Meta-Learning with Variational Bayes
This addresses the challenge of meta-learning in domains with unlabeled data, which is important for applications requiring human-level cognitive flexibility, though it appears incremental as it builds on existing variational Bayes frameworks.
The paper tackles the problem of generative meta-learning for unlabeled data by introducing a variational Bayes approach that creates fast-adapting latent-space generative models, achieving theoretical and empirical improvements without dependency on generative neural networks.
The field of meta-learning seeks to improve the ability of today's machine learning systems to adapt efficiently to small amounts of data. Typically this is accomplished by training a system with a parametrized update rule to improve a task-relevant objective based on supervision or a reward function. However, in many domains of practical interest, task data is unlabeled, or reward functions are unavailable. In this paper we introduce a new approach to address the more general problem of generative meta-learning, which we argue is an important prerequisite for obtaining human-level cognitive flexibility in artificial agents, and can benefit many practical applications along the way. Our contribution leverages the AEVB framework and mean-field variational Bayes, and creates fast-adapting latent-space generative models. At the heart of our contribution is a new result, showing that for a broad class of deep generative latent variable models, the relevant VB updates do not depend on any generative neural network. The theoretical merits of our approach are reflected in empirical experiments.