Two Routes to Scalable Credit Assignment without Weight Symmetry

Abstract

The neural plausibility of backpropagation has long been disputed, primarily for its use of non-local weight transport $-$ the biologically dubious requirement that one neuron instantaneously measure the synaptic weights of another. Until recently, attempts to create local learning rules that avoid weight transport have typically failed in the large-scale learning scenarios where backpropagation shines, e.g. ImageNet categorization with deep convolutional networks. Here, we investigate a recently proposed local learning rule that yields competitive performance with backpropagation and find that it is highly sensitive to metaparameter choices, requiring laborious tuning that does not transfer across network architecture. Our analysis indicates the underlying mathematical reason for this instability, allowing us to identify a more robust local learning rule that better transfers without metaparameter tuning. Nonetheless, we find a performance and stability gap between this local rule and backpropagation that widens with increasing model depth. We then investigate several non-local learning rules that relax the need for instantaneous weight transport into a more biologically-plausible "weight estimation" process, showing that these rules match state-of-the-art performance on deep networks and operate effectively in the presence of noisy updates. Taken together, our results suggest two routes towards the discovery of neural implementations for credit assignment without weight symmetry: further improvement of local rules so that they perform consistently across architectures and the identification of biological implementations for non-local learning mechanisms.

ICML 2020 Camera Ready Version, 19 pages including supplementary information, 10 figures

References in corpus (10)

• Adam: A Method for Stochastic Optimization
• Exact solutions to the nonlinear dynamics of learning in deep linear neural networks
• Direct Feedback Alignment Provides Learning in Deep Neural Networks
• How Auto-Encoders Could Provide Credit Assignment in Deep Networks via Target Propagation
• Deep Learning without Weight Transport
• How Important is Weight Symmetry in Backpropagation?
• Feedback alignment in deep convolutional networks
• Two Routes to Scalable Credit Assignment without Weight Symmetry
• Loss Landscapes of Regularized Linear Autoencoders
• TF-Replicator: Distributed Machine Learning for Researchers

Cited by in corpus (9)

• A Theoretical Framework for Target Propagation
• Two Routes to Scalable Credit Assignment without Weight Symmetry
• On the relationship between predictive coding and backpropagation
• Local plasticity rules can learn deep representations using self-supervised contrastive predictions
• Learning to Learn with Feedback and Local Plasticity
• Identifying Learning Rules From Neural Network Observables
• Direct Feedback Alignment Scales to Modern Deep Learning Tasks and Architectures
• Tightening the Biological Constraints on Gradient-Based Predictive Coding
• How and When Random Feedback Works: A Case Study of Low-Rank Matrix Factorization