Luzi, Simone (A.A. 2024/2025) Explainable AI: opening the black box for reliable decision making. Tesi di Laurea in Machine learning, Luiss Guido Carli, relatore Giuseppe Francesco Italiano, pp. 175. [Master's Degree Thesis]

PDF (Full text) Restricted to Registered users only Download (5MB) | Request a copy

Abstract/Index

Fundamentals of ML and XAI. Historical roots and learning paradigms. The transparency spectrum. Why explainability matters: the rationale for XAI. Taxonomy and evaluation metrics. The landscape of interpretability methods. Ante-hoc interpretability. Post-hoc attribution. Counterfactuals and actionable recourse. Attention as an explanation. Verifiable explainability: tracing the architecture. When explanations heal or deceive. Anatomy of models: interpretability frontiers. Failure models. Computational budget of an explanation. Governance and compliance. Principles and compliance tools. Decision chains and signing responsibilities. Explainability Ops. Explainability assurance: quality, privacy and security. Multi-level explainability: opening ResNet-18_224 on BreastMNIST. Operational context: data, pre-processing and environment. Model, training strategy and calibration plan. Concept-based explanations.

References

Bibliografia: pp. 166-174.

Thesis Type:	Master's Degree Thesis
Institution:	Luiss Guido Carli
Degree Program:	Master's Degree Programs > Master's Degree Program in Data Science e Management (LM-91)
Chair:	Machine learning
Thesis Supervisor:	Italiano, Giuseppe Francesco
Thesis Co-Supervisor:	Coppa, Emilio
Academic Year:	2024/2025
Session:	Autumn
Deposited by:	Alessandro Perfetti
Date Deposited:	19 Mar 2026 11:12
Last Modified:	19 Mar 2026 11:12
URI:	https://tesi.luiss.it/id/eprint/45198

Downloads

Repository Staff Only

View Item