Table 2.

Multiple taxonomies-part 2

Interpretable	Explainable	Transparent	Intelligible	Refs.
Used interchangeably with explainable	Post-hoc explanations involve an auxiliary method after a model is trained. Self-explaining models generate local explanations that may not be directly interpretable	Not mentioned	A “directly interpretable" model, namely intrinsically understandable by most consumers	Arya et al. (2019)
It is a domain-specific notion that does not allow a general-purpose definition. An interpretable ML model is constrained in model form so that it is either useful to someone, or obeys structural knowledge of the domain [...]	Possibly unreliable and misleading, explanations are not faithful to what the original model computes. Often, they do not make sense nor do they provide enough detail to understand what the black box is doing	Fully transparent models are allowed to understand their variables and the related correlations	Not mentioned.	Rudin (2019)
It refers to the degree of human comprehensibility of a given black-box model or decision	It refers to the numerous ways of exchanging information about a phenomenon (a model’s functionality or the rationale and criteria for a decision) with multiple stakeholders	A model is transparent if its functionality can be comprehended in its entirety by a person	Not mentioned	Mittelstadt et al. (2019)
It is a passive characteristic of a model referring to the level at which it makes sense for a human observer (also referred to as transparency)	Any action or procedure to clarify the internal model functions	As in Lipton, described by Simulability, Decomposability and Algorithmic Transparency	Not mentioned. Understandable is different from intelligible	Chromik and Schuessler (2020)
It encompasses multiple concepts and definitions. Generally, it is associated with models with inherently interpretable behavior	It is intended as the generation of post-hoc explanations for black-box models	It is intended as an explanation of how the system works	Not mentioned	Arrieta et al. (2020)
Assigning meaning to an explanation	Process of describing one or more facts, facilitating the understanding of said facts by a human consumer	Not mentioned	Not mentioned	Palacio et al. (2021)
Assigning a subjective meaning to a model, object, or variable that is possible to be interpreted by the explainee	The activity of producing more interpretable objects manipulating symbolic information	Providing a clear representation of the black-box dynamics	Concerning the explainee, it is intended a successful consumption of an explanation	Ciatto et al. (2020)