I recently published a paper on just this topic where we juxtapose replay in the mammalian brain and deep neural networks: https://arxiv.org/abs/2104.04132
There are many salient differences, and the paper is a collaboration between deep learning researchers and neuroscientists. I have developed some state-of-the-art deep learning algorithms that use replay for continual learning, but we believe there is a lot to be further gained by studying the brain.
There are many salient differences, and the paper is a collaboration between deep learning researchers and neuroscientists. I have developed some state-of-the-art deep learning algorithms that use replay for continual learning, but we believe there is a lot to be further gained by studying the brain.