Measuring Technical Debt in AI-Based Competition Platforms

TL;DR

This paper identifies, categorizes, and proposes a framework for measuring and managing technical debt in AI-based competition platforms to enhance their sustainability and quality.

Contribution

It introduces a novel categorization of technical debt specific to AI competition platforms and develops a questionnaire for assessment.

Findings

Categorized types of technical debt in AI systems.

Developed a questionnaire for technical debt assessment.

Highlighted challenges related to Accessibility Debt.

Abstract

Advances in AI have led to new types of technical debt in software engineering projects. AI-based competition platforms face challenges due to rapid prototyping and a lack of adherence to software engineering principles by participants, resulting in technical debt. Additionally, organizers often lack methods to evaluate platform quality, impacting sustainability and maintainability. In this research, we identify and categorize types of technical debt in AI systems through a scoping review. We develop a questionnaire for assessing technical debt in AI competition platforms, categorizing debt into various types, such as algorithm, architectural, code, configuration, data etc. We introduce Accessibility Debt, specific to AI competition platforms, highlighting challenges participants face due to inadequate platform usability. Our framework for managing technical debt aims to improve the sustainability and effectiveness of these platforms, providing tools for researchers, organizers, and participants.