D-CBRS: Accounting For Intra-Class Diversity in Continual Learning
This work addresses forgetting in continual learning for scenarios with imbalanced and diverse data, representing an incremental improvement over existing replay-based approaches.
The paper tackles the problem of forgetting in continual learning with class-imbalanced data by proposing D-CBRS, an algorithm that accounts for intra-class diversity when storing instances in memory, and shows it outperforms state-of-the-art methods on datasets with high intra-class diversity.
Continual learning -- accumulating knowledge from a sequence of learning experiences -- is an important yet challenging problem. In this paradigm, the model's performance for previously encountered instances may substantially drop as additional data are seen. When dealing with class-imbalanced data, forgetting is further exacerbated. Prior work has proposed replay-based approaches which aim at reducing forgetting by intelligently storing instances for future replay. Although Class-Balancing Reservoir Sampling (CBRS) has been successful in dealing with imbalanced data, the intra-class diversity has not been accounted for, implicitly assuming that each instance of a class is equally informative. We present Diverse-CBRS (D-CBRS), an algorithm that allows us to consider within class diversity when storing instances in the memory. Our results show that D-CBRS outperforms state-of-the-art memory management continual learning algorithms on data sets with considerable intra-class diversity.