Dahua Lin

Achieving Olympia-Level Geometry Large Language Model Agent via Complexity Boosting Reinforcement Learning

RoboInter: A Holistic Intermediate Representation Suite Towards Robotic Manipulation

MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence

Beyond Fixed: Training-Free Variable-Length Denoising for Diffusion Large Language Models

From Pixels to Words -- Towards Native Vision-Language Primitives at Scale

STAR-Bench: Probing Deep Spatio-Temporal Reasoning as Audio 4D Intelligence

Visual Self-Refine: A Pixel-Guided Paradigm for Accurate Chart Parsing

DiCache: Let Diffusion Model Determine Its Own Cache

SIM-CoT: Supervised Implicit Chain-of-Thought

Advancing Complex Video Object Segmentation via Progressive Concept Construction

Learning Video Generation for Robotic Manipulation with Collaborative Trajectory Control

Scaling Large Vision-Language Model RL Training via Efficient Load Balancing

ChangingGrounding: 3D Visual Grounding in Changing Scenes

FlexLinearAttention: Compiling a Unified Abstraction into Scalable Kernels for Linear Attention

InterActHuman: Multi-Concept Human Animation with Layout-Aligned Audio Conditions

CODA: Coordinating the Cerebrum and Cerebellum for a Dual-Brain Computer Use Agent with Decoupled Reinforcement Learning

ScaleCap: Scalable Image Captioning via Dual-Modality Debiasing

CapRL: Stimulating Dense Image Caption Capabilities via Reinforcement Learning

OPV: Outcome-based Process Verifier for Efficient Long Chain-of-Thought Verification

SEAgent: Self-Evolving Computer Use Agent with Autonomous Learning from Experience

SPARK: Synergistic Policy And Reward Co-Evolving Framework

LONG-HORIZON REASONING AGENT FOR OLYMPIAD- LEVEL MATHEMATICAL PROBLEM SOLVING

MCPVerse: An Expansive, Real-World Benchmark for Agentic Tool Use

The Imitation Game: Turing Machine Imitator is Length Generalizable Reasoner

BoostStep: Boosting Mathematical Capability of Large Language Models via Step-aligned In Context Learning

BridgEAD: A Vision-Language Framework for Action Modeling in End-to-End Autonomous Driving

Phased DMD: Few-step Distribution Matching Distillation via Score Matching within Subintervals

Windtalkers: Watermarking Open-Source LLMs with Ciphered-Instruction

VideoRoPE: What Makes for Good Video Rotary Position Embedding?

Video World Models with Long-term Spatial Memory

HiFlow: Training-free High-Resolution Image Generation with Flow-Aligned Guidance

LOKI: A Comprehensive Synthetic Data Detection Benchmark using Large Multimodal Models

Predictive Inverse Dynamics Models are Scalable Learners for Robotic Manipulation

OmniCorpus: A Unified Multimodal Corpus of 10 Billion-Level Images Interleaved with Text

Imagine360: Immersive 360 Video Generation from Perspective Anchor

Semi-off-Policy Reinforcement Learning for Vision-Language Slow-Thinking Reasoning

3DTrajMaster: Mastering 3D Trajectory for Multi-Entity Motion in Video Generation

IDArb: Intrinsic Decomposition for Arbitrary Number of Input Views and Illuminations

Exploring the Limit of Outcome Reward for Learning Mathematical Reasoning

Mask-DPO: Generalizable Fine-grained Factuality Alignment of LLMs

LV-Eval: A Balanced Long-Context Benchmark with 5 Length Levels Up to 256K

Hierachical Balance Packing: Towards Efficient Supervised Fine-tuning for Long-Context LLM

Training Language Models to Critique with Multi-Agent Feedback

Mixing Expert Knowledge: Bring Human Thoughts Back To the Game of Go

OmniBal: Towards Fast Instruction-Tuning for Vision-Language Models via Omniverse Computation Balance

What are the Essential Factors in Crafting Effective Long Context Multi-Hop Instruction Datasets? Insights and Best Practices

RAR: Retrieving And Ranking Augmented MLLMs for Visual Recognition

SongGen: A Single Stage Auto-regressive Transformer for Text-to-Song Generation

MIA-DPO: Multi-Image Augmented Direct Preference Optimization For Large Vision-Language Models

LV-Eval: A Balanced Long-Context Benchmark with 5 Length Levels Up to 256K

Deciphering Cross-Modal Alignment in Large Vision-Language Models with Modality Integration Rate

SAM2Long: Enhancing SAM2 for Long Video Segmentation with a Training-Free Memory Tree

VEnhancer: Generative Space-Time Enhancement for Video Generation

Unearthing Large Scale Domain-Specific Knowledge from Public Corpora

OMNIBAL: TOWARDS FAST INSTRUCT-TUNING FOR VISION-LANGUAGE MODELS VIA OMNIVERSE COMPUTATION BALANCE

MxMoE: Mixed-precision Quantization for MoE with Accuracy and Performance Co-Design

Tailor3D: Customized 3D Assets Editing and Generation with Dual-Side Images

3DTopia-XL: Scaling High-quality 3D Asset Generation via Primitive Diffusion

DualFocus: Integrating Macro and Micro Perspectives in Multi-modal Large Language Models

BroadWay: Boost Your Text-to-Video Generation Model in a Training-free Way

SongComposer: A Large Language Model for Lyric and Melody Composition in Song Generation

Trustworthy Dataset Proof: Certifying the Authentic Use of Dataset in Training Models for Enhanced Trust

Bootstrap3D: Improving Multi-view Diffusion Model with Synthetic Data

PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reduction