Modular Belief Updates and Confusion about Measures of Certainty in Artificial Intelligence Research

TL;DR

This paper examines the concept of modular belief updates in AI, highlighting their historical roots, defining their properties, and critiquing their misuse in expert systems, thereby clarifying their role in reasoning under uncertainty.

Contribution

It formally defines modular belief updates, explores their historical context, and critiques their improper application in influential AI expert systems.

Findings

Modular updates are rooted in 19th-century probability theory.

Misuse of modular updates occurs in two influential expert systems.

Clarification of belief update measures enhances understanding of reasoning under uncertainty.

Abstract

Over the last decade, there has been growing interest in the use or measures or change in belief for reasoning with uncertainty in artificial intelligence research. An important characteristic of several methodologies that reason with changes in belief or belief updates, is a property that we term modularity. We call updates that satisfy this property modular updates. Whereas probabilistic measures of belief update - which satisfy the modularity property were first discovered in the nineteenth century, knowledge and discussion of these quantities remains obscure in artificial intelligence research. We define modular updates and discuss their inappropriate use in two influential expert systems.