Spectral Dynamics of Learning Restricted Boltzmann Machines
This provides theoretical insights into RBM training dynamics, which is incremental for researchers in unsupervised learning.
The authors tackled the problem of understanding the learning dynamics of Restricted Boltzmann Machines (RBMs) by analyzing their spectral properties, showing that data statistics drive the selection of unstable modes in the linear regime and deriving equations for mode interactions in the non-linear regime to define a deterministic learning curve.
The Restricted Boltzmann Machine (RBM), an important tool used in machine learning in particular for unsupervized learning tasks, is investigated from the perspective of its spectral properties. Starting from empirical observations, we propose a generic statistical ensemble for the weight matrix of the RBM and characterize its mean evolution. This let us show how in the linear regime, in which the RBM is found to operate at the beginning of the training, the statistical properties of the data drive the selection of the unstable modes of the weight matrix. A set of equations characterizing the non-linear regime is then derived, unveiling in some way how the selected modes interact in later stages of the learning procedure and defining a deterministic learning curve for the RBM.