A DSL for Explaining Probabilistic Reasoning

Abstract

Abstract. We propose a new focus in language design where languages provide constructs that not only describe the computation of results, but also produce explanations of how and why those results were obtained. We posit that if users are to understand computations produced by a language, that language should provide explanations to the user.As an example of such an explanation-oriented language we present a domain-specific language for explaining probabilistic reasoning, a domain that is not well understood by non-experts. We show the design of the DSL in several steps. Based on a story-telling metaphor of explanations, we identify generic constructs for building stories out of events, and obtaining explanations by applying stories to specific examples. These generic constructs are then adapted to the particular explanation domain of probabilistic reasoning. Finally, we develop a visual notation for explaining probabilistic reasoning.

Visual explanations of probabilistic reasoning

“…In this language, probabilistic values are represented as probability distributions, essentially lists of values paired with associated probabilities. Building on this work, we designed another DSEL for describing explanations of probabilistic reasoning which is presented in [5], which also contains the initial ideas for the visual notation presented here.…”

“…For a complete listing, please refer to our previous work in [5]. For each operation we give a high-level description and define an explanation semantics which illustrates the evolution of distributions over the operation.…”

“…Each operation has a single distribution as input and a single distribution as output, allowing only the possibility of linear graphs. In [5] we discuss branching operations which would make these probability flow graphs somewhat more complex.…”

A visual language for explaining probabilistic reasoning

“…This work is an extension and consolidation of two earlier papers on explaining probabilistic reasoning [12,13]. Prior to these papers, we presented a domain-specific embedded language (DSEL) in Haskell for creating and manipulating (but not explaining) probabilistic values [10].…”

“…Prior to these papers, we presented a domain-specific embedded language (DSEL) in Haskell for creating and manipulating (but not explaining) probabilistic values [10]. In [12] we reused this DSEL in the design of a second Haskell DSEL for explaining probabilistic reasoning that became the basis for Probula. That paper also provides the first examples of Probula's visual notation along with an extended discussion of philosophical research on explanations and a lengthier motivation for the story-telling metaphor than is provided here.…”