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

Sep 30

Submitted by

jt-zhang

SLA: Beyond Sparsity in Diffusion Transformers via Fine-Tunable Sparse-Linear Attention

Tsinghua University

Submitted by

QbethQ

StableToken: A Noise-Robust Semantic Speech Tokenizer for Resilient SpeechLLMs

·
7 authors

2

Submitted by

fangwu97

Multiplayer Nash Preference Optimization

stanfordnlp

2

Submitted by

DogNeverSleep

OpenGPT-4o-Image: A Comprehensive Dataset for Advanced Image Generation and Editing

·
12 authors

2

Submitted by

MasterVito

Beyond the Exploration-Exploitation Trade-off: A Hidden State Approach for LLM Reasoning in RLVR

Tsinghua University

Submitted by

DogNeverSleep

RealUnify: Do Unified Models Truly Benefit from Unification? A Comprehensive Benchmark

·
26 authors

Submitted by

Yuyang-z

SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformer

nvidia

Submitted by

yuecao0119

Sequential Diffusion Language Models

OpenGVLab

Submitted by

shgao

Democratizing AI scientists using ToolUniverse

Harvard

Harvard University

Submitted by

Nicolas-BZRD

When Does Reasoning Matter? A Controlled Study of Reasoning's Contribution to Model Performance

When-Does-Reasoning-Matter

When Does Reasoning Matter ?

3

Submitted by

craigwu

Visual Jigsaw Post-Training Improves MLLMs

·
6 authors

Submitted by

taesiri

GSM8K-V: Can Vision Language Models Solve Grade School Math Word Problems in Visual Contexts

zju

Zhejiang University

Submitted by

adamdad

SparseD: Sparse Attention for Diffusion Language Models

·
5 authors

Submitted by

haoranhe

Random Policy Valuation is Enough for LLM Reasoning with Verifiable Rewards

·
7 authors

Submitted by

sienna223

EditScore: Unlocking Online RL for Image Editing via High-Fidelity Reward Modeling

BAAI

Beijing Academy of Artificial Intelligence

Submitted by

zjuxhl

EasySteer: A Unified Framework for High-Performance and Extensible LLM Steering

zju

Zhejiang University

Submitted by

Ningyu

Towards Personalized Deep Research: Benchmarks and Evaluations

OPPOer

OPPO

1

Submitted by

wenhu

VideoScore2: Think before You Score in Generative Video Evaluation

TIGER-Lab

Submitted by

taesiri

Rolling Forcing: Autoregressive Long Video Diffusion in Real Time

TencentARC

ARC Lab, Tencent PCG

Submitted by

taesiri

HunyuanImage 3.0 Technical Report

tencent

Submitted by

wenhu

Critique-Coder: Enhancing Coder Models by Critique Reinforcement Learning

TIGER-Lab

Submitted by

TeslaYang123

VGGT-X: When VGGT Meets Dense Novel View Synthesis

·
5 authors

Submitted by

Chuanyang-Jin

The Era of Real-World Human Interaction: RL from User Conversations

3

Submitted by

weizechen

From f(x) and g(x) to f(g(x)): LLMs Learn New Skills in RL by Composing Old Ones

·
10 authors

Submitted by

Ningyu

Scaling Generalist Data-Analytic Agents

Qwen

Qwen

2

Submitted by

Steven-Shaobo

Socratic-Zero : Bootstrapping Reasoning via Data-Free Agent Co-evolution

alibaba-inc

Submitted by

XINLI1997

WirelessMathLM: Teaching Mathematical Reasoning for LLMs in Wireless Communications with Reinforcement Learning

·
7 authors

2

Submitted by

LiamLian0727

Euclid's Gift: Enhancing Spatial Perception and Reasoning in Vision-Language Models via Geometric Surrogate Tasks

ZGCA

Zhongguancun Academy

Submitted by

wcy1122

MGM-Omni: Scaling Omni LLMs to Personalized Long-Horizon Speech

CUHK-CSE

The Chinese University of Hong Kong

Submitted by

Dingning

BRIDGE - Building Reinforcement-Learning Depth-to-Image Data Generation Engine for Monocular Depth Estimation

Submitted by

jaeikkim

MMPB: It's Time for Multi-Modal Personalization

snu-aidas

AI, Big Data, and System Laboratory

Submitted by

xcjthu

InfLLM-V2: Dense-Sparse Switchable Attention for Seamless Short-to-Long Adaptation

openbmb

2

Submitted by

zhangboguodong

Toward Effective Tool-Integrated Reasoning via Self-Evolved Preference Learning

RUC

Renmin University of China

Submitted by

changdae

Understanding Language Prior of LVLMs by Contrasting Chain-of-Embedding

huggingface-wisc

University of Wisconsin-Madison

Submitted by

bys0318

SIRI: Scaling Iterative Reinforcement Learning with Interleaved Compression

zai-org

Z.ai

Submitted by

taesiri

Pretraining Large Language Models with NVFP4

nvidia

Submitted by

li-qing

Efficient Multi-turn RL for GUI Agents via Decoupled Training and Adaptive Data Curation

bigai

Beijing Institute for General Artificial Intelligence

Submitted by

KunlunZhu

Where LLM Agents Fail and How They can Learn From Failures

UIUC-CS

University of Illinois at Urbana-Champaign

Submitted by

MatthieuZ

Rethinking Large Language Model Distillation: A Constrained Markov Decision Process Perspective

huawei-noah

HUAWEI Noah's Ark Lab

2

Submitted by

limuloo1999

Dynamic Experts Search: Enhancing Reasoning in Mixture-of-Experts LLMs at Test Time

·
4 authors

Submitted by

sundrops

GRPO-MA: Multi-Answer Generation in GRPO for Stable and Efficient Chain-of-Thought Training

·
5 authors

2

Submitted by

haonan3

From Harm to Help: Turning Reasoning In-Context Demos into Assets for Reasoning LMs

·
11 authors

Submitted by

samuelyeh

LUMINA: Detecting Hallucinations in RAG System with Context-Knowledge Signals

·
3 authors

Submitted by

Xianhang

Rethinking JEPA: Compute-Efficient Video SSL with Frozen Teachers

apple

2

Submitted by

desimfj

SCI-Verifier: Scientific Verifier with Thinking

·
11 authors

Submitted by

VsonicV

Evolution Strategies at Scale: LLM Fine-Tuning Beyond Reinforcement Learning

·
7 authors

Submitted by

JY-Young

Taming Masked Diffusion Language Models via Consistency Trajectory Reinforcement Learning with Fewer Decoding Step

Fudan-University

Fudan University

Submitted by

samuelyeh

Clean First, Align Later: Benchmarking Preference Data Cleaning for Reliable LLM Alignment

·
2 authors

Submitted by

zcai

DepthLM: Metric Depth From Vision Language Models

Submitted by

han-cai

DC-Gen: Post-Training Diffusion Acceleration with Deeply Compressed Latent Space

nvidia

Submitted by

guolinke

Hyperspherical Latents Improve Continuous-Token Autoregressive Generation

·
2 authors

Submitted by

yczhuang

AceSearcher: Bootstrapping Reasoning and Search for LLMs via Reinforced Self-Play

·
10 authors

Submitted by

XINLI1997

Local Success Does Not Compose: Benchmarking Large Language Models for Compositional Formal Verification

·
5 authors

2

Submitted by

SugerWu

MultiCrafter: High-Fidelity Multi-Subject Generation via Spatially Disentangled Attention and Identity-Aware Reinforcement Learning

·
7 authors

2

Submitted by

fushh7

LOVE-R1: Advancing Long Video Understanding with an Adaptive Zoom-in Mechanism via Multi-Step Reasoning

AlibabaTongyiLab

2

Submitted by

taesiri

IWR-Bench: Can LVLMs reconstruct interactive webpage from a user interaction video?

IWR-Bench

Submitted by

jmyang

Alignment through Meta-Weighted Online Sampling: Bridging the Gap between Data Generation and Preference Optimization

·
5 authors

1

Submitted by

zli999

PixelCraft: A Multi-Agent System for High-Fidelity Visual Reasoning on Structured Images

MicrosoftResearch

Microsoft Research

Submitted by

zhongwenxu

Cogito, Ergo Ludo: An Agent that Learns to Play by Reasoning and Planning

tencent

2

Submitted by

Cauthyyy

Advantage Weighted Matching: Aligning RL with Pretraining in Diffusion Models

adobe

Submitted by

weizhoudb

PARROT: A Benchmark for Evaluating LLMs in Cross-System SQL Translation

SJTU

Shanghai Jiao Tong University

Submitted by

charleslwang

MathBode: Frequency-Domain Fingerprints of LLM Mathematical Reasoning

ai-metrology-lab

AI Metrology Lab

Submitted by

HwanChang0106

ChatInject: Abusing Chat Templates for Prompt Injection in LLM Agents

Chung-AngUniversity

Chung-Ang University

Submitted by

hsi1032

Scalable GANs with Transformers

·
3 authors

2

Submitted by

SongzeLi

Learning Goal-Oriented Language-Guided Navigation with Self-Improving Demonstrations at Scale

OpenGVLab

Submitted by

ZihaoZhu

AdvChain: Adversarial Chain-of-Thought Tuning for Robust Safety Alignment of Large Reasoning Models

The Chinese University of Hongkong,Shenzhen

2

Submitted by

SteveZeyuZhang

UniVid: The Open-Source Unified Video Model

·
7 authors

Submitted by

lin-tan

TENET: Leveraging Tests Beyond Validation for Code Generation

lt-asset

Purdue ASSET Research Group | AI-Software Synergy

2

Submitted by

HelenMao

UniMIC: Token-Based Multimodal Interactive Coding for Human-AI Collaboration

CUC-MIPG

Multimedia Intelligent Processing Group in Communication University of China

3

Submitted by

versae

BOE-XSUM: Extreme Summarization in Clear Language of Spanish Legal Decrees and Notifications

bertin-project

2

Submitted by

xjh19972

ThermalGen: Style-Disentangled Flow-Based Generative Models for RGB-to-Thermal Image Translation

·
5 authors

Submitted by

s-jse

Detecting Corpus-Level Knowledge Inconsistencies in Wikipedia with Large Language Models

stanford-oval

Stanford Open Virtual Assistant Lab (OVAL)

1

Submitted by

robinzixuan

RHYTHM: Reasoning with Hierarchical Temporal Tokenization for Human Mobility

northwestern-university

Northwestern University

Submitted by

liboaccn

REMA: A Unified Reasoning Manifold Framework for Interpreting Large Language Model

·
8 authors

2

Submitted by

compulsi0n

Combinatorial Creativity: A New Frontier in Generalization Abilities

spiralworks

2

Submitted by

Franck-Dernoncourt

The Photographer Eye: Teaching Multimodal Large Language Models to See and Critique like Photographers

·
8 authors

Submitted by

jtlicardo

BPMN Assistant: An LLM-Based Approach to Business Process Modeling

·
3 authors

Submitted by

omidgh

ADAM: A Diverse Archive of Mankind for Evaluating and Enhancing LLMs in Biographical Reasoning

·
5 authors

2

Submitted by

pranamanam

TR2-D2: Tree Search Guided Trajectory-Aware Fine-Tuning for Discrete Diffusion

ChatterjeeLab

Programmable Biology Group

Submitted by

vaidehi99

Generalized Correctness Models: Learning Calibrated and Model-Agnostic Correctness Predictors from Historical Patterns

·
5 authors

2

Submitted by

alemiaschi

Charting a Decade of Computational Linguistics in Italy: The CLiC-it Corpus

·
8 authors

Submitted by

dipta007

Advancing Reference-free Evaluation of Video Captions with Factual Analysis

·
3 authors

1