Explaining Predictions from Machine Learning Models: Algorithms, Users, and Pedagogy

TL;DR

This paper explores the explainability of machine learning models from algorithms, user perspectives, and educational methods, offering new solutions to improve understanding and trust in ML predictions.

Contribution

It introduces novel solutions addressing ML explainability from multiple perspectives, enhancing user understanding and model transparency.

Findings

Proposes new algorithms for explainability

Analyzes user needs and comprehension strategies

Develops pedagogical approaches for teaching ML explainability

Abstract

Model explainability has become an important problem in machine learning (ML) due to the increased effect that algorithmic predictions have on humans. Explanations can help users understand not only why ML models make certain predictions, but also how these predictions can be changed. In this thesis, we examine the explainability of ML models from three vantage points: algorithms, users, and pedagogy, and contribute several novel solutions to the explainability problem.