Yanwei Fu

Full Professor@Fudan University·中国·OpenReview

研究方向

Image Inpainting · Robotic Grasping · Sparsity in Neural Network · learning based 3D reconstruction · Facial Analysis and Person Understanding · zero-shot learning and Attribute Learning · few-shot learning

Yanwei Fu

VTBench: Comprehensive Benchmark Suite Towards Real-World Virtual Try-on Models

Spatially Guided Training for Vision-Language-Action Model

Beyond 'Templates': Category-Agnostic Object Pose, Size, and Shape Estimation from a Single View

Why Reinforcement Fine-Tuning Enables MLLMs Preserve Prior Knowledge Better: A Data Perspective

TriVLA: A Triple-System-Based Unified Vision-Language-Action Model with Episodic World Modeling for General Robot Control

Conformal Reliability: A New Evaluation Metric for Conditional Generation

VTPerception-R1: Enhancing Multimodal Reasoning via Explicit Visual and Textual Perceptual Grounding

UniPruning: Unifying Local Metric and Global Feedback for Scalable Sparse LLMs

Brain Signal Rendering: Unifying EEG Video Representations for Subject-level Few-shot Learning

CoSMo-RL: Towards Trustworthy LMRMs via Joint Safety and Stability

Revisiting Large Language Model Pruning using Neuron Semantic Attribution

VidSplice: Towards Coherent Video Inpainting via Explicit Spaced Frame Guidance

UniDB: A Unified Diffusion Bridge Framework via Stochastic Optimal Control

PPMStereo: Pick-and-Play Memory Construction for Consistent Dynamic Stereo Matching

TP-MDDN: Task-Preferenced Multi-Demand-Driven Navigation with Autonomous Decision-Making

Towards Reliable and Holistic Visual In-Context Learning Prompt Selection

Domain-RAG: Retrieval-Guided Compositional Image Generation for Cross-Domain Few-Shot Object Detection

Bivariate Causal Discovery with Proxy Variables: Integral Solving and Beyond

3D StreetUnveiler with Semantic-aware 2DGS - a simple baseline

Adaptive Pruning of Pretrained Transformer via Differential Inclusions

StyleMaster: Towards Flexible Stylized Image Generation with Diffusion Models

Splitted Wavelet Differential Inclusion for neural signal processing

Enhancing Human Body Generation in Diffusion Models with Dual-Level Prior Knowledge

Combating inherent noise for direct preference optimization