Random feedback weights support learning in deep neural networks

Abstract

The brain processes information through many layers of neurons. This deep architecture is representationally powerful, but it complicates learning by making it hard to identify the responsible neurons when a mistake is made. In machine learning, the backpropagation algorithm assigns blame to a neuron by computing exactly how it contributed to an error. To do this, it multiplies error signals by matrices consisting of all the synaptic weights on the neuron's axon and farther downstream. This operation requires a precisely choreographed transport of synaptic weight information, which is thought to be impossible in the brain. Here we present a surprisingly simple algorithm for deep learning, which assigns blame by multiplying error signals by random synaptic weights. We show that a network can learn to extract useful information from signals sent through these random feedback connections. In essence, the network learns to learn. We demonstrate that this new mechanism performs as quickly and accurately as backpropagation on a variety of problems and describe the principles which underlie its function. Our demonstration provides a plausible basis for how a neuron can be adapted using error signals generated at distal locations in the brain, and thus dispels long-held assumptions about the algorithmic constraints on learning in neural circuits.

14 pages, 5 figures in main text; 13 pages appendix

References in corpus (1)

• Deep Big Simple Neural Nets Excel on Handwritten Digit Recognition

Cited by in corpus (33)

• Synaptic Plasticity Dynamics for Deep Continuous Local Learning (DECOLLE)
• Towards Biologically Plausible Deep Learning
• Uncertainty-Aware Reinforcement Learning for Collision Avoidance
• Direct Feedback Alignment Provides Learning in Deep Neural Networks
• Unsupervised Learning by Competing Hidden Units
• A Theory of Local Learning, the Learning Channel, and the Optimality of Backpropagation
• Towards an integration of deep learning and neuroscience
• Trainable and Dynamic Computing: Error Backpropagation through Physical Media
• How Important is Weight Symmetry in Backpropagation?
• Learning to solve the credit assignment problem
• Revisiting Locally Supervised Learning: an Alternative to End-to-end Training
• Meta-Learning through Hebbian Plasticity in Random Networks
• Equilibrium Propagation: Bridging the Gap Between Energy-Based Models and Backpropagation
• Relaxing the Constraints on Predictive Coding Models
• Efficient Convolutional Neural Network Training with Direct Feedback Alignment
• Variational Walkback: Learning a Transition Operator as a Stochastic Recurrent Net
• Activation Relaxation: A Local Dynamical Approximation to Backpropagation in the Brain
• Learning to Adapt by Minimizing Discrepancy
• Improving Gradient Estimation in Evolutionary Strategies With Past Descent Directions
• Learning in the Machine: Random Backpropagation and the Deep Learning Channel
• ToyArchitecture: Unsupervised Learning of Interpretable Models of the World
• DANTE: Deep AlterNations for Training nEural networks
• Adaptive Bidirectional Backpropagation: Towards Biologically Plausible Error Signal Transmission in Neural Networks
• Continual Learning with Deep Artificial Neurons
• MPLP: Learning a Message Passing Learning Protocol
• Reinforcement learning with a network of spiking agents
• Benchmarking the Accuracy and Robustness of Feedback Alignment Algorithms
• Semantics, Representations and Grammars for Deep Learning
• MAP Propagation Algorithm: Faster Learning with a Team of Reinforcement Learning Agents
• Training DNNs in O(1) memory with MEM-DFA using Random Matrices
• CATERPILLAR: Coarse Grain Reconfigurable Architecture for Accelerating the Training of Deep Neural Networks
• A More Biologically Plausible Local Learning Rule for ANNs
• Efficient Training Convolutional Neural Networks on Edge Devices with Gradient-pruned Sign-symmetric Feedback Alignment