Titulo Estágio
Explainable Artificial Intelligence Model for Decision Support in Software Development
Local do Estágio
CISUC/DEI
Enquadramento
Modern software engineering involves a continuous stream of decision-making activities throughout the software development lifecycle, from initial planning and architecture to maintenance and evolution. As software systems grow in complexity and development practices become more agile and dynamic, software teams face a growing number of decisions that must be made rapidly, accurately, and in context.
Recent research has explored the use of Artificial Intelligence (AI) to support decision-making in software engineering. For instance, Alcaide et al. [1] investigated the application of explainable AI to assist software modelers in understanding complex models, while Chen et al. [2] applied reinforcement learning to improve effort estimation for software tasks. These studies highlight the promise of AI in enhancing decision quality and transparency.
Despite these advances, a fundamental challenge remains: how to effectively automate and contextualize decision-making throughout the software development lifecycle. Existing solutions are often isolated, addressing specific tasks or phases, and rarely provide seamless, explainable, and adaptive decision support across the entire workflow.
This can lead to delays in project delivery, developer overload due to poorly balanced task allocation, accumulation of risk in critical software components, decreased software quality and developer satisfaction.
To address this challenge, there is a clear need for intelligent, context-aware, and explainable decision support systems that can be embedded in development workflows. AI-powered decision models with explainability capabilities offer a promising approach to improving how decisions are made, understood, and trusted by software teams.
Objetivo
This master’s thesis aims to investigate the development of explainable AI-based decision models as intelligent decision-making agents integrated into software development workflows. The goal is to make decision-making more contextualized, transparent, and adaptable, thus enhancing the quality and traceability of decisions in software projects.
Based on this research aim, the following research objectives can be identified:
1. To conduct a critical review of the literature on AI explainability, recommendation systems, reinforcement learning, and decision support in software engineering.
2. To identify key decision points in the software development process that could benefit from automated and explainable support (e.g., task assignment, backlog prioritization, architectural trade-offs).
3. To design and implement a prototype of an intelligent, explainable decision-making agent (available datasets such as that of [3]).
4. To integrate the prototype into a real or simulated development workflow, allowing for the collection of contextual data and real-time decisions for fine tunning the model.
5. To evaluate the effectiveness of the proposed model in terms of decision accuracy, transparency, user trust, and adaptability.
Plano de Trabalhos - Semestre 1
1- State of the art [Sept – Oct]
2- Problem statement, research aims and objectives [Nov]
3- Design and first implementation of the AI system [Nov – Jan]
4- Thesis proposal writing [Dec – Jan]
Plano de Trabalhos - Semestre 2
5- Improvement of the AI system [Feb – Apr]
6- Experimental Tests [Apr – May]
7- Paper writing [May – Jun]
8- Thesis writing [Jan – Jul]
Condições
The work should take place at the Centre for Informatics and Systems of the University of Coimbra (CISUC) at the Department of Informatics Engineering of the University of Coimbra.
Observações
References
[1] F. J. Alcaide, J. R. Romero, and A. Ramírez, “Can explainable artificial intelligence support software modelers in model comprehension?”, Software and Systems Modeling, vol. 23, no. 3, pp. 615–640, May 2025. doi: 10.1007/s10270-024-01251-4.
Available: https://link.springer.com/article/10.1007/s10270-024-01251-4
[2] Z. Zhong, Y. Wang, X. Li, J. Zhang, and Q. Liu, “Enhancing software effort estimation through reinforcement learning,” Int. J. Managing Projects Bus., ahead-of-print, Mar. 2025. doi: 10.1108/IJMPB-03-2024-0065.
Available: https://www.emerald.com/insight/content/doi/10.1108/ijmpb-03-2024-0065/full/html
[3] Perez, Q., Urtado, C., & Vauttier, S. (2023). Dataset of open-source software developers labeled by their experience level in the project and their associated software metrics. Data in Brief, 46, 108842. https://doi.org/10.1016/j.dib.2022.108842
Orientador
Luís Miguel Machado Lopes Macedo (co-orientação com Naghmeh Ivaki)
macedo@dei.uc.pt 📩