Assessing the Use of AutoML for Data-Driven Software Engineering

TL;DR

This paper addresses the adoption and effectiveness of AutoML in data-driven software engineering by combining a 12-tool AutoML benchmark on SE sentiment/emotion datasets with a survey and interviews of software engineers. It finds that AutoML can reach or surpass SE-specific baselines in text classification while automation remains uneven across the ML workflow, excelling in analysis but offering limited data-preparation and deployment support. The results yield actionable insights for SE researchers and AutoML tool builders, highlighting the need for human-in-the-loop designs, better dissemination automation, and workflow-aware tooling. Overall, the work suggests AutoML as a practical starting point for SE research and a misnomer for broader audiences, acting as a generator of tailored workflows rather than a fully automated, all-in-one solution.

Abstract

Background. Due to the widespread adoption of Artificial Intelligence (AI) and Machine Learning (ML) for building software applications, companies are struggling to recruit employees with a deep understanding of such technologies. In this scenario, AutoML is soaring as a promising solution to fill the AI/ML skills gap since it promises to automate the building of end-to-end AI/ML pipelines that would normally be engineered by specialized team members. Aims. Despite the growing interest and high expectations, there is a dearth of information about the extent to which AutoML is currently adopted by teams developing AI/ML-enabled systems and how it is perceived by practitioners and researchers. Method. To fill these gaps, in this paper, we present a mixed-method study comprising a benchmark of 12 end-to-end AutoML tools on two SE datasets and a user survey with follow-up interviews to further our understanding of AutoML adoption and perception. Results. We found that AutoML solutions can generate models that outperform those trained and optimized by researchers to perform classification tasks in the SE domain. Also, our findings show that the currently available AutoML solutions do not live up to their names as they do not equally support automation across the stages of the ML development workflow and for all the team members. Conclusions. We derive insights to inform the SE research community on how AutoML can facilitate their activities and tool builders on how to design the next generation of AutoML technologies.

Figures (3)

• Figure 1: A typical ML workflow with activities and roles (adapted from Amershi2019Amershi2019).
• Figure 2: The two experimental stages executed in the study.
• Figure 3: Features more relevant for choosing an AutoML solution.