Author
Şensoy, Murat, Saleki, Maryam, Julier, S., Aydoğan, Reyhan, Reid, J.
Publication Date
2021
Publication Place
-
IEEE
Type
Document
Language
English
Digital
Yes
Manuscript
No
Library
Özyeğin University
Library Asset ID
978-073814266-1
Record ID
a10c90b0-a05c-4cb1-925d-d4f245c96b8d
Library Location
Computer Science
Date
2021
Notes
United States Department of Defense US Army Research Laboratory (ARL)
Sample Text
In this paper, we propose risk-calibrated evidential deep classifiers to reduce the costs associated with classification errors. We use two main approaches. The first is to develop methods to quantify the uncertainty of a classifier’s predictions and reduce the likelihood of acting on erroneous predictions. The second is a novel way to train the classifier such that erroneous classifications are biased towards less risky categories. We combine these two approaches in a principled way. While doing this, we extend evidential deep learning with pignistic probabilities, which are used to quantify uncertainty of classification predictions and model rational decision making under uncertainty.We evaluate the performance of our approach on several image classification tasks. We demonstrate that our approach allows to (i) incorporate misclassification cost while training deep classifiers, (ii) accurately quantify the uncertainty of classification predictions, and (iii) simultaneously learn how to make classification decisions to minimize expected cost of classification errors.
DOI
10.1109/WACV48630.2021.00253
