DeepMind, London, UK; UCL, UK.
DeepMind, London, UK.
Curr Opin Neurobiol. 2019 Apr;55:82-89. doi: 10.1016/j.conb.2019.01.011. Epub 2019 Mar 7.
It has long been speculated that the backpropagation-of-error algorithm (backprop) may be a model of how the brain learns. Backpropagation-through-time (BPTT) is the canonical temporal-analogue to backprop used to assign credit in recurrent neural networks in machine learning, but there's even less conviction about whether BPTT has anything to do with the brain. Even in machine learning the use of BPTT in classic neural network architectures has proven insufficient for some challenging temporal credit assignment (TCA) problems that we know the brain is capable of solving. Nonetheless, recent work in machine learning has made progress in solving difficult TCA problems by employing novel memory-based and attention-based architectures and algorithms, some of which are brain inspired. Importantly, these recent machine learning methods have been developed in the context of, and with reference to BPTT, and thus serve to strengthen BPTT's position as a useful normative guide for thinking about temporal credit assignment in artificial and biological systems alike.
长期以来，人们一直推测误差反向传播算法（backprop）可能是大脑学习的一种模型。时间反向传播（BPTT）是反向传播的典型时间模拟，用于在机器学习中的递归神经网络中分配信用，但对于 BPTT 是否与大脑有关，人们的信心甚至更少。即使在机器学习中，在经典神经网络架构中使用 BPTT 已被证明不足以解决一些具有挑战性的时间信用分配（TCA）问题，而我们知道大脑能够解决这些问题。尽管如此，机器学习领域的最新工作通过采用基于记忆和基于注意力的新型架构和算法在解决困难的 TCA 问题方面取得了进展，其中一些是受大脑启发的。重要的是，这些最近的机器学习方法是在 BPTT 的背景下和参考 BPTT 开发的，因此有助于加强 BPTT 作为思考人工和生物系统中时间信用分配的有用规范指南的地位。
