Learning Expressive Priors for Generalization and Uncertainty Estimation in Neural Networks
This work addresses generalization and uncertainty issues in neural networks, which is an incremental advance in Bayesian deep learning.
The paper tackles the problem of improving generalization and uncertainty estimation in deep neural networks by learning expressive priors from scalable structured posteriors, achieving non-vacuous generalization bounds and demonstrating effectiveness in empirical tests.
In this work, we propose a novel prior learning method for advancing generalization and uncertainty estimation in deep neural networks. The key idea is to exploit scalable and structured posteriors of neural networks as informative priors with generalization guarantees. Our learned priors provide expressive probabilistic representations at large scale, like Bayesian counterparts of pre-trained models on ImageNet, and further produce non-vacuous generalization bounds. We also extend this idea to a continual learning framework, where the favorable properties of our priors are desirable. Major enablers are our technical contributions: (1) the sums-of-Kronecker-product computations, and (2) the derivations and optimizations of tractable objectives that lead to improved generalization bounds. Empirically, we exhaustively show the effectiveness of this method for uncertainty estimation and generalization.