Is Class-Incremental Enough for Continual Learning?
This work critiques current continual learning benchmarks for being unrealistic and incremental, potentially limiting progress in the field.
The paper challenges the predominant focus on class-incremental scenarios in continual learning research, arguing that this setting artificially worsens catastrophic forgetting and overlooks other objectives like forward transfer and efficiency. It advocates for exploring alternative scenarios with repetition to better assess models in real-world environments.
The ability of a model to learn continually can be empirically assessed in different continual learning scenarios. Each scenario defines the constraints and the opportunities of the learning environment. Here, we challenge the current trend in the continual learning literature to experiment mainly on class-incremental scenarios, where classes present in one experience are never revisited. We posit that an excessive focus on this setting may be limiting for future research on continual learning, since class-incremental scenarios artificially exacerbate catastrophic forgetting, at the expense of other important objectives like forward transfer and computational efficiency. In many real-world environments, in fact, repetition of previously encountered concepts occurs naturally and contributes to softening the disruption of previous knowledge. We advocate for a more in-depth study of alternative continual learning scenarios, in which repetition is integrated by design in the stream of incoming information. Starting from already existing proposals, we describe the advantages such class-incremental with repetition scenarios could offer for a more comprehensive assessment of continual learning models.