Exploiting Explainable AI tools for improving model classification explainability

Rodriguez Lopez, Ivan Misael (2025) Exploiting Explainable AI tools for improving model classification explainability. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (2MB) | Preview
Preview	PDF (Configuration Manual) Download (2MB) | Preview

Abstract

As the black-box nature of complex AI models presents challenges for trust and debugging, the field of Explainable AI (XAI) seeks to provide methods for interpreting their decisions. This research investigates how prominent XAI tools can be exploited to improve the explainability of model classifications for stakeholders.

Three XAI tools, LIME, SHAP, and Grad-Cam, were applied to three different CNN models built for healthcare image classification: diabetic retinopathy,
COVID19, and blood cells. The tools were evaluated and compared the basis of their performance and the quality of their visual explanations.

The findings show that no single tool is superior in all cases. However, for medical imaging, a combination of Grad-Cam and SHAP provides the most effective explanation; Grad-Cam offers a high-level visual heatmap of relevant areas, while SHAP adds valuable pixel-level precision. Although these tools require additional computational resources, they significantly enhance model transparency and can help justify diagnostic decisions.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Vamadevan, Arundev UNSPECIFIED
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Artificial Intelligence
Depositing User:	Ciara O'Brien
Date Deposited:	12 Aug 2026 09:56
Last Modified:	12 Aug 2026 09:56
URI:	https://norma.ncirl.ie/id/eprint/9513

Actions (login required)

View Item