Diving into Class-Incremental Learning from Better Balancing Old and New knowledge
Keywords: class incremental learning, catastrophic forgetting, complementary learning system, knowledge representation
TL;DR: We propose a novel CLS-based method to acquire more new knowledge from new tasks while consolidating the old knowledge so as to make a better balance between them.
Abstract: Class-Incremental Learning (Class-IL) aims to continuously learn new knowledge without forgetting old knowledge from a given data stream using deep neural networks. Recent Class-IL methods strive to balance old and new knowledge and have achieved excellent results in mitigating the forgetting by mainly employing the rehearsal-based strategy. However, the representation learning on new tasks is often impaired since the trade-off is hard to taken between old and new knowledge. To overcome this challenge, based on the Complementary Learning System (CLS) theory, we propose a novel CLS-based method by focusing on the representation of old and new knowledge in Class-IL, which can acquire more new knowledge from new tasks while consolidating the old knowledge so as to make a better balance between them. Specifically, our proposed method has two novel components: (1) To effectively mitigate the forgetting, we first propose a bidirectional transport (BDT) strategy between old and new models, which can better integrate the old knowledge into the new knowledge and meanwhile enforce the old knowledge to be better consolidated by bidirectionally transferring parameters across old and new models. (2) To ensure that the representation of new knowledge is not impaired by the old knowledge, we further devise a selective momentum (SMT) mechanism to give parameters greater flexibility to learn new knowledge while transferring important old knowledge, which is achieved by selectively (momentum) updating network parameters through parameter importance evaluation. Extensive experiments on four benchmarks show that our proposed method significantly outperforms the state-of-the-arts under the Class-IL setting.
Primary Area: transfer learning, meta learning, and lifelong learning
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 3554
Loading