Reviews: Continual Learning with Deep Generative Replay

Quality: The paper is technically sound. Extensive comparisons between the proposed approach and other baseline approaches are being made. Clarity: The paper is well-organized. However, insufficient details are provided on the architectures of the discriminator, generator and the used optimizer. I would be surprised if the cited papers refer to the hippocampus as a short-term memory system since the latter typically refers to working memroy.