few-Step Diffusion Distillation

alibaba-inc

Submitted by

ChengsongHuang

Nonsense Helps: Prompt Space Perturbation Broadens Reasoning Exploration

·
6 authors

Submitted by

ethanning

Auto Research with Specialist Agents Develops Effective and Non-Trivial Training Recipes

CarnegieMellonU

Carnegie Mellon University

Submitted by

scofield7419

Audio-Visual Intelligence in Large Foundation Models

NationalUniversityofSingapore

National University of Singapore

Submitted by

CuSO4-Chen

A^2TGPO: Agentic Turn-Group Policy Optimization with Adaptive Turn-level Clipping

tencent

Submitted by

ZhuofengLi

Beyond Semantic Similarity: Rethinking Retrieval for Agentic Search via Direct Corpus Interaction

TIGER-Lab

Submitted by

centaurus-alpha

UniPool: A Globally Shared Expert Pool for Mixture-of-Experts

Submitted by

taesiri

SkillOS: Learning Skill Curation for Self-Evolving Agents

·
16 authors

Submitted by

pengxiang

ReflectDrive-2: Reinforcement-Learning-Aligned Self-Editing for Discrete Diffusion Driving

·
10 authors

Submitted by

taiganga

TabEmbed: Benchmarking and Learning Generalist Embeddings for Tabular Understanding

·
8 authors

Submitted by

LazySheeep

SwiftI2V: Efficient High-Resolution Image-to-Video Generation via Conditional Segment-wise Generation

·
6 authors

Submitted by

1e12Leon

RemoteZero: Geospatial Reasoning with Zero Human Annotations

·
7 authors

Submitted by

taesiri

AI Co-Mathematician: Accelerating Mathematicians with Agentic AI

·
18 authors

Submitted by

wtl666wtl

Can RL Teach Long-Horizon Reasoning to LLMs? Expressiveness Is Key

·
7 authors

Submitted by

J017athan

The Granularity Axis: A Micro-to-Macro Latent Direction for Social Roles in Language Models

HKUCDS

University of Hong Kong

Submitted by

hdong51

Are We Making Progress in Multimodal Domain Generalization? A Comprehensive Benchmark Study

ethz

Submitted by

taesiri

Think, then Score: Decoupled Reasoning and Scoring for Video Reward Modeling

·
11 authors

Submitted by

stdKonjac

Sparkle: Realizing Lively Instruction-Guided Video Background Replacement via Decoupled Guidance

showlab

Submitted by

SushantGautam

When No Benchmark Exists: Validating Comparative LLM Safety Scoring Without Ground-Truth Labels

SimulaMet

Simula Metropolitan Center for Digital Engineering (SimulaMet)

Submitted by

Jim137

Generative Quantum-inspired Kolmogorov-Arnold Eigensolver

·
12 authors

Submitted by

jt-zhang

KernelBench-X: A Comprehensive Benchmark for Evaluating LLM-Generated GPU Kernels

Tsinghua University

Submitted by

envomp

The Scaling Properties of Implicit Deductive Reasoning in Transformers

taltechappai

Applied AI group of the Tallinn University of Technology

Submitted by

jl3353

Prescriptive Scaling Laws for Data Constrained Training

·
5 authors

Submitted by

sotaagi

Recovering Hidden Reward in Diffusion-Based Policies

·
10 authors

2

Submitted by

SWY666

EDU-CIRCUIT-HW: Evaluating Multimodal Large Language Models on Real-World University-Level STEM Student Handwritten Solutions

·
6 authors

Submitted by

zyzeng

Balanced Aggregation: Understanding and Fixing Aggregation Bias in GRPO

OpenMOSS-Team

Submitted by

gxx27

BioTool: A Comprehensive Tool-Calling Dataset for Enhancing Biomedical Capabilities of Large Language Models

University of California San Diego

Submitted by

Ajay1994

TIDE: Every Layer Knows the Token Beneath the Context

apple