Jan 5, 2025. Congrats to Xovee Xu and Wenxin Tai, for being selected to the Young Elite Scientists Sponsorship Program by CAST, Doctoral Student Special Plan. 

Nov 7, 2024. Congrats to Yi, Yanlong, Jian, Bin, Kai, and Jienan for awarding the National Scholarship, one of the highest honors for university students in China. 

July, 2024. Our paper "RRE: A relevance relation extraction framework for cross-domain recommender system at Alipay" is selected as the one of the 15 best paper candidates of ICME 2024. Congrats to Xovee and other co-authors.

Oct 23, 2023. Our lab won 2 runners-up and 3 third places in the Artificial Intelligence Challenge hosted by People's Daily Online, standing out among 852 teams from 176 schools. Congrats!

Trustworthy neural networks are the bedrock of responsible AI, forging a path where innovation meets ethics, reliability, and accountability. We primarily focus on improving the NN's interpretability and generalization.

We aim to bridge the gaps between spatial and temporal data and design novel spatial-temporal models for real-world applications such as human trajectories and smart-cities. 

We focus on analyzing the capabilities of large foundation models in various tasks, such as social network analysis and spatial-temporal data modeling.

We study large-scale social networks, such as information diffusion, user interaction, and rumor propagation within the network. 

Seeing the Unseen in Micro-Video Popularity Prediction: Self-Correlation Retrieval for Missing Modality Generation

Biting Off More Than You Can Detect: Retrieval-Augmented Multimodal Experts for Short Video Hate Detection

Improving Multimodal Social Media Popularity Prediction via Selective Retrieval Knowledge Augmentation

DOSE: Diffusion Dropout with Adaptive Prior for Speech Enhancement