Bayesian Optimization Meets Laplace Approximation for Robotic Introspection

TL;DR

This paper introduces a scalable Bayesian optimization method combined with Laplace approximation to improve the confidence estimates of deep neural networks in robotics, enhancing their reliability for autonomous systems.

Contribution

It presents a novel Bayesian optimization algorithm that better calibrates DNNs' uncertainty estimates using Laplace approximation, scalable to large datasets and architectures.

Findings

BO requires fewer iterations than random search

Framework scales to large datasets and architectures

Improves confidence calibration of DNNs in robotics

Abstract

In robotics, deep learning (DL) methods are used more and more widely, but their general inability to provide reliable confidence estimates will ultimately lead to fragile and unreliable systems. This impedes the potential deployments of DL methods for long-term autonomy. Therefore, in this paper we introduce a scalable Laplace Approximation (LA) technique to make Deep Neural Networks (DNNs) more introspective, i.e. to enable them to provide accurate assessments of their failure probability for unseen test data. In particular, we propose a novel Bayesian Optimization (BO) algorithm to mitigate their tendency of under-fitting the true weight posterior, so that both the calibration and the accuracy of the predictions can be simultaneously optimized. We demonstrate empirically that the proposed BO approach requires fewer iterations for this when compared to random search, and we show that the proposed framework can be scaled up to large datasets and architectures.