Three Learning Stages and Accuracy-Efficiency Tradeoff of Restricted Boltzmann Machines
This work addresses the accuracy-efficiency tradeoff in RBM training, which is crucial for practitioners in unsupervised machine learning, though it is incremental as it builds on existing RBM theory.
The paper identifies three learning regimes in Restricted Boltzmann Machines (RBMs) where accuracy and efficiency trade off, showing that independent learning improves accuracy without losing efficiency, correlation learning trades higher accuracy for lower efficiency, and degradation leads to no improvement or deterioration in both.
Restricted Boltzmann Machines (RBMs) offer a versatile architecture for unsupervised machine learning that can in principle approximate any target probability distribution with arbitrary accuracy. However, the RBM model is usually not directly accessible due to its computational complexity, and Markov-chain sampling is invoked to analyze the learned probability distribution. For training and eventual applications, it is thus desirable to have a sampler that is both accurate and efficient. We highlight that these two goals generally compete with each other and cannot be achieved simultaneously. More specifically, we identify and quantitatively characterize three regimes of RBM learning: independent learning, where the accuracy improves without losing efficiency; correlation learning, where higher accuracy entails lower efficiency; and degradation, where both accuracy and efficiency no longer improve or even deteriorate. These findings are based on numerical experiments and heuristic arguments.