Defining Explanation in Probabilistic Systems

TL;DR

This paper critically examines existing explanation methods for probabilistic systems, identifies their limitations, and proposes a new approach that integrates features from prior models and recent causality research.

Contribution

It introduces a novel framework for defining 'better explanations' in probabilistic systems, combining elements from Gärdenfors, Pearl, and recent causality work.

Findings

Both Gärdenfors and Pearl's approaches have significant issues.

The proposed method offers a more robust way to rank explanations.

Integrates causality concepts to improve explanation quality.

Abstract

As probabilistic systems gain popularity and are coming into wider use, the need for a mechanism that explains the system's findings and recommendations becomes more critical. The system will also need a mechanism for ordering competing explanations. We examine two representative approaches to explanation in the literature - one due to G\"ardenfors and one due to Pearl - and show that both suffer from significant problems. We propose an approach to defining a notion of "better explanation" that combines some of the features of both together with more recent work by Pearl and others on causality.