Reproduce “Learning PAC-Bayes Priors for Probabilistic Neural Networks” 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
Maria Perez-Ortiz, Omar Rivasplata, Benjamin Guedj, Matthew Gleeson, Jingyu Zhang, John Shawe-Taylor, Miroslaw Bober, Josef Kittler
The authors study how data-dependent priors and data allocation affect PAC-Bayes certificates for randomized neural classifiers. They learn a prior on one subset, certify on separate examples, and train the posterior on all training data. Across six datasets, data-free priors yield much looser certificates, while the best allocation varies. On Spambase, 90% prior allocation slightly improves test error but worsens the certificate versus 50%.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.