Beneficial and harmful explanatory machine learning

Lun Ai,Céline Hocquette,Ute Schmid,Stephen H Muggleton,Mark Gromowski

doi:10.1007/s10994-020-05941-0

Lun Ai, Céline Hocquette + Show 3 more

Open Access

https://doi.org/10.1007/s10994-020-05941-0

Copy DOI

Journal: Machine Learning	Publication Date: Mar 11, 2021
Citations: 19	License type: open-access

Affiliation: Imperial College London, University of Bamberg

Abstract

Given the recent successes of Deep Learning in AI there has been increased interest in the role and need for explanations in machine learned theories. A distinct notion in this context is that of Michie’s definition of ultra-strong machine learning (USML). USML is demonstrated by a measurable increase in human performance of a task following provision to the human of a symbolic machine learned theory for task performance. A recent paper demonstrates the beneficial effect of a machine learned logic theory for a classification task, yet no existing work to our knowledge has examined the potential harmfulness of machine’s involvement for human comprehension during learning. This paper investigates the explanatory effects of a machine learned theory in the context of simple two person games and proposes a framework for identifying the harmfulness of machine explanations based on the Cognitive Science literature. The approach involves a cognitive window consisting of two quantifiable bounds and it is supported by empirical evidence collected from human trials. Our quantitative and qualitative results indicate that human learning aided by a symbolic machine learned theory which satisfies a cognitive window has achieved significantly higher performance than human self learning. Results also demonstrate that human learning aided by a symbolic machine learned theory that fails to satisfy this window leads to significantly worse performance than unaided human learning.

Highlights

In a recent paper (Muggleton et al 2018) the authors provided an operational definition for comprehensibility of logic programs and used this, in experiments with humans, to provide the first demonstration of Michie’s Ultra-Strong Machine Learning (USML)
Based on the analogy between declarative understanding of a logic program and understanding of a natural language explanation, we describe measures for estimating the degree to which the output of a symbolic machine learning algorithm2 can be simulated by humans and aid comprehension
While the focus of explainable AI approaches has been on explanations of classifications (Adadi and Berrada 2018), we have investigated explanations in the context of game strategy learning

Summary

Introduction

In a recent paper (Muggleton et al 2018) the authors provided an operational definition for comprehensibility of logic programs and used this, in experiments with humans, to provide the first demonstration of Michie’s Ultra-Strong Machine Learning (USML). The authors demonstrated USML via empirical evidence that humans improve out-of-sample. Editors: Nikos Katzouris, Alexander Artikis, Luc De Raedt, Artur d’Avila Garcez, Sebastijan Dumančić, Ute Schmid, Jay Pujara. Extended author information available on the last page of the article. You select this territory and obtain 1 pair (Island 3). Opponent conquers and prevent you from getting a triplet (Island 3)

Methods

Findings

Discussion

Conclusion