Definitions, methods, and applications in interpretable machine learning

M learning (ML) has recently received considerable attention for its ability to accurately predict a wide variety of complex phenomena. However, there is a growing realization that, in addition to predictions, ML models are capable of producing knowledge about domain relationships contained in data, often referred to as interpretations. These interpretations have found uses both in their own right, e.g. medicine (1), policy-making (2), and science (3, 4), as well as in auditing the predictions themselves in response to issues such as regulatory pressure (5) and fairness (6). In the absence of a well-formed definition of interpretability, a broad range of methods with a correspondingly broad range of outputs (e.g. visualizations, natural language, mathematical equations) have been labeled as interpretation. This has led to considerable confusion about the notion of interpretability. In particular, it is unclear what it means to interpret something, what common threads exist among disparate methods, and how to select an interpretation method for a particular problem/audience. In this paper, we attempt to address these concerns. To do so, we first define interpretability in the context of machine learning and place it within a generic data science life cycle. This allows us to distinguish between two main classes of interpretation methods: model-based∗ and post hoc. We then introduce the Predictive, Descriptive, Relevant (PDR) framework, consisting of three desiderata for evaluating and constructing interpretations: predictive accuracy, descriptive