Lifelong adaptation, continual learning, incremental learning, and unlearning for evolving foundation models.
To enable foundation models to learn continuously from new data while retaining prior knowledge and forgetting sensitive information when required.
This direction studies how large models adapt over time without catastrophic forgetting. Research covers rehearsal-free continual learning, document retrieval under distribution shift, and principled machine unlearning for safe AI.
Paper An Optimal Transport-driven Approach for Cultivating Latent Space in Online Incremental Learning accepted at CVPR.
Paper A Good Teacher Adapts Their Knowledge for Distillation accepted at ICCV.
Paper Boosting Multiple Views for pretrained-based Continual Learning accepted at ICLR.
Quyen Tran, Ngoc-Hai Nguyen, Quan Dao, Hoang Phan, Ngo Van-Linh, Khoat Than, Dinh Phung, Dimitris N. Metaxas, Trung Le
Quyen Tran, Ngoc-Hai Nguyen, Quan Dao, Zhuowei Li, Nam Le Hai, Trung Le, Dimitris N. Metaxas
Chengyao Qian, Trung Le, Mehrtash Harandi
Quyen Tran, Tung Lam Tran, Khanh Doan, Toan Tran, Khoat Than, Dinh Phung, Trung Le
Quyen Tran, Hoang Phan, Minh Le, Tuan Truong, Dinh Phung, Linh Ngo, Thien Nguyen, Nhat Ho, Trung Le
Quyen Tran, Tung Lam Tran, Khoat Than, Toan Tran, Dinh Phung, Trung Le
Khanh Doan, Quyen Tran, Tuan Nguyen, Dinh Phung, Trung Le