Explainable Artificial Intelligence for Breast Tumour Classification: Helpful or Harmful

Amy Rafferty; Rudolf Nenutil; Ajitha Rajan

doi:10.1007/978-3-031-17976-1_10

Explainable Artificial Intelligence for Breast Tumour Classification: Helpful or Harmful

Amy Rafferty, Rudolf Nenutil, Ajitha Rajan

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Explainable Artificial Intelligence (XAI) is the field of AI dedicated to promoting trust in machine learning models by helping us to understand how they make their decisions. For example, image explanations show us which pixels or segments were deemed most important by a model for a particular classification decision. This research focuses on image explanations generated by LIME, RISE and SHAP for a model which classifies breast mammograms as either benign or malignant. We assess these XAI techniques based on (1) the extent to which they agree with each other, as decided by One-Way ANOVA, Kendall’s Tau and RBO statistical tests, and (2) their agreement with the diagnostically important areas as identified by a radiologist on a small subset of mammograms. The main contribution of this research is the discovery that the 3 techniques consistently disagree both with each other and with the medical truth. We argue that using these off-shelf techniques in a medical context is not a feasible approach, and discuss possible causes of this problem, as well as some potential solutions

Original language	English
Title of host publication	Interpretability of Machine Intelligence in Medical Image Computing: 5th International Workshop, iMIMIC 2022, Held in Conjunction with MICCAI 2022, Singapore, Singapore, September 22, 2022, Proceedings
Editors	Mauricio Reyes, Pedro Henriques Abreu, Jaime Cardoso
Publisher	Springer, Cham
Pages	104-123
Number of pages	18
ISBN (Electronic)	978-3-031-17976-1
ISBN (Print)	978-3-031-17975-4
DOIs	https://doi.org/10.1007/978-3-031-17976-1_10
Publication status	Published - 7 Oct 2022
Event	Workshop on Interpretability of Machine Intelligence in Medical Image Computing at MICCAI 2022 - , Singapore Duration: 22 Sep 2022 → 22 Sep 2022 https://imimic-workshop.com/

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer Cham
Volume	13611
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Workshop

Workshop	Workshop on Interpretability of Machine Intelligence in Medical Image Computing at MICCAI 2022
Abbreviated title	iMIMIC 2022
Country/Territory	Singapore
Period	22/09/22 → 22/09/22
Internet address	https://imimic-workshop.com/

Keywords

Machine Learning
Breast Tumour Classification
Explainable AI
LIME
RISE
SHAP

Access to Document

10.1007/978-3-031-17976-1_10Licence: All Rights Reserved

Explainable Artifcial_RAFFERTY_DOA20072022_AFVAccepted author manuscript, 4.45 MBLicence: Creative Commons: Attribution (CC-BY)

https://link.springer.com/chapter/10.1007/978-3-031-17976-1_10Licence: All Rights Reserved

Cite this

Rafferty, A., Nenutil, R., & Rajan, A. (2022). Explainable Artificial Intelligence for Breast Tumour Classification: Helpful or Harmful. In M. Reyes, P. H. Abreu, & J. Cardoso (Eds.), Interpretability of Machine Intelligence in Medical Image Computing: 5th International Workshop, iMIMIC 2022, Held in Conjunction with MICCAI 2022, Singapore, Singapore, September 22, 2022, Proceedings (pp. 104-123). (Lecture Notes in Computer Science; Vol. 13611). Springer, Cham. https://doi.org/10.1007/978-3-031-17976-1_10

Rafferty, Amy ; Nenutil, Rudolf ; Rajan, Ajitha. / Explainable Artificial Intelligence for Breast Tumour Classification: Helpful or Harmful. Interpretability of Machine Intelligence in Medical Image Computing: 5th International Workshop, iMIMIC 2022, Held in Conjunction with MICCAI 2022, Singapore, Singapore, September 22, 2022, Proceedings. editor / Mauricio Reyes ; Pedro Henriques Abreu ; Jaime Cardoso. Springer, Cham, 2022. pp. 104-123 (Lecture Notes in Computer Science).

@inproceedings{3b1f99021bb84bf2a1857b053f7eb81d,

title = "Explainable Artificial Intelligence for Breast Tumour Classification: Helpful or Harmful",

abstract = "Explainable Artificial Intelligence (XAI) is the field of AI dedicated to promoting trust in machine learning models by helping us to understand how they make their decisions. For example, image explanations show us which pixels or segments were deemed most important by a model for a particular classification decision. This research focuses on image explanations generated by LIME, RISE and SHAP for a model which classifies breast mammograms as either benign or malignant. We assess these XAI techniques based on (1) the extent to which they agree with each other, as decided by One-Way ANOVA, Kendall{\textquoteright}s Tau and RBO statistical tests, and (2) their agreement with the diagnostically important areas as identified by a radiologist on a small subset of mammograms. The main contribution of this research is the discovery that the 3 techniques consistently disagree both with each other and with the medical truth. We argue that using these off-shelf techniques in a medical context is not a feasible approach, and discuss possible causes of this problem, as well as some potential solutions",

keywords = "Machine Learning, Breast Tumour Classification, Explainable AI, LIME, RISE, SHAP",

author = "Amy Rafferty and Rudolf Nenutil and Ajitha Rajan",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; Workshop on Interpretability of Machine Intelligence in Medical Image Computing at MICCAI 2022<br/>, iMIMIC 2022 ; Conference date: 22-09-2022 Through 22-09-2022",

year = "2022",

month = oct,

day = "7",

doi = "10.1007/978-3-031-17976-1_10",

language = "English",

isbn = "978-3-031-17975-4",

series = "Lecture Notes in Computer Science",

publisher = "Springer, Cham",

pages = "104--123",

editor = "Mauricio Reyes and Abreu, {Pedro Henriques} and Jaime Cardoso",

booktitle = "Interpretability of Machine Intelligence in Medical Image Computing: 5th International Workshop, iMIMIC 2022, Held in Conjunction with MICCAI 2022, Singapore, Singapore, September 22, 2022, Proceedings",

url = "https://imimic-workshop.com/",

}

Rafferty, A, Nenutil, R & Rajan, A 2022, Explainable Artificial Intelligence for Breast Tumour Classification: Helpful or Harmful. in M Reyes, PH Abreu & J Cardoso (eds), Interpretability of Machine Intelligence in Medical Image Computing: 5th International Workshop, iMIMIC 2022, Held in Conjunction with MICCAI 2022, Singapore, Singapore, September 22, 2022, Proceedings. Lecture Notes in Computer Science, vol. 13611, Springer, Cham, pp. 104-123, Workshop on Interpretability of Machine Intelligence in Medical Image Computing at MICCAI 2022
, Singapore, 22/09/22. https://doi.org/10.1007/978-3-031-17976-1_10

Explainable Artificial Intelligence for Breast Tumour Classification: Helpful or Harmful. / Rafferty, Amy; Nenutil, Rudolf; Rajan, Ajitha.

Interpretability of Machine Intelligence in Medical Image Computing: 5th International Workshop, iMIMIC 2022, Held in Conjunction with MICCAI 2022, Singapore, Singapore, September 22, 2022, Proceedings. ed. / Mauricio Reyes; Pedro Henriques Abreu; Jaime Cardoso. Springer, Cham, 2022. p. 104-123 (Lecture Notes in Computer Science; Vol. 13611).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Explainable Artificial Intelligence for Breast Tumour Classification: Helpful or Harmful

AU - Rafferty, Amy

AU - Nenutil, Rudolf

AU - Rajan, Ajitha

PY - 2022/10/7

Y1 - 2022/10/7

N2 - Explainable Artificial Intelligence (XAI) is the field of AI dedicated to promoting trust in machine learning models by helping us to understand how they make their decisions. For example, image explanations show us which pixels or segments were deemed most important by a model for a particular classification decision. This research focuses on image explanations generated by LIME, RISE and SHAP for a model which classifies breast mammograms as either benign or malignant. We assess these XAI techniques based on (1) the extent to which they agree with each other, as decided by One-Way ANOVA, Kendall’s Tau and RBO statistical tests, and (2) their agreement with the diagnostically important areas as identified by a radiologist on a small subset of mammograms. The main contribution of this research is the discovery that the 3 techniques consistently disagree both with each other and with the medical truth. We argue that using these off-shelf techniques in a medical context is not a feasible approach, and discuss possible causes of this problem, as well as some potential solutions

AB - Explainable Artificial Intelligence (XAI) is the field of AI dedicated to promoting trust in machine learning models by helping us to understand how they make their decisions. For example, image explanations show us which pixels or segments were deemed most important by a model for a particular classification decision. This research focuses on image explanations generated by LIME, RISE and SHAP for a model which classifies breast mammograms as either benign or malignant. We assess these XAI techniques based on (1) the extent to which they agree with each other, as decided by One-Way ANOVA, Kendall’s Tau and RBO statistical tests, and (2) their agreement with the diagnostically important areas as identified by a radiologist on a small subset of mammograms. The main contribution of this research is the discovery that the 3 techniques consistently disagree both with each other and with the medical truth. We argue that using these off-shelf techniques in a medical context is not a feasible approach, and discuss possible causes of this problem, as well as some potential solutions

KW - Machine Learning

KW - Breast Tumour Classification

KW - Explainable AI

KW - LIME

KW - RISE

KW - SHAP

U2 - 10.1007/978-3-031-17976-1_10

DO - 10.1007/978-3-031-17976-1_10

M3 - Conference contribution

SN - 978-3-031-17975-4

T3 - Lecture Notes in Computer Science

SP - 104

EP - 123

BT - Interpretability of Machine Intelligence in Medical Image Computing: 5th International Workshop, iMIMIC 2022, Held in Conjunction with MICCAI 2022, Singapore, Singapore, September 22, 2022, Proceedings

A2 - Reyes, Mauricio

A2 - Abreu, Pedro Henriques

A2 - Cardoso, Jaime

PB - Springer, Cham

T2 - Workshop on Interpretability of Machine Intelligence in Medical Image Computing at MICCAI 2022<br/>

Y2 - 22 September 2022 through 22 September 2022

ER -

Rafferty A, Nenutil R, Rajan A. Explainable Artificial Intelligence for Breast Tumour Classification: Helpful or Harmful. In Reyes M, Abreu PH, Cardoso J, editors, Interpretability of Machine Intelligence in Medical Image Computing: 5th International Workshop, iMIMIC 2022, Held in Conjunction with MICCAI 2022, Singapore, Singapore, September 22, 2022, Proceedings. Springer, Cham. 2022. p. 104-123. (Lecture Notes in Computer Science). https://doi.org/10.1007/978-3-031-17976-1_10

Explainable Artificial Intelligence for Breast Tumour Classification: Helpful or Harmful

Abstract

Publication series

Workshop

Keywords

Access to Document

Fingerprint

Cite this