A domain-specific language for describing machine learning datasets

Abstract

Datasets are essential for training and evaluating machine learning (ML) models. However, they are also at the root of many undesirable model behaviors, such as biased predictions. To address this issue, the machine learning community is proposing a data-centric cultural shift, where data issues are given the attention they deserve and more standard practices for gathering and describing datasets are discussed and established. So far, these proposals are mostly high-level guidelines described in natural language and, as such, they are difficult to formalize and apply to particular datasets. In this sense, and inspired by these proposals, we define a new domain-specific language (DSL) to precisely describe machine learning datasets in terms of their structure, provenance, and social concerns. We believe this DSL will facilitate any ML initiative to leverage and benefit from this data-centric shift in ML (e.g., selecting the most appropriate dataset for a new project or better replicating other ML results). The DSL is implemented as a Visual Studio Code plugin, and it has been published under an open-source license.