PAC-Bayesian Meta-Learning: From Theory to Practice
PAC-Bayesian Meta-Learning: From Theory to Practice
Jonas Rothfuss,Martin Josifoski,Vincent Fortuin,Andreas Krause
2022 · DOI: 10.48550/arXiv.2211.07206
arXiv.org · 12 Citations
TLDR
A theoretical analysis using the PAC-Bayesian framework is provided and the first bound for meta-learners with unbounded loss functions is derived, thereby avoiding the reliance on nested optimization and giving rise to an optimization problem amenable to standard variational methods that scale well.