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 10

Submitted by

JokerJan

Beyond Simply Environment Scaling: Designing Effective Environment Distributions for Multimodal Agent Learning

CASIA

Chinese Academic of Science Institute of Automation

Submitted by

JokerJan

SFT Conflicts, RL Coexists: A Theoretical and Empirical Analysis of Multi-Task Learning for LLMs

CASIA

Chinese Academic of Science Institute of Automation

Submitted by

dkliang

SimWAM: A Simple World Action Model for End-to-End Autonomous Driving

H-EmbodVis

Submitted by

gatilin

YOLO-PEFT: Parameter-Efficient Fine-Tuning on YOLO Family

tencent

1

Submitted by

hkuzxc

StreamArena: Toward Continuous, Interactive, and Long-Horizon Agentic Streaming Video Understanding

XiaohongshuAI

Submitted by

ParadiseYu

Reinforcement Learning with Evolving Rubrics as Rewards for Audio Reasoning

UMCP

University of Maryland College Park

Submitted by

Atmyre

When Activation Oracles Learn Not to Read: Concept-Specific Blind Spots in Fine-Tuned Oracles

QMUL

Queen Mary University of London

Submitted by

Iker

Efficient Knowledge Distillation for LLMs: Offline Top-K Logits and a Fused Chunked KL Loss

MultiverseComputingCAI

Multiverse Computing

Submitted by

DurYi

Uncertainty-Aware World Model for Aerial Image-Goal Navigation

Tsinghua University

Submitted by

Haon-Chen

Douyin Multimodal Embedding Model Technical Report

ByteDance

1

Submitted by

JusperLee

Relevant but Incomplete: Referential Dangling as a Paradigm-Level Failure Mode in Hard Prompt Compression

Tsinghua University

Submitted by

taesiri

Modular TTT: Rethinking Test-Time Training as Composable Modules

ByteDance-Seed

Submitted by

taesiri

Skaling: Chinchilla's Exponents Meet Kaplan's Coupling

Submitted by

taesiri

Addressable Memory for Video World Models

nvidia

Submitted by

AScheinker

Round-Trip Consistency: Bidirectional Diffusion Models Can Predict Their Own Rollout Errors

·
1 authors

Submitted by

taesiri

The Optimizer Is the Agent: Reasoning-Driven Search across Prompts, Programs, and ML Workflows

·
8 authors

Submitted by

taesiri

Characterizing the Quality Profile of AI-Generated C++ in Production

google

Submitted by

nick-sh-oh

Small Foundation Models of Human Cognition and Behaviour

socius

Submitted by

sci-m-wang

Do AI Personas Grow? Analyzing and Benchmarking Personality Evolution in LLM Agents After Life Events

kinamind

Submitted by

Junzhuo

When Privileged Guidance Misaligns: State-Matched Routing and Contextualized Self-Distillation for Multi-Turn Agents

·
4 authors

Submitted by

onnookk

Zero Gap Is Not Restoration: Stratified Per-Question Probability Evaluation and Step-wise Mitigation of Benchmark Contamination

zju

Zhejiang University

Submitted by

xuan269

PrivacyPeek: Auditing What LLM-Based Agents Acquire, Not Just What They Say

·
8 authors

Submitted by

sci-m-wang

Can MLLMs Decode the Creative Leap? Introducing C4 for Cross-Concept Understanding

kinamind

1

Submitted by

hanxunh

Adversarial Attacks for Good: A Survey of Proactive Protection across the Visual Content Lifecycle

·
17 authors

Submitted by

Guan123

FATE: Frame-Level Audio-Visual Temporal Embedding

RUC

Renmin University of China

Submitted by

zhehuderek

Capek 0.5: An Execution-Centric Vision-Language Model for Embodied Intelligence

Submitted by

Jiaha0Hu4ng

OneEmo: A Unified Multimodal Reasoning Model for Emotion Perception, Understanding, and Interaction

·
6 authors

Submitted by

CamilaR20

Towards Interpretable Foundation Models for Retinal Fundus Images

UniTuebingen

Eberhard Karls Universität Tübingen