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 6

Submitted by

taesiri

ToolArtist: Tool-Using Unified Multimodal Models for Agentic Image Generation

·
8 authors

Submitted by

ruiye-sjtu

ABSeeker: Training Long-Horizon Search Agents via Answer-Backtracked Credit Assignment

SJTU

Shanghai Jiao Tong University

Submitted by

Yushi98

The Personalization Mirage: How LLMs Fabricate User Profiles, and Why Self-Monitoring Misleads

Submitted by

Junlinh

Towards Physics of Multimodal Pretraining: Knowledge Flow, Modality Synergy, Early Unification, and Recipes

Submitted by

Ningyu

OneDayAgent: Towards a Long-Horizon Harness for Autonomous Agents

zjunlp

Submitted by

hiyouga

GDPevo: Evaluating Agent Self-Evolution on Real Business Tasks

PrismShadow

Submitted by

jenoj

When Teachers Mislead: Spurious-Signal-Aware On-Policy Distillation

ByteDance

Submitted by

Lingaaaaaaa

Toward Skill-Native LLMs: Skill Entropy for Benchmarking and Training Long-Horizon Reasoning

princetonu

Princeton University

Submitted by

Dasool

NOLLI: A Difficulty-Calibrated Puzzle Benchmark for Diagnosing the English-Korean Performance Gap

HAERAE-HUB

Submitted by

mistletoe111

AVE-Compass: Towards Holistic Evaluation for Audio-Video Editing Abilities

NJU-LINK

Submitted by

wwwyyy

Ego2Robot: Scalable Robot Data Synthesis from Egocentric Human Data

·
15 authors

Submitted by

Yushi98

When Memory Lies: An Empirical Study of Spatial Memory Staleness in VLM Agents

tencent

Submitted by

LZXzju

Distill Where You Fail: Recovering Learning Signals of Negative RL-Groups from Adaptive Teacher Guidance

·
12 authors

1

Submitted by

DogNeverSleep

FocusMem: Factorizing Content, Readout, and Trust in Latent GUI Memory

·
8 authors

1

Submitted by

taesiri

HelloWorld: Enabling Socially Interactive Characters in Video World Models

·
5 authors

Submitted by

MarcusGu

WorldCycle: Self-Verifiable Reinforcement Learning for Long-Horizon Video World Models

tencent

Submitted by

taesiri

K-EXAONE 2.0 Technical Report

LGAI-EXAONE

Submitted by

LPY

BridgeVLA++: A Data-Efficient, Generalizable, and Memory-Augmented Vision-Language-Action Framework for 3D Manipulation

·
13 authors

Submitted by

simingfu

Poly-OPD: Heterogeneous Multi-Teacher On-Policy Distillation for Capability-Selectable Flow Models

jingdong1

Submitted by

taesiri

OPD-V: Visual On-Policy Self-Distillation with Modality Balance

·
8 authors

Submitted by

SecureLLMSys

Agent Against Agent: An Agentic System for Automatic Prompt Injection Red Teaming

Pennsylvania State University

Submitted by

Marblueocean

UniWorld-View: Large-Baseline View Synthesis via Video Diffusion Models

·
6 authors

Submitted by

Goal-aweiluo

TriGlue: a Biology-Inspired Generative Model for Generating Molecular Glue-Induced Ternary Complex

·
5 authors

Submitted by

KhanCold

SKILL-KD: Contrastive Skill Distillation for LLM Agents

Zhejiang University

Submitted by

xuehang

Consistency-Driven Co-Evolution for Self-Supervised Cross-Representation Learning

·
6 authors

1

Submitted by

tomhu

Self-Evolving Coding Agents

NJUST1412

Nanjing University of Science and Technology