Reproduce “Learning Restricted Boltzmann Machines with Sparse Latent Variables” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Guy Bresler, Rares-Darius Buhai
The authors learn dependencies among observed variables in Boltzmann machines from their marginal distribution, rather than recovering hidden factors. Their algorithm uses latent sparsity: when each observed variable’s marginal Markov random field neighborhood connects to at most s latent variables, searching groups of at most 2^s yields runtime approximately n^(2^s+1). This improves the n^d search when s < log2(d−1), but exact recovery depends on minimum potential α.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.