Chenqi Presents AnchorInv at AAAI-25

From 25 February to 4 March 2025, the Thirty-Ninth AAAI Conference on Artificial Intelligence (AAAI-25) was held at the Pennsylvania Convention Centre in Philadelphia, bringing together the international AI research community for a week of scientific exchange. Our group was delighted to see Chenqi attend this event and present his latest work, AnchorInv, in the main conference programme.

Chenqi’s paper addresses the challenge of learning from limited and incrementally available biomedical data—a common hurdle in real-world applications. AnchorInv introduces an efficient buffer-replay strategy that generates synthetic samples guided by anchor points in feature space, rather than storing raw data. This method offers both privacy protection and improved model adaptability, enabling deep learning models to retain previous knowledge while adapting to novel classes. Experiments on three public physiological time series datasets demonstrated that AnchorInv effectively prevents knowledge forgetting and achieves superior performance compared to state-of-the-art baselines. 

For more details, please refer to our group’s research direction: Generative AI for Personlisation