Based in Naples, Italy. MSc in Applied Computer Science (Machine Learning & Big Data) from the University of Naples "Parthenope", graduated with 110/110 Summa Cum Laude (April 2026), after earning my BSc with 110/110 Summa Cum Laude with Special Mention.
-
Master's Thesis (CI&SS Lab): Feature-Driven Bias Detection: An Association Rule Mining Approach to Analyze Feature Importance. A pipeline combining counterfactual-based feature importance (BoCSoR, adapted to fully categorical data) with hierarchical Association Rule Mining (FP-Growth) to produce global, human-readable explanations of tabular classifiers. Validated on five U.S. Census ACS 2024 benchmarks (~1.75M records) with CatBoost and MLP. → Source code
-
Generative AI for tabular data: TabDiff-based diffusion model for synthetic data augmentation on mixed numerical/categorical census data, with classifier-free guidance and EMA. Distributional fidelity validated via Pearson/Spearman correlation and fairness metrics (Equalized Odds, Error Rate Balance, Overall Accuracy Equality).
-
Publication: Co-author of Emotivation in Human-Robot Interaction for Affective Behavioral Adaptation: Springer LNCS vol. 16133, ICSR+AI 2025. [DOI]
- Research Intern at the National Research Council (CNR), where I developed AUDO4RIAS: an open-source Python application that automates the full molecular docking workflow (ligand/receptor preparation, docking, post-docking interaction analysis) for pesticide-bee receptor risk-assessment studies, integrating MGLTools, AutoDock Vina, and GNINA.
- Programming Tutor (2023–2025) at the University of Naples "Parthenope" (Nola Branch), supporting 30+ BSc students per semester in Computer Engineering and Cybersecurity on C/C++ programming, algorithms, and data structures.
Languages:
Machine Learning & Data Science:
Tools & Environment:
- LinkedIn: Alfredo Mungari
- Email: a.mungari@gmail.com
- ResearchGate: Alfredo Mungari