Wellawatte Geemi P, Seshadri Aditi, White Andrew D
Department of Chemistry, University of Rochester Rochester NY USA.
Department of Chemical Engineering, University of Rochester Rochester NY USA
Chem Sci. 2022 Feb 16;13(13):3697-3705. doi: 10.1039/d1sc05259d. eCollection 2022 Mar 30.
An outstanding challenge in deep learning in chemistry is its lack of interpretability. The inability of explaining why a neural network makes a prediction is a major barrier to deployment of AI models. This not only dissuades chemists from using deep learning predictions, but also has led to neural networks learning spurious correlations that are difficult to notice. Counterfactuals are a category of explanations that provide a rationale behind a model prediction with satisfying properties like providing chemical structure insights. Yet, counterfactuals have been previously limited to specific model architectures or required reinforcement learning as a separate process. In this work, we show a universal model-agnostic approach that can explain any black-box model prediction. We demonstrate this method on random forest models, sequence models, and graph neural networks in both classification and regression.
化学深度学习中的一个突出挑战是其缺乏可解释性。无法解释神经网络为何做出预测是人工智能模型部署的一个主要障碍。这不仅使化学家不愿使用深度学习预测，还导致神经网络学习难以察觉的虚假相关性。反事实解释是一类解释，它为模型预测背后提供了一个基本原理，并具有诸如提供化学结构见解等令人满意的特性。然而，反事实解释以前仅限于特定的模型架构，或者需要强化学习作为一个单独的过程。在这项工作中，我们展示了一种通用的模型无关方法，它可以解释任何黑箱模型的预测。我们在分类和回归中的随机森林模型、序列模型和图神经网络上演示了这种方法。
