Identifying concerns when specifying machine learning-enabled systems: A perspective-based approach

Hugo Villamizar, Marcos Kalinowski, Hélio Lopes, Daniel Méndez

Abstract: Highlights•The perspective-based approach called PerSpecML helps identify concerns involved in the development of ML-enabled systems.•60 concerns related to 28 tasks should be analyzed by business owners, domain experts, designers, software/ML engineers, and data scientists.•Three formative evaluations were conducted in both academia and industry.•PerSpecML was seen as useful by practitioners when specifying ML-enabled systems.