Explanation in Artificial Intelligence: A Human-Centred Approach

Grant number: DP190103414 | Funding period: 2019 - 2022

Completed

Abstract

This project aims to produce validated methods for creating human-centred explanations of decisions made by artificial intelligence (AI). Trial deployment of AI devices has resulted in the requirement for explanations of how AI makes decisions, where developed AI systems gave insufficient consideration of how decision logic would be explained to people. This project positions 'explainable AI' within the intersection of human-computer interaction, computer science and cognitive psychology. The expected outcomes of this project are new methods, models and algorithms for explaining different types of AI models to people. This project should result in improved understanding and trust of decision..

View full description

Related publications (3)

Conceptualising Contestability: Perspectives on Contesting Algorithmic Decisions

H Lyons, E Velloso, T Miller

2021-04-22

As the use of algorithmic systems in high-stakes decision-making increases, the ability to contest algorithmic decisions is being ..

Invertible Concept-based Explanations for CNN Models with Non-negative Concept Activation Vectors

Ruihan Zhang, Prashan Madumal, Tim Miller, Krista A Ehinger, Benjamin IP Rubinstein

2021-01-01

Convolutional neural network (CNN) models for computer vision are powerful but lack explainability in their most basic form. This ..

Explainable Reinforcement Learning through a Causal Lens

Prashan Madumal, Tim Miller, Liz Sonenberg, Frank Vetere

2020-01-01

Prominent theories in cognitive science propose that humans understand and represent the knowledge of the world through causal rel..

University of Melbourne Researchers

Frank Vetere Computing and Information Systems

Timothy Miller Computing and Information Systems

Eduardo Velloso Computing and Information Systems

Piers Howe Melbourne School of Psychological Sciences