Xiangbo Shu

Bayesian Primitive Distributing for Compositional Zero-shot Learning

Toward Spatial Intelligence: A Unified Self-supervised Framework for 3D Representation Learning from Unposed Multi-View Images

Routing-Deconstructed LoRA in Federated Fine-Tuning

Cognitive-Awakening Chain-of-Surgery for Compositional Zero-Shot Surgical Triplet Recognition

CODA: Commonsense-Driven Autoregressive Human Interaction Generation

Probabilistic Prototype Generation Network for Cross-Domain Few-Shot Semantic Segmentation

IMAGEdit: Let Any Subject Transform

LGR: Local Geometric Refinement in High-Fidelity Surgical Scene Reconstruction

OmniGaze: Reward-inspired Generalizable Gaze Estimation in the Wild

Vision-centric Token Compression in Large Language Model

Learning Clustering-based Prototypes for Compositional Zero-Shot Learning

You Only Communicate Once: One-shot Federated Low-Rank Adaptation of MLLM

Spatio-temporal Decoupled Knowledge Compensator for Few-Shot Action Recognition

TrackMamba: Mamba-Transformer Tracking