A Diagnostic Study of Explainability Techniques for Text Classification

Publikation: Bidrag til bog/antologi/rapportKonferencebidrag i proceedingsForskningfagfællebedømt

Dokumenter

Recent developments in machine learning have introduced models that approach human performance at the cost of increased architectural complexity. Efforts to make the rationales behind the models’ predictions transparent have inspired an abundance of new explainability techniques. Provided with an already trained model, they compute saliency scores for the words of an input instance. However, there exists no definitive guide on (i) how to choose such a technique given a particular application task and model architecture, and (ii) the benefits and drawbacks of using each such technique. In this paper, we develop a comprehensive list of diagnostic properties for evaluating existing explainability techniques. We then employ the proposed list to compare a set of diverse explainability techniques on downstream text classification tasks and neural network architectures. We also compare the saliency scores assigned by the explainability techniques with human annotations of salient input regions to find relations between a model’s performance and the agreement of its rationales with human ones. Overall, we find that the gradient-based explanations perform best across tasks and model architectures, and we present further insights into the properties of the reviewed explainability techniques.
OriginalsprogEngelsk
TitelProceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)
ForlagAssociation for Computational Linguistics
Publikationsdato2020
Sider3256-3274
DOI
StatusUdgivet - 2020
BegivenhedThe 2020 Conference on Empirical Methods in Natural Language Processing - online
Varighed: 16 nov. 202020 nov. 2020
http://2020.emnlp.org

Konference

KonferenceThe 2020 Conference on Empirical Methods in Natural Language Processing
Lokationonline
Periode16/11/202020/11/2020
Internetadresse

Antal downloads er baseret på statistik fra Google Scholar og www.ku.dk


Ingen data tilgængelig

ID: 254783374