An Initial Study of Machine Learning Underspecification Using Feature Attribution Explainable AI Algorithms: A COVID-19 Virus Transmission Case Study

James Hinns, Siyuan Liu, Veera Raghava Reddy Kovvuri, Mehmet Orcun Yalcin, Markus Roggenbach

Research output: Chapter in Book/Report/Conference proceedingChapterAcademic


From a dataset, one can construct different machine learning (ML) models with different parameters and/or inductive biases. Although these models give similar prediction performances when tested on data that are currently available, they may not generalise equally well on unseen data. The existence of multiple equally performing models exhibits underspecification of the ML pipeline used for producing such models. In this work, we propose identifying underspecification using feature attribution algorithms developed in Explainable AI. Our hypothesis is: by studying the range of explanations produced by ML models, one can identify underspecification. We validate this by computing explanations using the Shapley additive explainer and then measuring statistical correlations between them. We experiment our approach on multiple datasets drawn from the literature, and in a COVID-19 virus transmission case study.
Original languageEnglish
Title of host publicationPRICAI 2021: Trends in Artificial Intelligence
EditorsDuc Ngia Pham, Thanaruk Theeramunkong, Guido Governatori, Fenrong Liu
PublisherSpringer, Cham
Number of pages12
ISBN (Electronic)978-3-030-89188-6
ISBN (Print)978-3-030-89187-9
Publication statusPublished - 25 Oct 2021
Event18th Pacific Rim International Conference on Artificial Intelligence - Hanoi, Viet Nam
Duration: 8 Nov 202112 Nov 2021

Publication series

SeriesLecture Notes in Computer Science


Conference18th Pacific Rim International Conference on Artificial Intelligence
Abbreviated titlePRICAI 2021
Country/TerritoryViet Nam

Cite this