Analytical Learning

This talk is based on

1 Introduction

1.1 New Learning Problem

1.2 EBL Example

1.3 Explanations

1.4 Inductive vs. Analytical Learning

1.5 Example

2 Learning With Perfect Domain Theories

2.1 Prolog-EBG

Prolog-EGB(TargetConcept, TraningExamples, DomainTheory)
  1. LearnedRules = {}
  2. Pos = the positive examples from TraningExamples.
  3. for each PositiveExample in Pos that is not covered by LearnedRules do
    1. Explanation = an explanation in terms of DomainTheory that Pos satisfies the TargetConcept.
    2. SufficientConditions = the most general set of features of PositiveExample sufficient to satisfy the TargetConcept according to the Explanation.
    3. LearnedRules = LearnedRules + {TargetConcept ← SufficientConditions}.
  4. return LearnedRules

2.1.1 Explaining the Example

2.1.2 Analyze the Explanation

2.1.3 Weakest Preimage

2.1.4 Inductive Bias of Prolog-EBG

3 Thinking about EBL

3.1 Knowledge Level Learning

4 EBL of Search Control Knowledge

4.1 PRODIGY

4.2 SOAR

4.3 The Allure of Numbers

URLs

  1. Machine Learning book at Amazon, http://www.amazon.com/exec/obidos/ASIN/0070428077/multiagentcom/
  2. Deep Blue homepage, http://www.research.ibm.com/deepblue/
  3. Prodigy Homepage, http://www.cs.cmu.edu/~prodigy/
  4. Soar Homepage, http://ai.eecs.umich.edu/soar/

This talk available at http://jmvidal.cse.sc.edu/talks/analyticallearning/
Copyright © 2009 José M. Vidal . All rights reserved.

10 April 2003, 12:16PM