Options
Ramifications of Approximate Posterior Inference for Bayesian Deep Learning in Adversarial and Out-of-Distribution Settings
Date Issued
2020-08-23
Date Available
2021-10-26T11:22:09Z
Abstract
Deep neural networks have been successful in diverse discriminative classification tasks, although, they are poorly calibrated often assigning high probability to misclassified predictions. Potential consequences could lead to trustworthiness and accountability of the models when deployed in real applications, where predictions are evaluated based on their confidence scores. Existing solutions suggest the benefits attained by combining deep neural networks and Bayesian inference to quantify uncertainty over the models’ predictions for ambiguous data points. In this work we propose to validate and test the efficacy of likelihood based models in the task of out of distribution detection (OoD). Across different datasets and metrics we show that Bayesian deep learning models indeed outperform conventional neural networks but in the event of minimal overlap between in/out distribution classes, even the best models exhibit a reduction in AUC scores in detecting OoD data. We hypothesise that the sensitivity of neural networks to unseen inputs could be a multi-factor phenomenon arising from the different architectural design choices often amplified by the curse of dimensionality. Preliminary investigations indicate the potential inherent role of bias due to choices of initialisation, architecture or activation functions. Furthermore, we perform an analysis on the effect of adversarial noise resistance methods regarding in and out-of-distribution performance when combined with Bayesian deep learners.
Sponsorship
Science Foundation Ireland
Type of Material
Book Chapter
Publisher
Springer
Series
Lecture Notes in Computer Science
12535
Copyright (Published Version)
2020 Springer
Web versions
Language
English
Status of Item
Peer reviewed
Part of
Bartoli, A. and Fusiello, A. (eds.).Computer Vision – ECCV 2020 Workshops. ECCV 2020: Glasgow, UK, August 23–28, 2020, Proceedings, Part I
Conference Details
The 16th European Conference on Computer Vision (ECCV 2020), Online Conference, 23-28 August 2020
ISBN
978-3-030-66414-5
This item is made available under a Creative Commons License
File(s)
Owning collection
Scopus© citations
0
Acquisition Date
Mar 28, 2024
Mar 28, 2024
Views
388
Last Week
3
3
Last Month
3
3
Acquisition Date
Mar 28, 2024
Mar 28, 2024
Downloads
544
Last Week
2
2
Last Month
10
10
Acquisition Date
Mar 28, 2024
Mar 28, 2024