We Know Where We Don't Know: 3D Bayesian CNNs for Credible Geometric Uncertainty
- Sandia National Lab. (SNL-NM), Albuquerque, NM (United States); Univ. of Southern California, Los Angeles, CA (United States)
- Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
Deep learning has been successfully applied to the segmentation of 3D Computed Tomography (CT) scans. Establishing the credibility of these segmentations requires uncertainty quantification (UQ) to identify untrustworthy predictions. Recent UQ architectures include Monte Carlo dropout networks (MCDNs), which approximate deep Gaussian processes, and Bayesian neural networks (BNNs), which learn the distribution of the weight space. BNNs are advantageous over MCDNs for UQ but are thought to be computationally infeasible in high dimension, and neither architecture has produced interpretable geometric uncertainty maps. We propose a novel 3D Bayesian convolutional neural network (BCNN), the first deep learning method which generates statistically credible geometric uncertainty maps and scales for application to 3D data. We present experimental results on CT scans of graphite electrodes and laser-welded metals and show that our BCNN outperforms an MCDN in recent uncertainty metrics. The geometric uncertainty maps generated by our BCNN capture distributions of sigmoid values that are interpretable as confidence intervals, critical for applications that rely on deep learning for high-consequence decisions.
- Research Organization:
- Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
- Sponsoring Organization:
- USDOE National Nuclear Security Administration (NNSA); USDOE Laboratory Directed Research and Development (LDRD) Program
- DOE Contract Number:
- AC04-94AL85000; NA0003525
- OSTI ID:
- 1605518
- Report Number(s):
- SAND-2020-3269R; 684831
- Country of Publication:
- United States
- Language:
- English
Similar Records
Deeply uncertain: comparing methods of uncertainty quantification in deep learning algorithms
Semi-supervised Bayesian Low-shot Learning