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

Nov 10

Submitted by

cherubicxn

Learning Heterogeneous Tactile Representation for General Dexterous Manipulation

robbyant

Submitted by

Juanxi

WorldFoundry: Towards Unified Infrastructure for World Intelligence

OpenEnvisionLab

Submitted by

weidawang

Sage-Taiyi: A Metacognitive Architecture for Scientific Discovery

Shanghai-AI-Laboratory

Shanghai AI Laboratory

Submitted by

EliYuan00

Reframe: Observation-Aligned Multi-Teacher On-Policy Distillation for Agentic Visual Reasoning

NJU

Nanjing University

Submitted by

BradNLP

Flide: Accelerating Decision Inference with Exact State Forking

NationalUniversityofSingapore

National University of Singapore

Submitted by

ubowang

ReproBridge: Building Reproducibility into ML Research Artifacts

google

Submitted by

djalexj

Longtail-SWE: High-Quality SWE Tasks for Underrepresented Languages

nebius

Submitted by

yuchuantian

InfraRSI: Harnessing Recursive Self-Improvement for Training Infrastructure

TokenRhythm

Submitted by

zomss

SIRIS: Can LLMs Tell When Simulation Errors Matter in AI-Driven Systems Research

MicrosoftResearch

Microsoft Research

2

Submitted by

Benyucong

How Much Training-Inference Mismatch Can On-Policy Distillation Tolerate?

·
8 authors

Submitted by

An998

TRIO: Task Refinement with Individual Optimization for Retrieval-Augmented Generation

uts-research

University of Technology Sydney

Submitted by

Delcher

FewStepRL: Fine-tuning Few-step Diffusion Models with Reinforcement Learning

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

jama1017

MotionGraphicsBench: Evaluating Motion Graphics Generation with Programmatic Verification

figma-ai

Submitted by

KhanCold

AutoVend: Benchmarking LLM Agents on Long-Horizon Autonomous Vending Operations

·
10 authors

Submitted by

yuchuantian

Auriga Technical Report: A Task-Agnostic Agent Harness for Zero-Shot Embodied Manipulation

·
21 authors

Submitted by

ngyouyang

Training Vision Transformers to Focus: Emergence of Spatial Attention Head Specialization

·
3 authors

Submitted by

pranamhan

FlyHash for LLM Agent Memory: An Empirical Validation and a New Application

·
2 authors

Submitted by

Bryceee

SkinMind: Equipping Agentic RAG with Human-Inspired Graph Probing for Reliable Dermatology AI

antgroup

Submitted by

KuOnoda

How Many Exposures Does a World Model Need? Learning Unseen Game Mechanics at Test Time

third-intelligence

Third Intelligence, Inc.

Submitted by

xiangjuntang

JoVox: Joint Hierarchical Flow Matching for Lightweight Voxel Generation

KAUST

King Abdullah University of Science and Technology

Submitted by

wdlctc

onóna: Twin Embeddings with Tied Unembeddings

·
7 authors

Submitted by

Dyluhn

RidgeFill: Training-Free Late-Layer Cache Fill for Faster Prefill

TheCacheIsALie

Approximate Science

Submitted by

alon-albalak

Creative Priming: Decoupling Process from Product to Escape the Creative Paradox

LilaSciences

Submitted by

FreDERicia

SWAM: Scaling World Action Models with Cross-Modal Alignment

SJTU

Shanghai Jiao Tong University

Submitted by

nielsr

InstantHMR : Real-Time Human Mesh Recovery on Your Phone

Submitted by

Aboriginer

Spatial Harness: Grounding Spatial Reasoning in Executable World States with Physics Engines

TMLR-Group-HF

Submitted by

gurayozgur

Stochastic Deep Supervision for Gallery-Consistent Early-Exit Face Recognition

FraunhoferIGD

Submitted by

emilioferrara

All Quiet on the Neural Front: Language Models Do Not Report Injected Pain States

·
1 authors

Submitted by

ForeverBlue

PhysioWave-T: Token-Efficient Wavelet Foundation Models for Biosignals

NanyangTechnologicalUniversity

Nanyang Technological University

Submitted by

kevinlwei

FraudBench: A Legal Evaluation of AI Deception

·
20 authors

Submitted by

yaojin17

Click3R: Interactive Stereo 3D Reconstruction with Sparse Correspondence Clicks

uva-cv-lab

UVA Computer Vision Lab

Submitted by

NestaM

LCM-Promo: An LLM-Based Uplift Model for Promotion Allocation

ProsusAI

1

Submitted by

dylan-ma

Knowledge Graph-based Retrieval-Augmented Generation for Schema Matching

·
4 authors

Submitted by

andleb

Energy-Tweedie: Score meets Score, Energy meets Energy

umichigan

University of Michigan

Submitted by

ryanxhr

Horizon-Scalable Q-Learning

UTEXAS

University of Texas at Austin

Submitted by

swa2n

e-means: outperforming k-means (in one-tenth the epochs)

meta