Area of research
Artificial Intelligence
Research interest
Research interests include Machine Learning and Data Classification, Domain Adaptation and Few-Shot Learning, Adversarial Robustness in Machine Learning, and Anomaly Detection Techniques and Applications.
AI-Accelerated Discovery of Electrocatalyst Materials
LaVin-DiT: Large Vision Diffusion Transformer
BadLabel: A Robust Perspective on Evaluating and Enhancing Label-Noise Learning
E2HQV: High-Quality Video Generation from Event Camera via Theory-Inspired Model-Aided Deep Learning
Quantization Aware Attack: Enhancing Transferable Adversarial Attacks by Model Quantization
DeepSolo: Let Transformer Decoder with Explicit Points Solo for Text Spotting
Recent Advances for Quantum Neural Networks in Generative Learning
HumanMAC: Masked Motion Completion for Human Motion Prediction
Graph Pooling for Graph Neural Networks: Progress, Challenges, and Opportunities
Point-Query Quadtree for Crowd Counting, Localization, and More
Combating Noisy Labels with Sample Selection by Mining High-Discrepancy Examples
ALIP: Adaptive Language-Image Pre-training with Synthetic Caption
BiCro: Noisy Correspondence Rectification for Multi-modality Data via Bi-directional Cross-modal Similarity Consistency
Joint Admission Control and Resource Allocation of Virtual Network Embedding via Hierarchical Deep Reinforcement Learning
Dynamics-aware loss for learning with label noise
CRIS: CLIP-Driven Referring Image Segmentation
Selective-Supervised Contrastive Learning with Noisy Labels
Killing Two Birds with One Stone: Efficient and Robust Training of Face Recognition CNNs by Partial FC
Instance-Dependent Label-Noise Learning with Manifold-Regularized Transition Matrix Estimation
Mutual Quantization for Cross-Modal Search with Noisy Labels
SimT: Handling Open-set Noise for Domain Adaptive Semantic Segmentation
Quantum noise protects quantum classifiers against adversaries
Robust early-learning: Hindering the memorization of noisy labels
International Conference on Learning Representations 2021cited by 124position: middle
Learnability of Quantum Neural Networks
HRSiam: High-Resolution Siamese Network, Towards Space-Borne Satellite Video Tracking
A Second-Order Approach to Learning with Instance-Dependent Label Noise
Transferable Coupled Network for Zero-Shot Sketch-Based Image Retrieval
Bridging the Gap Between Few-Shot and Many-Shot Learning via Distribution Calibration
Removing Adversarial Noise in Class Activation Feature Space
Why ResNet Works? Residuals Generalize