Learning to Adapt by Minimizing Discrepancy

Abstract

We explore whether useful temporal neural generative models can be learned from sequential data without back-propagation through time. We investigate the viability of a more neurocognitively-grounded approach in the context of unsupervised generative modeling of sequences. Specifically, we build on the concept of predictive coding, which has gained influence in cognitive science, in a neural framework. To do so we develop a novel architecture, the Temporal Neural Coding Network, and its learning algorithm, Discrepancy Reduction. The underlying directed generative model is fully recurrent, meaning that it employs structural feedback connections and temporal feedback connections, yielding information propagation cycles that create local learning signals. This facilitates a unified bottom-up and top-down approach for information transfer inside the architecture. Our proposed algorithm shows promise on the bouncing balls generative modeling problem. Further experiments could be conducted to explore the strengths and weaknesses of our approach.

Note: Additional experiments in support of this paper are still running (updates will be made as they are completed)

References in corpus (19)

• Adam: A Method for Stochastic Optimization
• Deeply-Supervised Nets
• Modeling Temporal Dependencies in High-Dimensional Sequences: Application to Polyphonic Music Generation and Transcription
• Deep Predictive Coding Networks for Video Prediction and Unsupervised Learning
• Towards Biologically Plausible Deep Learning
• Fast Inference in Sparse Coding Algorithms with Applications to Object Recognition
• Direct Feedback Alignment Provides Learning in Deep Neural Networks
• Deep Learning of Representations: Looking Forward
• Random feedback weights support learning in deep neural networks
• Decoupled Neural Interfaces using Synthetic Gradients
• How Important is Weight Symmetry in Backpropagation?
• Spike-and-Slab Sparse Coding for Unsupervised Feature Discovery
• Deep Predictive Coding Networks
• Online Semi-Supervised Learning with Deep Hybrid Boltzmann Machines and Denoising Autoencoders
• Reweighted Wake-Sleep
• Training recurrent networks online without backtracking
• Learning in the Machine: Random Backpropagation and the Deep Learning Channel
• Training Language Models Using Target-Propagation
• A Theory of Feature Learning

Cited by in corpus (7)

• Conducting Credit Assignment by Aligning Local Representations
• A Unified Framework of Online Learning Algorithms for Training Recurrent Neural Networks
• Predictive Coding Can Do Exact Backpropagation on Convolutional and Recurrent Neural Networks
• Activation Relaxation: A Local Dynamical Approximation to Backpropagation in the Brain
• Spiking Neural Predictive Coding for Continual Learning from Data Streams
• Lifelong Neural Predictive Coding: Learning Cumulatively Online without Forgetting
• Large-Scale Gradient-Free Deep Learning with Recursive Local Representation Alignment