Why do you think that? Exploring Faithful Sentence-Level Rationales Without Supervision

TL;DR

This paper introduces a differentiable framework for training models to produce faithful sentence-level rationales without supervision on the rationales, improving interpretability and maintaining competitive accuracy.

Contribution

It presents a novel training method that enables models to generate faithful rationales solely supervised on the target task, avoiding costly rationale annotations.

Findings

Achieves competitive performance with standard BERT models.

Outperforms pipeline-based rationale extraction in two datasets.

Enhances rationale quality through direct supervision.

Abstract

Evaluating the trustworthiness of a model's prediction is essential for differentiating between `right for the right reasons' and `right for the wrong reasons'. Identifying textual spans that determine the target label, known as faithful rationales, usually relies on pipeline approaches or reinforcement learning. However, such methods either require supervision and thus costly annotation of the rationales or employ non-differentiable models. We propose a differentiable training-framework to create models which output faithful rationales on a sentence level, by solely applying supervision on the target task. To achieve this, our model solves the task based on each rationale individually and learns to assign high scores to those which solved the task best. Our evaluation on three different datasets shows competitive results compared to a standard BERT blackbox while exceeding a pipeline counterpart's performance in two cases. We further exploit the transparent decision-making process of these models to prefer selecting the correct rationales by applying direct supervision, thereby boosting the performance on the rationale-level.