Learning Invariant Features for Online Continual Learning

Yiduo Guo, Bing Liu, Dongyan Zhao

Keywords: continual learning, online continual learning

Abstract: It has been shown recently that learning only discriminative features that are sufficient to separate the classes in a task using a traditional learning method has a major shortcoming for continual learning (CL). This is because many features that are not learned may be necessary for distinguishing classes of some future tasks. When such a future task arrives, these features have to be learned by updating the network, which causes catastrophic forgetting (CF). A recent work on online CL showed that if the learning method can learn as many features as possible from each class, called holistic representations, CF can be significantly reduced to achieve a large performance gain. This paper argues that learning only holistic representations is still insufficient. The learned representations should also be invariant and those features that are present in the data but are irrelevant to the class (e.g., the background information) should be ignored for better generalization across tasks. This new condition further boosts the performance significantly. This paper proposes several strategies and a loss to learn holistic and invariant representations and evaluates their effectiveness in online CL.

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.

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: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning