Kai Han

Assistant Professor@The University of Hong Kong·中国香港·OpenReview

研究方向

vision language representation · foundation models · generative AI · large language models · semi-supervised learning · self-supervised learning · open-world learning · category discovery · deep learning · visual correspondence · representation learning · computer vision · 3D vision

Kai Han

PartCo: Part-Level Correspondence Priors Enhance Category Discovery

PhyCo: Physics-Consistent Learning of Implicit Constitutive Laws via Monocular Observations of 3D Gaussians

Prompt-SSLC: A Unified Framework for Dual Prompt-Augmented Semi-Supervised Sequential Leader Clustering in On-the-Fly Category Discovery

VEFlow: Training-free Text to Video Editing via Inversion-Free Video Editing Flow

MLLMs Need 3D-Aware Representation Supervision for Scene Understanding

Fin3R: Fine-tuning Feed-forward 3D Reconstruction Models via Monocular Knowledge Distillation

VaMP: Variational Multi-Modal Prompt Learning for Vision-Language Models

Wukong's 72 Transformations: High-fidelity Textured 3D Morphing via Flow Models

GSPN-2: Efficient Parallel Sequence Modeling

HiLo: A Learning Framework for Generalized Category Discovery Robust to Domain Shifts

BiGR: Harnessing Binary Latent Codes for Image Generation and Improved Visual Representation Capabilities

Panoptic Captioning: An Equivalence Bridge for Image and Text

Needle Threading: Can LLMs Follow Threads Through Near-Million-Scale Haystacks?

DebGCD: Debiased Learning with Distribution Guidance for Generalized Category Discovery

LLM-VTP: LLM-Reasoned Visual Token Pruning for Efficient Multi-Modal Video Understanding

SEAL: Semantic-Aware Hierarchical Learning for Generalized Category Discovery

AvatarGO: Zero-shot 4D Human-Object Interaction Generation and Animation

GAMEBOT: Gaming Arena for Model Evaluation - Battle of Tactics