Lightweight Probabilistic Deep Networks

TL;DR

This paper introduces two lightweight methods for incorporating uncertainty estimation into deep neural networks, enhancing their reliability and robustness with minimal modifications.

Contribution

The paper proposes probabilistic output layers and assumed density filtering to efficiently propagate uncertainties in deep networks, improving practical applicability.

Findings

Uncertainties correlate well with empirical errors.

Robustness to adversarial examples is significantly increased.

Methods require only minor modifications to existing architectures.

Abstract

Even though probabilistic treatments of neural networks have a long history, they have not found widespread use in practice. Sampling approaches are often too slow already for simple networks. The size of the inputs and the depth of typical CNN architectures in computer vision only compound this problem. Uncertainty in neural networks has thus been largely ignored in practice, despite the fact that it may provide important information about the reliability of predictions and the inner workings of the network. In this paper, we introduce two lightweight approaches to making supervised learning with probabilistic deep networks practical: First, we suggest probabilistic output layers for classification and regression that require only minimal changes to existing networks. Second, we employ assumed density filtering and show that activation uncertainties can be propagated in a practical fashion through the entire network, again with minor changes. Both probabilistic networks retain the predictive power of the deterministic counterpart, but yield uncertainties that correlate well with the empirical error induced by their predictions. Moreover, the robustness to adversarial examples is significantly increased.