MLLGSep 6, 2016

Direct Feedback Alignment Provides Learning in Deep Neural Networks

arXiv:1609.01596v5555 citations
AI Analysis

This provides a biologically plausible alternative to back-propagation for machine learning, though it is incremental as it builds on feedback alignment principles.

The paper tackles the problem of training deep neural networks without back-propagation by using direct feedback alignment with fixed random feedback connections, achieving zero training error in convolutional and deep networks and competitive test performance on MNIST and CIFAR, with 1.45% error on permutation-invariant MNIST when combined with dropout.

Artificial neural networks are most commonly trained with the back-propagation algorithm, where the gradient for learning is provided by back-propagating the error, layer by layer, from the output layer to the hidden layers. A recently discovered method called feedback-alignment shows that the weights used for propagating the error backward don't have to be symmetric with the weights used for propagation the activation forward. In fact, random feedback weights work evenly well, because the network learns how to make the feedback useful. In this work, the feedback alignment principle is used for training hidden layers more independently from the rest of the network, and from a zero initial condition. The error is propagated through fixed random feedback connections directly from the output layer to each hidden layer. This simple method is able to achieve zero training error even in convolutional networks and very deep networks, completely without error back-propagation. The method is a step towards biologically plausible machine learning because the error signal is almost local, and no symmetric or reciprocal weights are required. Experiments show that the test performance on MNIST and CIFAR is almost as good as those obtained with back-propagation for fully connected networks. If combined with dropout, the method achieves 1.45% error on the permutation invariant MNIST task.

Code Implementations5 repos
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes