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

Jun 8

Submitted by

shwu

Your UnEmbedding Matrix is Secretly a Feature Lens for Text Embeddings

·
6 authors

Submitted by

hamzzi

SoCRATES: Towards Reliable Automated Evaluation of Proactive LLM Mediation across Domains and Socio-cognitive Variations

DISLab

Data Intelligence System Lab

Submitted by

darlednik

GENEB: Why Genomic Models Are Hard to Compare

·
3 authors

Submitted by

BoJack

MMAE: A Massive Multitask Audio Editing Benchmark

·
38 authors

Submitted by

lyabc

AnchorWorld: Embodied Egocentric World Simulation with View-based Evolution Customization

KlingTeam

Submitted by

superGong

Direct 3D-Aware Object Insertion via Decomposed Visual Proxies

·
9 authors

Submitted by

hba123

Robots Need More than VLA and World Models

·
9 authors

Submitted by

dongsheng

When Tools Fail: Benchmarking Dynamic Replanning and Anomaly Recovery in LLM Agents

baidu

Submitted by

taesiri

OpenSkill: Open-World Self-Evolution for LLM Agents

·
11 authors

Submitted by

Yummytanmo

SubtleMemory: A Benchmark for Fine-Grained Relational Memory Discrimination in Long-Horizon AI Agents

·
7 authors

Submitted by

Insta360-Research

UniSHARP: Universal Sharp Monocular View Synthesis

·
7 authors

Submitted by

qizekun

LIMMT: Less is More for Motion Tracking

GalaxyGeneralRobotics

1

Submitted by

marinero4972

Watch, Remember, Reason: Human-View Video Understanding with MLLMs

PekingUniversity

Peking University

Submitted by

taesiri

dots.tts Technical Report

·
9 authors

Submitted by

AmirhosseinAbaskohi

UnpredictaBench: A Benchmark for Evaluating Distributional Randomness in LLMs

UBC-VisionLanguage

Vision and NLP Group at University of British Columbia

Submitted by

nitay

LLM Explainability with Counterfactual Chains and Causal Graphs

Technion

Technion Israel institute of technology

Submitted by

ChaimZhu

Thinking with Imagination: Agentic Visual Spatial Reasoning with World Simulators

liu-vision-group

HKU LIU Vision Group

Submitted by

dnwjddl

Physics in 2-Steps: Locking Motion Priors Before Visual Refinement Erases Them

·
6 authors

Submitted by

chengtan9907

PaperFlow: Profiling, Recommending, and Adapting Across Daily Paper Streams

opendatalab

Submitted by

taesiri

Socratic-SWE: Self-Evolving Coding Agents via Trace-Derived Agent Skills

·
8 authors

Submitted by

VigneshHexo

SIA: Self Improving AI with Harness & Weight Updates

Submitted by

Sterzhang

SPACENUM: Revisiting Spatial Numerical Understanding in VLMs

·
7 authors

Submitted by

JonnyYu828

Stream3D-VLM: Online 3D Spatial Understanding with Incremental Geometry Priors

Tencent-Hunyuan

Tencent Hunyuan

Submitted by

aboots

Almieyar-Oryx-BloomBench: A Bilingual Multimodal Benchmark for Cognitively Informed Evaluation of Vision-Language Models

QCRI

Qatar Computing Research Institute

Submitted by

mingju0303

HarnessForge: Joint Harness and Policy Evolution for Adaptive Agent Systems

·
5 authors

Submitted by

rish-1086

Reinforcement Learning from Rich Feedback with Distributional DAgger

UniversityofSouthernCalifornia

University of Southern California

Submitted by

adivekar

When Gradients Collide: Failure Modes of Multi-Objective Prompt Optimization for LLM Judges

Indian Institute of Technology Jodhpur

Submitted by

OzTianlu

Entropy as a Structural Prior: How a Log-Barrier on DiT Belief Space Drives Musical Diversity and Development

ReasoningKingdom

Reasoning Kingdom

1

Submitted by

zli12321

A Cookbook of 3D Vision: Data, Learning Paradigms, and Application

brownu

Brown University

Submitted by

prateeksikdar

LayerRoute: Input-Conditioned Adaptive Layer Skipping via LoRA Fine-Tuning for Agentic Language Models

Accenture

Submitted by

ArvinZhuang

Towards Retrieving Interaction Spaces for Agentic Search

·
5 authors

Submitted by

taesiri

Streaming Video Generation with Streaming Force Control

·
6 authors

Submitted by

chungimungi

ECI_{sem}: Semantic Residual Effective Contrastive Information for Evaluating Hard Negatives

·
3 authors

1

Submitted by

shubhamprshr

Data-Efficient Autoregressive-to-Diffusion Language Models via On-Policy Distillation

divelab

DIVE Lab, Texas A&M University

Submitted by

maximegmd

Compress-Distill: Reasoning Trace Compression for Efficient Knowledge Distillation

UCLouvain

Université Catholique de Louvain

Submitted by

weikaih

Imaginative Perception Tokens Enhance Spatial Reasoning in Multimodal Language Models

University of Washington

Submitted by

pcczc15

Measuring Model Robustness via Fisher Information: Spectral Bounds, Theoretical Guarantees, and Practical Algorithms

XJTLU-AI

XJTLU Student AI Lab

Submitted by

hexi222

Parametric Social Identity Injection and Diversification in Public Opinion Simulation

Tsinghua University

Submitted by

Andrew0425

Towards Human-Like Interactive Speech Recognition With Agentic Correction and Semantic Evaluation

X-LANCE

SJTU Cross Media Language Intelligence Lab

Submitted by

linasmn

CORE: Contrastive Reflection Enables Rapid Improvements in Reasoning

StanfordUniversity

Stanford University

Submitted by

imranraad

Empirical Study on the Characteristics and Evolution of AI-usage in GitHub Repositories: Evidence from Code Comments

·
3 authors

Submitted by

PearlLeeStudio

How Far Can Chord-Symbol Time-Series Adaptation Carry Genre Identity? Capabilities and Boundaries in Multi-Genre Chord-Symbol Modeling

·
1 authors

Submitted by

taesiri

WorldBench: A Challenging and Visually Diverse Multimodal Reasoning Benchmark

·
12 authors

Submitted by

zarif98sjs

Critic-R: Improving Agentic Search using Instruction-tuned Retrievers with Natural Language Introspective Feedback

University of Massachusetts Amherst

Submitted by

barpitf

Augmenting Attention with Exponentially Decaying Memory Improves Query-Aware KV Sparsity

·
2 authors