学会用结构忠实地表达
Learning to Faithfully Rationalize by Construction
摘要：
在许多情况下，重要的是要能够理解为什么一个模型做出了特定的预测。在NLP中，这通常需要提取“负责”相应模型输出的输入文本的片段;当这样的片段包含确实告知模型预测的标记时，它是一个忠实的解释。在某些情况下，忠实可能是确保透明度的关键。Lei et al.（2016）提出了一种模型，通过定义独立的片段提取和预测模块，为神经文本分类提供可靠的理论基础。然而，这种方法对输入标记进行的离散选择使训练复杂化，导致高方差并需要仔细的超参数调整。我们提出了一个更简单的变体，这种方法提供了忠实的解释建设。在我们的方案中，命名为FRESH，任意特征重要性得分（例如，来自训练模型的梯度）用于在令牌输入上引入二进制标签，提取器可以被训练来预测该二进制标签。然后，一个独立的分类器模块专门训练提取器提供的片段;这些片段因此构成忠实的解释，即使分类器是任意复杂的。在自动和手动评估中，我们发现这个简单框架的变体产生优于“端到端”方法的预测性能，同时更通用，更容易训练。代码可在https://github.com/successar/FRESH上获得
英文摘要：
In many settings it is important for one to be able to understand why a model made a particular prediction. In NLP this often entails extracting snippets of an input text `responsible for' corresponding model output; when such a snippet comprises tokens that indeed informed the model's prediction, it is a faithful explanation. In some settings, faithfulness may be critical to ensure transparency. Lei et al. (2016) proposed a model to produce faithful rationales for neural text classification by defining independent snippet extraction and prediction modules. However, the discrete selection over input tokens performed by this method complicates training, leading to high variance and requiring careful hyperparameter tuning. We propose a simpler variant of this approach that provides faithful explanations by construction. In our scheme, named FRESH, arbitrary feature importance scores (e.g., gradients from a trained model) are used to induce binary labels over token inputs, which an extractor can be trained to predict. An independent classifier module is then trained exclusively on snippets provided by the extractor; these snippets thus constitute faithful explanations, even if the classifier is arbitrarily complex. In both automatic and manual evaluations we find that variants of this simple framework yield predictive performance superior to `end-to-end' approaches, while being more general and easier to train. Code is available at https://github.com/successar/FRESH
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
