2022-12-31-trombley22a.md

abstract	booktitle	title	volume	year	layout	series	publisher	issn	id	month	tex_title	firstpage	lastpage	page	order	cycles	bibtex_author	author	date	address	container-title	genre	issued	pdf	extras
Deep learning has demonstrated impressive accuracy for prostate lesion identification and classification. Deep learning algorithms are considered black-box methods therefore they require explanation methods to gain insight into the model’s classification. For high stakes tasks such as medical diagnosis, it is important that explanation methods are able to estimate explanation uncertainty. Recently, there have been various methods proposed for providing uncertainty-based explanations. However, the clinical effectiveness of uncertainty-based explanation methods and what radiologists deem explainable within this context is still largely unknown. To that end, this pilot study investigates the effectiveness of uncertainty-based prostate lesion detection explanations. It also attempts to gain insight into what radiologists consider explainable. An experiment was conducted with a cohort of radiologists to determine if uncertainty-based explanation methods improve prostate lesion detection. Additionally, a qualitative assessment of each method was conducted to gain insight into what characteristics make an explanation method suitable for radiology end use. It was found that uncertainty-based explanation methods increase lesion detection performance by up to 20%. It was also found that perceived explanation quality is related to actual explanation quality. This pilot study demonstrates the potential use of explanation methods for radiology end use and gleans insight into what radiologists deem explainable.	Proceedings of the 7th Machine Learning for Healthcare Conference	Evaluating Uncertainty-Based Deep Learning Explanations for Prostate Lesion Detection	182	2022	inproceedings	Proceedings of Machine Learning Research	PMLR	2640-3498	trombley22a	0	Evaluating Uncertainty-Based Deep Learning Explanations for Prostate Lesion Detection	874	891	874-891	874	false	Trombley, Christopher M and Gulum, Mehmet Akif and Ozen, Merve and Esen, Enes and Aksamoglu, Melih and Kantardzic, Mehmed	given family Christopher M Trombley given family Mehmet Akif Gulum given family Merve Ozen given family Enes Esen given family Melih Aksamoglu given family Mehmed Kantardzic	2022-12-31		Proceedings of the 7th Machine Learning for Healthcare Conference	inproceedings	date-parts 2022 12 31	https://proceedings.mlr.press/v182/trombley22a/trombley22a.pdf