ERASER: A Benchmark to Evaluate Rationalized NLP Models

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4443–4458 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 4443 ERASER : A Benchmark to Evaluate Rationalized NLP Models Jay DeYoung⋆Ψ, Sarthak Jain⋆Ψ, Nazneen Fatema Rajani⋆Φ, Eric LehmanΨ, Caiming XiongΦ, Richard SocherΦ, and Byron C. WallaceΨ ⋆Equal contribution. ΨKhoury College of Computer Sciences, Northeastern University ΦSalesforce Research, Palo Alto, CA, 94301 Abstract State-of-the-art models in NLP are now pre- dominantly based on deep neural networks that are opaque in terms of how they come to make predictions. This limitation has increased interest in designing more inter- pretable deep models for NLP that reveal the ‘reasoning’ behind model outputs. But work in this direction has been conducted on dif- ferent datasets and tasks with correspondingly unique aims and metrics; this makes it difficult to track progress. We propose the Evaluating Rationales And Simple English Reasoning (ERASER ) benchmark to advance research on interpretable models in NLP. This bench- mark comprises multiple datasets and tasks for which human annotations of “rationales” (sup- porting evidence) have been collected. We pro- pose several metrics that aim to capture how well the rationales provided by models align with human rationales, and also how faithful these rationales are (i.e., the degree to which provided rationales influenced the correspond- ing predictions). Our hope is that releasing this benchmark facilitates progress on designing more interpretable NLP systems. The bench- mark, code, and documentation are available at https://www.eraserbenchmark.com/ 1 Introduction Interest has recently grown in designing NLP sys- tems that can reveal why models make specific predictions. But work in this direction has been conducted on different datasets and using different metrics to quantify performance; this has made it difficult to compare methods and track progress. We aim to address this issue by releasing a stan- dardized benchmark of datasets — repurposed and augmented from pre-existing corpora, spanning a range of NLP tasks — and associated metrics for measuring different properties of rationales. We re- fer to this as the Evaluating Rationales And Simple English Reasoning (ERASER ) benchmark. Commonsense Explanations (CoS-E) Where do you find the most amount of leafs? (a) Compost pile (b) Flowers (c) Forest (d) Field (e) Ground Movie Reviews In this movie, … Plots to take over the world. The acting is great! The soundtrack is run-of-the-mill, but the action more than makes up for it (a) Positive (b) Negative Evidence Inference Article Patients for this trial were recruited … Compared with 0.9% saline, 120 mg of inhaled nebulized furosemide had no effect on breathlessness during exercise. (a) Sig. decreased (b) No sig. difference (c) Sig. increased Prompt With respect to breathlessness, what is the reported difference between patients receiving placebo and those receiving furosemide? e-SNLI H A man in an orange vest leans over a pickup truck P A man is touching a truck (a) Entailment (b) Contradiction (c) Neutral Figure 1: Examples of instances, labels, and rationales illustrative of four (out of seven) datasets included in ERASER. …

Content truncated to 500 words