Hidden Markov Neural Networks
This addresses a crucial problem in time-series forecasting and continual learning for practitioners needing robust models, though it appears incremental as it builds on existing Bayesian neural network and regularization techniques.
The paper tackles the challenge of balancing adaptation to new data with forgetting outdated information in time-series forecasting and continual learning by introducing Hidden Markov Neural Networks, which model neural network weights as hidden states in a Hidden Markov model. Experiments on MNIST, dynamic classification, and video forecasting show strong predictive performance and effective uncertainty quantification.
We define an evolving in-time Bayesian neural network called a Hidden Markov Neural Network, which addresses the crucial challenge in time-series forecasting and continual learning: striking a balance between adapting to new data and appropriately forgetting outdated information. This is achieved by modelling the weights of a neural network as the hidden states of a Hidden Markov model, with the observed process defined by the available data. A filtering algorithm is employed to learn a variational approximation of the evolving-in-time posterior distribution over the weights. By leveraging a sequential variant of Bayes by Backprop, enriched with a stronger regularization technique called variational DropConnect, Hidden Markov Neural Networks achieve robust regularization and scalable inference. Experiments on MNIST, dynamic classification tasks, and next-frame forecasting in videos demonstrate that Hidden Markov Neural Networks provide strong predictive performance while enabling effective uncertainty quantification.