Looking through the past: better knowledge retention for generative replay in continual learning
This work addresses the challenge of knowledge retention in continual learning for AI systems, but it is incremental as it builds upon existing VAE-based generative replay methods.
The authors tackled the problem of generative replay in continual learning, which struggles with complex datasets, by proposing three modifications to reduce feature drift and improve alignment, resulting in their method outperforming other generative replay approaches in various scenarios.
In this work, we improve the generative replay in a continual learning setting to perform well on challenging scenarios. Current generative rehearsal methods are usually benchmarked on small and simple datasets as they are not powerful enough to generate more complex data with a greater number of classes. We notice that in VAE-based generative replay, this could be attributed to the fact that the generated features are far from the original ones when mapped to the latent space. Therefore, we propose three modifications that allow the model to learn and generate complex data. More specifically, we incorporate the distillation in latent space between the current and previous models to reduce feature drift. Additionally, a latent matching for the reconstruction and original data is proposed to improve generated features alignment. Further, based on the observation that the reconstructions are better for preserving knowledge, we add the cycling of generations through the previously trained model to make them closer to the original data. Our method outperforms other generative replay methods in various scenarios. Code available at https://github.com/valeriya-khan/looking-through-the-past.