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

Feb 10

Submitted by

yangzhi1

QuantaAlpha: An Evolutionary Framework for LLM-Driven Alpha Mining

QuantaAlpha

Submitted by

Yikunb

Weak-Driven Learning: How Weak Agents make Strong Agents Stronger

·
11 authors

Submitted by

tianyilt

MOVA: Towards Scalable and Synchronized Video-Audio Generation

OpenMOSS-Team

Submitted by

Yu2020

Modality Gap-Driven Subspace Alignment Training Paradigm For Multimodal Large Language Models

·
15 authors

Submitted by

Jiafei1224

Recurrent-Depth VLA: Implicit Test-Time Compute Scaling of Vision-Language-Action Models via Latent Iterative Reasoning

allenai

Ai2

Submitted by

bhavul

AIRS-Bench: a Suite of Tasks for Frontier AI Research Science Agents

Submitted by

taesiri

InternAgent-1.5: A Unified Agentic Framework for Long-Horizon Autonomous Scientific Discovery

InternScience

Submitted by

utdawn

LLaDA2.1: Speeding Up Text Diffusion via Token Editing

inclusionAI

Submitted by

zoeyuchao

RLinf-USER: A Unified and Extensible System for Real-World Online Policy Learning in Embodied AI

RLinf

Submitted by

Canyu

Alleviating Sparse Rewards by Modeling Step-Wise and Long-Term Sampling Effects in Flow-Based GRPO

zju

Zhejiang University

Submitted by

Huan0825

Towards Agentic Intelligence for Materials Science

·
21 authors

Submitted by

mickyhimself

GEBench: Benchmarking Image Generation Models as GUI Environments

stepfun-ai

2

Submitted by

Zijie-chen

Improving Data and Reward Design for Scientific Reasoning in Large Language Models

microsoft

Submitted by

THUdyh

Demo-ICL: In-Context Learning for Procedural Video Knowledge Acquisition

mmlab-ntu

Submitted by

XaiverZ

Learning Query-Aware Budget-Tier Routing for Runtime Agent Memory

NanyangTechnologicalUniversity

Nanyang Technological University

Submitted by

AndrewZeng

LOCA-bench: Benchmarking Language Agents Under Controllable and Extreme Context Growth

HKUST NLP Group

Submitted by

yutaozhu94

GISA: A Benchmark for General Information-Seeking Assistant

RUC

Renmin University of China

Submitted by

March07

TermiGen: High-Fidelity Environment and Robust Trajectory Synthesis for Terminal Agents

ucsbnlp

UC Santa Barbara NLP Group

Submitted by

yiye2023

AgentCPM-Report: Interleaving Drafting and Deepening for Open-Ended Deep Research

openbmb

Submitted by

jieyuz2

Theory of Space: Can Foundation Models Construct Spatial Beliefs through Active Exploration?

·
14 authors

Submitted by

sleetwang6

WorldCompass: Reinforcement Learning for Long-Horizon World Models

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

RTT1

LatentChem: From Textual CoT to Latent Thinking in Chemical Reasoning

·
19 authors

Submitted by

Jiang-nan

Context Compression via Explicit Information Transmission

KingsCollegeLondon

King's College London

2

Submitted by

XingweiT

Fundamental Reasoning Paradigms Induce Out-of-Domain Generalization in Language Models

·
7 authors

Submitted by

d7chong

NanoQuant: Efficient Sub-1-Bit Quantization of Large Language Models

SamsungResearch

Samsung Research

3

Submitted by

jiwonsong

RelayGen: Intra-Generation Model Switching for Efficient Reasoning

SNU-VLSI

Seoul National University VLSI Lab

Submitted by

peterljq

Reliable and Responsible Foundation Models: A Comprehensive Survey

CarnegieMellonU

Carnegie Mellon University

Submitted by

taesiri

How2Everything: Mining the Web for How-To Procedures to Evaluate and Improve LLMs

allenai

Ai2

Submitted by

Shoubin

When and How Much to Imagine: Adaptive Test-Time Scaling with World Models for Visual Spatial Reasoning

UNC-ChapelHill

University of North Carolina at Chapel Hill

Submitted by

changdae

Thinking Makes LLM Agents Introverted: How Mandatory Thinking Can Backfire in User-Engaged Agents

uw-madison

University of Wisconsin - Madison

2

Submitted by

taesiri

Rolling Sink: Bridging Limited-Horizon Training and Open-Ended Testing in Autoregressive Video Diffusion

adobe-research

Submitted by

BigDong

Data Science and Technology Towards AGI Part I: Tiered Data Management

openbmb

Submitted by

bruno888

CodeCircuit: Toward Inferring LLM-Generated Code Correctness via Attribution Graphs

UIUC-CS

University of Illinois at Urbana-Champaign

Submitted by

Weidong-Huang

Towards Bridging the Gap between Large-Scale Pretraining and Efficient Finetuning for Humanoid Control

·
6 authors

Submitted by

RuijieZhu

MotionCrafter: Dense Geometry and Motion Reconstruction with a 4D VAE

TencentARC

ARC Lab, Tencent PCG

Submitted by

Wesleythu

WildReward: Learning Reward Models from In-the-Wild Human Interactions

THU-KEG

Knowledge Engineer Group @ Tsinghua University

Submitted by

Weidong-Huang

ECO: Energy-Constrained Optimization with Reinforcement Learning for Humanoid Walking

·
9 authors

Submitted by

KunChen

Flexible Entropy Control in RLVR with Gradient-Preserving Perspective

·
7 authors

Submitted by

Xinsheng-Wang

SoulX-Singer: Towards High-Quality Zero-Shot Singing Voice Synthesis

Soul-AILab

Submitted by

andre15silva

On Randomness in Agentic Evals

ASSERT-KTH

ASSERT | Research group at KTH Royal Institute of Technology

Submitted by

BayanDuygu

Optimal Turkish Subword Strategies at Scale: Systematic Evaluation of Data, Vocabulary, Morphology Interplay

turkish-nlp-suite

Turkish NLP Suite

Submitted by

larry2210

Echoes as Anchors: Probabilistic Costs and Attention Refocusing in LLM Reasoning

·
6 authors

Submitted by

JacobBITLABS

FlexMoRE: A Flexible Mixture of Rank-heterogeneous Experts for Efficient Federatedly-trained Large Language Models

Schneider-Kamp Lab

2

Submitted by

zhongshsh

Agent Skills: A Data-Driven Analysis of Claude Skills for Extending Large Language Model Functionality

·
3 authors

Submitted by

Peter2023HuggingFace

Learning-guided Kansa collocation for forward and inverse PDEs beyond linearity

·
5 authors

2

Submitted by

mjbuehler

GraphAgents: Knowledge Graph-Guided Agentic AI for Cross-Domain Materials Design

lamm-mit

LAMM: MIT Laboratory for Atomistic and Molecular Mechanics

2

Submitted by

jacquelinehe

Anchored Decoding: Provably Reducing Copyright Risk for Any Language Model

uwnlp

University of Washington NLP

Submitted by

Roytsai27

Concept-Aware Privacy Mechanisms for Defending Embedding Inversion Attacks

·
4 authors

Submitted by

puar-playground

KV-CoRE: Benchmarking Data-Dependent Low-Rank Compressibility of KV-Caches in LLMs

Tele-AI

Submitted by

ctma

Cost-Efficient RAG for Entity Matching with LLMs: A Blocking-based Exploration

·
5 authors

Submitted by

chaubeyG

AVERE: Improving Audiovisual Emotion Reasoning with Preference Optimization

UniversityofSouthernCalifornia

University of Southern California

Submitted by

taesiri

Aster: Autonomous Scientific Discovery over 20x Faster Than Existing Methods

·
1 authors

Submitted by

roi-pony

Col-Bandit: Zero-Shot Query-Time Pruning for Late-Interaction Retrieval

ibm-research

2

Submitted by

MagicalChair

CauScale: Neural Causal Discovery at Scale

Submitted by

sandyherho

dewi-kadita: A Python Library for Idealized Fish Schooling Simulation with Entropy-Based Diagnostics

ITB

Institut Teknologi Bandung

Submitted by

HanSolo9682

Reasoning-Augmented Representations for Multimodal Retrieval

·
6 authors

Submitted by

rhaldar97

f-GRPO and Beyond: Divergence-Based Reinforcement Learning Algorithms for General LLM Alignment

·
5 authors

Submitted by

liminho123

Statistical Learning Theory in Lean 4: Empirical Processes from Scratch

·
3 authors