Stopping Criteria in Contrastive Divergence: Alternatives to the Reconstruction Error
This addresses a methodological issue for researchers using RBMs in unsupervised learning, but it is incremental as it builds on prior critiques of existing stopping criteria.
The paper tackles the problem of determining when to stop training Restricted Boltzmann Machines using the Contrastive Divergence algorithm, proposing simple alternatives to the reconstruction error to detect decreases in log-likelihood earlier.
Restricted Boltzmann Machines (RBMs) are general unsupervised learning devices to ascertain generative models of data distributions. RBMs are often trained using the Contrastive Divergence learning algorithm (CD), an approximation to the gradient of the data log-likelihood. A simple reconstruction error is often used to decide whether the approximation provided by the CD algorithm is good enough, though several authors (Schulz et al., 2010; Fischer & Igel, 2010) have raised doubts concerning the feasibility of this procedure. However, not many alternatives to the reconstruction error have been used in the literature. In this manuscript we investigate simple alternatives to the reconstruction error in order to detect as soon as possible the decrease in the log-likelihood during learning.