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

Oct 20

Submitted by

ysa2001

Learning from Fleet Experience: A Collective Learning Paradigm for Robots

·
42 authors

Submitted by

hhhllan

MemT-1.0: Towards Memory-Native Evolution

MemTensor

Submitted by

bryanzhou008

OpenVLHarness: Real World Visual Agentic Tool-Use with Multimodal Memory

UCLA

University of California, Los Angeles

Submitted by

PLUS-WAVE

ScaleNav: Scalable Internet Video Pretraining for Language-Conditioned Local Navigation

·
10 authors

Submitted by

weizhepei

Self-Evolving Coding Agents via Harness-Policy Optimization

MicrosoftResearch

Microsoft Research

Submitted by

lifuguan

EgoXperience: Learning World-Grounded Egocentric Representations from Human Experience

ropedia-ai

Submitted by

lifuguan

Holi-World: Building Holistic World Intelligence from Human-Ego Experience

ropedia-ai

Submitted by

houyuanchen

MT Forcing: Towards Stable and Consistent Multi-Turn Image Editing with Autoregressive Diffusion

sensenova

Submitted by

xtma

Scaling Parameter-Efficient Fine-Tuning for Brain Encoding

mindlab-research

Submitted by

zhongweixie

CoworkEnv: Synthesizing Training Environments for Long-Horizon Cowork Agents

·
13 authors

Submitted by

CedarWang

OR-License: Evaluating LLM Agents as Real-World Operations Research Practitioners

HKUST-DSAIL

Data Science & Artificial Intelligence Lab at HKUST

Submitted by

yueyang2000

RoAR-VAE: Improving Latent Diffusability through Robust Prediction

Tsinghua-LeapLab

Tsinghua-LeapLab

Submitted by

HazelNam

AnyGoal: On World Model Generalization Beyond Expert Rollouts

galilai-group

Submitted by

shwu

SpecMRL: Redesign Matryoshka Representation Learning within Parametric Spectrum

·
6 authors

Submitted by

yulunliu

InfiniDream: Open-Ended Photorealistic 3D Worlds Through Procedural Generation

AlayaLab

Submitted by

XaiverZ

MemPilot: Orchestrating On-Demand Multimodal Memory Curation for LLM Agents

NanyangTechnologicalUniversity

Nanyang Technological University

Submitted by

anamariarobertap

Kronecker Delta Rule for Long Context Linear Attention

NX-AI

Submitted by

qizekun

Where is Physical RSI capability from? An Observational Scaling Law of LLM-as-policy

Galbot

Submitted by

yulunliu

Manga2Animate: An Agentic Framework for Faithful Manga Animation with Script and Storyboard Planning

NYCU

National Yang Ming Chiao Tung University

Submitted by

JiayuJeff

AdaExploreBench: Benchmarking LLM Agents for Proactive Exploration and Test-time Adaptation in Evolving Environments

Submitted by

Skywalker0410

AffordanceWAM: Affordance-Aware Joint World-Action Modeling for Robot Manipulation

PKU

Peking University

Submitted by

BBBBCHAN

Think Visually Across Space and Time: Grounded Reasoning for Video Segmentation

·
6 authors

Submitted by

BBBBCHAN

DeOverlay: A Benchmark and Source-Anchored Flow for Video Overlay Removal

ByteDance

Submitted by

rsanayei

Evidence-First Agentic Verification

ScaleAI

Submitted by

panzs19

HarnessPro: Recursive Harness Self-Improvement with Pareto Optimization for Performance-Cost Trade-offs

·
9 authors

Submitted by

Yanran21

State World Models: Execute the Rules, Generate the Pixels

Tsinghua-IVG

Submitted by

zhaoyian01

Frechet Forcing: Training Long-Horizon Streaming World Models from Short Clips

·
6 authors

Submitted by

nebulae09

BlenderWorld: Benchmarking Visual Agents for 3D Creation from Real-World References

Qwen

Submitted by

OliverLee

Self-Evaluating Recursive Agents

TongjiUniversity

Submitted by

SapnapHalo

Interlock: From Local Correctness to Task Completion in Long-Horizon Agents

meituan-longcat

1

Submitted by

mengtanZ

Learning to Discriminate Scene Structures Makes Self-Supervised Depth Learning Scalable

Tongji Unversity

Submitted by

cindysung

LIBERO-Industry: Benchmarking Robotic Manipulation under Industrial Process Constraints

IAIE-PhysicalAI

Physical AI Group, Institute of AI for Engineering (IAIE), Tongji University

Submitted by

Cyan666

HandsOn: A Large-Scale Simulation Benchmark for Generalizable Dexterous Manipulation

·
17 authors

Submitted by

Vincent-Fan

Factorization-Guided Training and Inference for Diffusion-Drafted Speculative Decoding

ByteDance

1

Submitted by

mingju0303

RSIFlow: Adaptive Self-Improvement Workflows with an Evolving Meta-Agent

Beihang

Beihang University

Submitted by

Jize1

World Model Interpretability: From Decoding Representations to Understanding Latent Worlds

·
6 authors

Submitted by

teawhite

ActionRoPE: action as coordinates for spatially consistent control in planar-camera game world models

·
4 authors

Submitted by

kinneso

ECHO: Efficient Causal Self-Recursion Beyond Auxiliary Multi-Token Prediction

ernie-research

Submitted by

berryccc1

HSPR: Recovering Software-Agent Benchmark Scores via Hierarchical Posterior Inference

SUDA

Soochow University

Submitted by

lwaekfjlk

Co-Evolving Acting and Forecasting for Language Agents

·
9 authors

Submitted by

ShuyangXiang

A Conditional No-Hair Theorem for Transformers: How Language Models Forget Their Prompt

·
1 authors

Submitted by

TK3105

UniAVC: Unified Audio-Video Creation via Progressive Training and Asymmetric Scheduling

·
9 authors

Submitted by

HwasikJeong

FineSplat: Fine Patches, Local Groups, Detailed Splats

·
7 authors

Submitted by

nullnonenilNULL

Vision-Language-Delta-Action: Learning Compact Deltas as Actionable Transition Interfaces for Robotic Manipulation

LoongForge

Submitted by

chen-yingfa

Competing Learning in Hybrid Architectures

TsinghuaNLP

Tsinghua NLP Group

Submitted by

AIGNLAI

Remembering Has a Price: Learning Memory Worth under Resource Constraints

nanjinguniv

Nanjing University

Submitted by

tksii

QD-Rec: Open-Ended Evolution of Recommendations via Quality-Diversity

SakanaAI

Submitted by

chengpingan

Empowering Edge Agents: A Systematic Analysis of Edge-Cloud Collaboration in LLM Agent Execution

NortheasternUniversity1923

Northeastern University

Submitted by

arnabdotorg

RealityBench: Simulated Environments to Evaluate a Vision Language Model's Ability to Document Human Physical Work

relantic

1

Submitted by

wzlllong

RoboValue: A Fine-Grained Sim-and-Real Benchmark for Unified Evaluation of Robotic Value Models

·
19 authors

Submitted by

Moreno1

Harnessing Textual Refusal Directions for Multimodal Safety

MHUGLab

Multimedia and Human Understanding Group

Submitted by

VahidJbrl

Policy Is Not Capability: Some Safety Benchmarks Measure Which Definition of "Unsafe" You Trained On

·
5 authors

Submitted by

keshik6

Text-to-Image Generative Modeling Without Language Pretraining

StanfordUniversity

Stanford University

Submitted by

jiho-e

Sparse Feature Policy Unlearning Mitigates State Hallucination in Vision-Language-Action Models

·
5 authors

1

Submitted by

timtsapras23

SPARC: Spoken Profile Authorship Recognition Corpus

USC

University of Southern California

1