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

Aug 21

Submitted by

taesiri

EnvHarness: Awakening Static Worlds for Agent Learning

google

Submitted by

KouShi2

FACET: Preserving Source Intent and Executable State in Terminal Task Synthesis

ustc

university of science and technology of china

Submitted by

krahets

4DAnyone: Create Anyone in 4D from a Casual Monocular Video

AntResearch

Submitted by

yxzwang

SWE-bench Science: Can Coding Agents Resolve Engineering Tasks in Science?

OpenMOSS-Team

Submitted by

DobyXu

WithEveryone: Unified Planning and Identity Grounding for Group Image Generation

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

mengru

MemTrapBench: Benchmarking Cognitive Traps in LLM Memory Use

zjunlp

Submitted by

EasonUwU

SkillEvo: Self-Renewing Evolution Gradients from Multi-Turn Interaction Feedback

tencent

6

Submitted by

asdfo123

ForgeWM: Progressive Causal Training for Few-Step Action-Conditioned Video World Models

ForgeWM

Submitted by

Silin-Chen

Repo0: Design-Driven Zero-to-All Code Generation

SJTU

Shanghai Jiao Tong University

2

Submitted by

aldjalkdf

FlashPrefill V2: Block-Sparse Prefill Attention for Long-Context LLM Serving

tencent

Submitted by

ayoubkirouane

Thinking in a Low-Resource Language: What SFT Builds, What RL Fixes, What Accuracy Cannot See

KIEFERSA

3

Submitted by

XiaofengAlg

Inject, Align, Recover: Staged Post-Training for Retrieval-Free Document Knowledge Internalization

BAAI

Beijing Academy of Artificial Intelligence

2

Submitted by

taesiri

EXIMO: VLM Guided Exploration of VLA Policies

deepmind

Submitted by

TalentZHOU

Hierarchical Self-Improvement: A Framework for Task-Specific Evolvable Agent Harnesses

Submitted by

bezzam

Towards Quantifying Benchmark Optimization in ASR Models

HumeAI

Submitted by

J3rr1

τ_0-VLA: a Hierarchical Robot Foundation Model with World-Model-Guided Test-Time Computation

sii-research

Shanghai Innovation Institute

Submitted by

asteinh

TinyCast: Probabilistic Zero-Shot Forecasting with Computed Periodicity

raws-labs

Submitted by

yuhenghuang

NARU: A Benchmark for NARrative Evolution and Cultural Nuance Understanding in Japanese Extreme Long Video

utokyo-ai

The University of Tokyo

Submitted by

erjui

PolicyGuide: From Guarding One Action to Guiding the Whole Workflow for Policy-Compliant LLM Agents

kaist-ai

Submitted by

PahaII

Chain-of-Experience for Continual LLM Improvement

UCSC-VLAA

Submitted by

Muennighoff

The Embedder's Dilemma: LLMs Are Better, but at What Cost?

·
3 authors

Submitted by

lsamc

QuoteBench: How Matched Scores Can Hide Command-Path Failures

·
4 authors

Submitted by

JulienMERAND

GOAG: Generative and Object-Agnostic Grasp Planner for Dexterous Robotic Manipulation

·
4 authors

Submitted by

JulienMERAND

CoToGrasp: Contact-Topology-Conditioned Dexterous Grasp Synthesis via Canonical Workspace Learning

·
4 authors

Submitted by

hisoka94

Listening Forward: Next Patch Embedding Prediction Enables Scalable Audio Learners

ImperialCollegeLondon

Imperial College London

Submitted by

LeoYML

FlowEvo: Self-Evolving Agents through the Co-Evolution of Workflows and Executable Skills

SoutheastU

Southeast Univeristy