Thesis Proposes Methods to Identify and Fix Technical Debt in AI-Cyber-Physical Systems
A recent thesis in software engineering tackles the escalating issue of technical debt (TD) within AI-intensive Cyber-Physical Systems (AI-CPS), which merge hardware, AI elements, and traditional software components. This research, presented on arXiv, identifies the distinct technical debt that emerges in these systems, noting its increased complexity compared to standard software due to AI integration. The initial phase of the study includes examining AI ecosystems and AI-CPS repositories, along with developer interviews, to grasp the essence of this debt. The author suggests strategies for recognizing and alleviating AI-CPS TD based on these insights. Ultimately, the aim is to create an automated tool for managing and repaying this debt, particularly relevant in fields like autonomous vehicles, robotics, and healthcare. The work falls under the Computer Science > Software Engineering category on arXiv and includes references and tools for further investigation.
Key facts
- The thesis focuses on technical debt in AI-intensive Cyber-Physical Systems (AI-CPS).
- AI-CPS combine hardware, AI components, and conventional modules.
- The research characterizes AI-CPS technical debt by analyzing AI ecosystems and repositories.
- Interviews with developers are part of the data collection.
- The study proposes approaches to identify and mitigate AI-CPS technical debt.
- An automated tool supporting agentic AI solutions is planned for development.
- AI-CPS are used in autonomous vehicles, industry, home automation, robotics, and healthcare.
- The thesis is categorized under Computer Science > Software Engineering on arXiv.
Entities
Institutions
- arXiv