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

Mar 12

Submitted by

Lingaaaaaaa

OpenClaw-RL: Train Any Agent Simply by Talking

princeton-ai

Princeton AI Lab

Submitted by

xihc-ucb

Flash-KMeans: Fast and Memory-Efficient Exact K-Means

Berkeley

Submitted by

KangsanKim71

MA-EgoQA: Question Answering over Egocentric Videos from Multiple Embodied Agents

kaist-ai

Submitted by

parishadbehnam

LLM2Vec-Gen: Generative Embeddings from Large Language Models

McGill-NLP

McGill NLP Group

Submitted by

q-rz

ReMix: Reinforcement routing for mixtures of LoRAs in LLM finetuning

metaresearch

3

Submitted by

Yiran0924

In-Context Reinforcement Learning for Tool Use in Large Language Models

NationalUniversityofSingapore

National University of Singapore

Submitted by

jiyeonkim

Can Large Language Models Keep Up? Benchmarking Online Adaptation to Continual Knowledge Streams

kaist-ai

Submitted by

xyingzhang

RetroAgent: From Solving to Evolving via Retrospective Dual Intrinsic Feedback

Submitted by

YuyaoGe

Prism-Δ: Differential Subspace Steering for Prompt Highlighting in Large Language Models

ict-cas

Institute of Computing Technology, Chinese Academy of Sciences

Submitted by

TongkunGuan

CodePercept: Code-Grounded Visual STEM Perception for MLLMs

Qwen

Submitted by

taesiri

V_{0.5}: Generalist Value Model as a Prior for Sparse RL Rollouts

meituan-longcat

Submitted by

taesiri

Just-in-Time: Training-Free Spatial Acceleration for Diffusion Transformers

·
3 authors

Submitted by

shwu22

RbtAct: Rebuttal as Supervision for Actionable Review Feedback Generation

yale-nlp

Submitted by

Luckyyy

Bootstrapping Exploration with Group-Level Natural Language Feedback in Reinforcement Learning

rednote-hilab

Submitted by

AviadDahan

ID-LoRA: Identity-Driven Audio-Video Personalization with In-Context LoRA

tau

Tel Aviv University

Submitted by

MiaTiancai

UniCom: Unified Multimodal Modeling via Compressed Continuous Semantic Representations

Tencent-Hunyuan

Tencent Hunyuan

1

Submitted by

mi2547nyu

Causal Concept Graphs in LLM Latent Space for Stepwise Reasoning

newyorkuniversity

New York University

Submitted by

nthngdy

Lost in Backpropagation: The LM Head is a Gradient Bottleneck

cornell

Cornell University

Submitted by

yangxw

Hindsight Credit Assignment for Long-Horizon LLM Agents

·
10 authors

1

Submitted by

genjib

V2M-Zero: Zero-Pair Time-Aligned Video-to-Music Generation

adobe-research

1

Submitted by

taesiri

CLIPO: Contrastive Learning in Policy Optimization Generalizes RLVR

·
9 authors

Submitted by

susunghong

COMIC: Agentic Sketch Comedy Generation

University of Washington

Submitted by

taesiri

Code-Space Response Oracles: Generating Interpretable Multi-Agent Policies with Large Language Models

deepmind

Submitted by

Harry4582

EmboAlign: Aligning Video Generation with Compositional Constraints for Zero-Shot Manipulation

northwestern-university

Northwestern University

1

Submitted by

zhiyuandaily

Any to Full: Prompting Depth Anything for Depth Completion in One Stage

·
7 authors

1

Submitted by

Jingbiao

According to Me: Long-Term Personalized Referential Memory QA

CambUni

University of Cambridge

Submitted by

Yuan-avs

StyleVLA: Driving Style-Aware Vision Language Action Model for Autonomous Driving

·
7 authors