Diagnostics-Guided Explanation Generation

Research output: Contribution to journalConference article

Documents

Explanations shed light on a machine learning model's rationales and can aid in identifying deficiencies in its reasoning process. Explanation generation models are typically trained in a supervised way given human explanations. When such annotations are not available, explanations are often selected as those portions of the input that maximise a downstream task's performance, which corresponds to optimising an explanation's Faithfulness to a given model. Faithfulness is one of several so-called diagnostic properties, which prior work has identified as useful for gauging the quality of an explanation without requiring annotations. Other diagnostic properties are Data Consistency, which measures how similar explanations are for similar input instances, and Confidence Indication, which shows whether the explanation reflects the confidence of the model. In this work, we show how to directly optimise for these diagnostic properties when training a model to generate sentence-level explanations, which markedly improves explanation quality, agreement with human rationales, and downstream task performance on three complex reasoning tasks.
Original languageEnglish
JournalProceedings of the International Joint Conference on Artificial Intelligence
Volume36
Issue number10
Pages (from-to)10445-10453.
ISSN1045-0823
DOIs
Publication statusPublished - 2022
Event36th AAAI Conference on Artificial Intelligence (AAAI-22) - Vancouver, BC, Canada
Duration: 28 Feb 20221 Mar 2022

Conference

Conference36th AAAI Conference on Artificial Intelligence (AAAI-22)
CountryCanada
CityVancouver, BC
Period28/02/202201/03/2022

ID: 339344122