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 2

Submitted by

ZHZisZZ

dLLM: Simple Diffusion Language Modeling

Berkeley

Submitted by

Falcary

Enhancing Spatial Understanding in Image Generation via Reward Modeling

PKU

Peking University

Submitted by

hannayukhymenko

Recovered in Translation: Efficient Pipeline for Automated Translation of Benchmarks and Datasets

INSAIT-Institute

Institute for Computer Science, Artificial intelligence and Technology

Submitted by

taesiri

CUDA Agent: Large-Scale Agentic RL for High-Performance CUDA Kernel Generation

ByteDance-Seed

Submitted by

primecai

Mode Seeking meets Mean Seeking for Fast Long Video Generation

nvidia

2

Submitted by

djalexj

LK Losses: Direct Acceptance Rate Optimization for Speculative Decoding

nebius

2

Submitted by

Tyrannosaurus

CiteAudit: You Cited It, But Did You Read It? A Benchmark for Verifying Scientific References in the LLM Era

notredame

University of Notre Dame

Submitted by

Gigglingface

Compositional Generalization Requires Linear, Orthogonal Representations in Vision Embedding Models

·
3 authors

Submitted by

Yossilevii100

InfoNCE Induces Gaussian Distribution

Technion

Technion Israel institute of technology

Submitted by

Andrew613

Accelerating Masked Image Generation by Learning Latent Controlled Dynamics

·
11 authors

Submitted by

taesiri

Memory Caching: RNNs with Growing Memory

·
6 authors

Submitted by

dddraxxx

Ref-Adv: Exploring MLLM Visual Reasoning in Referring Expression Tasks

Northeastern

Northeastern University

Submitted by

ChurchillQAQ

LongVideo-R1: Smart Navigation for Low-cost Long Video Understanding

UCAS

Submitted by

taesiri

SenCache: Accelerating Diffusion Model Inference via Sensitivity-Aware Caching

·
2 authors

Submitted by

dyang39

Shared Nature, Unique Nurture: PRISM for Pluralistic Reasoning via In-context Structure Modeling

UCSC

University of California, Santa Cruz

Submitted by

tomhu

CL4SE: A Context Learning Benchmark For Software Engineering Tasks

PolyUHK

The Hong Kong Polytechnic University

Submitted by

taesiri

Vectorizing the Trie: Efficient Constrained Decoding for LLM-based Generative Retrieval on Accelerators

·
13 authors

Submitted by

laitifranz

How to Take a Memorable Picture? Empowering Users with Actionable Feedback

·
4 authors

Submitted by

Viani

DUET-VLM: Dual stage Unified Efficient Token reduction for VLM Training and Inference

amd

Submitted by

taesiri

DLEBench: Evaluating Small-scale Object Editing Ability for Instruction-based Image Editing Model

·
8 authors