Assessing Measures of Explanatory Power

<a href="https://embed.stanford.edu/iframe/?url=https%3A%2F%2Fpurl.stanford.edu%2Fxq488dg8193" class="su-underline">Show Content</a>

Abstract/Contents

Abstract: Measures of explanatory power have a key point of discussion recently in the context of epistemic research into abduction or inference to the best explanation. Many recent works have integrated explanatory power into Bayesian inference rules, arguing that such inference rules converge more quickly or are otherwise better than inference using Bayes' rule. This thesis proposes a pragmatic test, explanation task superiority, which attempts to determine if belief inference rules capture more or less valuable information than one another. This test borrows the decision theory literature's partially-observable Markov decision process (POMDP) formulation to separate epistemic and pragmatic rationality to assess a belief update rule by the value of its epistemic rationality. Through simulation, the thesis then shows that currently proposed inference rules which incorporate explanatory power fail to capture valuable information that inference by Bayes' rule does not and thus are not explanation task superior.

Type of resource	text
Date created	June 4, 2021

Subject	Epistemology
Subject	Philosophy
Subject	Symbolic Systems Program
Subject	School of Humanities & Sciences
Genre	Thesis

Location	https://purl.stanford.edu/xq488dg8193

Use and reproduction: User agrees that, where applicable, content will not be used to identify or to otherwise infringe the privacy or confidentiality rights of individuals. Content distributed via the Stanford Digital Repository may be subject to additional license and use restrictions applied by the depositor.
License: This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).

Preferred Citation: Beasley, Jack and Icard, Thomas and Gerstenberg, Tobias. (2021). Assessing Measures of Explanatory Power. Stanford Digital Repository. Available at: https://purl.stanford.edu/xq488dg8193

Master's Theses, Symbolic Systems Program, Stanford University

Loading usage metrics...