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

Sep 14

Submitted by

ktwu01

Benchmark Radar: A Living Database and Search Engine for AI Benchmarks and Evaluation

CarnegieMellonU

Carnegie Mellon University

Submitted by

zongkey

Feyospace-v1: How the Cyber Mercury Seven Trained Frontier Cyber Models

feyospace

3

Submitted by

lhpku20010120

DataFlex-RL: An Evaluation Platform for RLVR Data Policies

PekingUniversity

Peking University

2

Submitted by

Jiafei1224

Breaking the Vision-Action Shortcut: Latent Interface Training for Generalizable Robotics Foundation Models

NationalUniversityofSingapore

National University of Singapore

Submitted by

taesiri

SAS: Simple Attention Sparsification via End-to-End Optimization of Context Ranking

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

Alan123

ReactHuman: A Physics-Grounded Benchmark for Human-Like Reactive Decision-Making in Embodied Multimodal LLMs

·
8 authors

2

Submitted by

Lpc206

COBRA-Skills: Contextual Bandit-Guided Evolution for Agent Skill Optimization

CUHKSZ

Chinese University of Hong Kong, Shenzhen

Submitted by

giantPanda0906

StepAudio 3 Gen Technical Report

·
71 authors

Submitted by

junboolee

Occamy-1.0: Open Pareto-frontier 35B Intelligence for Co-work

Accio-Lab

Submitted by

VennTum

PLC-DPO: Posterior Label Correction in Noisy and Ambiguous Preference Optimization

kaist-ai

Submitted by

ShilongZou

Pelican-Sim 1.0: A General World Model Simulator for Embodied Intelligence

X-Humanoid

Submitted by

Franck-Dernoncourt

Beyond Top-k Skill Retrieval: Diversity-Aware Skill Routing for LLM Agents

·
10 authors

Submitted by

Franck-Dernoncourt

Online Learning with LLM Experts from Limited Feedback

·
7 authors

Submitted by

varun-scale

Studying Without a Syllabus: Task-Agnostic Environment Preprocessing

ScaleAI

Submitted by

taesiri

SNAP3D: Physically Grounded 3D Parts for Assembly from a Single Image

·
5 authors

Submitted by

infinitylogesh

Ambient @ EgoProactive 2026 : Proactive Egocentric Assistance with Visually Grounded Supervision

ambient-intelligence-labs

Ambient Intelligence

Submitted by

infinitylogesh

Ambient @ EgoLongQA 2026: Distilling Long-Video perception into a Sub-2B Model

ambient-intelligence-labs

Ambient Intelligence

Submitted by

PardisTaghavi

ActionSplice: In-Flight Action Editing for Interactive World Models

TexasAMUniversity

Texas A&M University

Submitted by

adititiwari19

Competence-Gated Pooling of Language Models and Priors for Event Forecasting

UIUC-CS

University of Illinois at Urbana-Champaign

Submitted by

YHY-Test

TRACE: Trajectory-robust Admission with Evidence Ordering for Efficient GUI Agents

·
6 authors

Submitted by

kunato

How Far Can Synthetic Data Take Thai OCR?

wayu-ai

Submitted by

adititiwari19

Feature Recovery for Object Understanding After Irreversible Fire Damage

UIUC-CS

University of Illinois at Urbana-Champaign

Submitted by

sebasmos

Towards a Deterministic Math Solver for Clinical Language Models

MIT

Massachusetts Institute of Technology

Submitted by

SteveZeyuZhang

ReMoMask-2: Latent Retrieval-Augmented Masked Motion Generation

PekingUniversity

Peking University

Submitted by

SteveZeyuZhang

MobileVLA-R1 2.0: RL-Enhanced Reasoning for Mobile Robot Control

PekingUniversity

Peking University

Submitted by

kunato

Building and Evaluating Fixed-Voice Thai TTS from Synthetic Speech

wayu-ai