Hi there!đź‘‹
About🤌Â
I am currently a third-year PhD student in Applied Mathematics at Inria and UniversitĂ© CĂ´te d'Azur, as a member of the Maasai project-team, located in Sophia Antipolis, and the J. A. DieudonnĂ© laboratory in Nice, France. My research focuses on the interpretability of machine learning models and algorithms, both by analyzing their theoretical foundations and by proposing new solutions. I am supervised by Damien Garreau and FrĂ©dĂ©ric Precioso. Previously, I got an MSc in Mathematical Engineering and a BSc in Applied Mathematics, both from Politecnico di Torino.Â
Education
Ph.D. in Applied Mathematics, 2024 (expected)
Inria & UniversitĂ© CĂ´te d'AzurÂM.Sc. in Mathematical Engineering, 2021
Politecnico di TorinoB.Sc. in Applied Mathematics, 2019
Politecnico di Torino
Research
You can find my code on Github and my publications on Google Scholar.
G. Lopardo, F. Precioso, D. Garreau, Attention Meets Post-hoc Interpretability: A Mathematical Perspective [preprint] [code]
G. Lopardo, F. Precioso, D. Garreau, Faithful and Robust Local Interpretability for Textual Predictions [preprint] [code]
G. Lopardo, F. Precioso, D. Garreau, Understanding Post-hoc Explainers: The Case of Anchors, 54es Journées de Statistique 2023 [paper][code]
G. Lopardo, F. Precioso, D. Garreau, A Sea of Words: An In-Depth Analysis of Anchors for Text Data, AISTATS 2023 [paper][code]
G. Lopardo, D. Garreau, Comparing Feature Importance and Rule Extraction for Interpretability on Text Data, ICPR 2nd Workshop on Explainable and Ethical AI, 2022 [paper][code]
G. Lopardo, D. Garreau, F. Precioso, G. Ottosson, SMACE: A New Method for the Interpretability of Composite Decision Systems, ECML 2022 [paper][code]
Teaching
During my PhD, I also teach undergraduate courses in science and economics, as well as master's level courses within the Mathematical Engineering master program of UniversitĂ© CĂ´te d'Azur.Â
2023-2024 (64 hours)
Mathematical Statistics (MSc, 1st year, Mathematical Engineering, 36 hours)
Statistics 2 (BSc, 2st year, Economics, 21 hours)
Probability and introduction to statistics (BSc, 2st year, 7 hours)
2022-2023 (64 hours)
Mathematical methods, classroom exercises and R laboratory (BSc, 2nd yer, 48 hours)
Introduction to Mathematics (BSc, 1st year, 16 hours)
2021-2022 (40 hours)
Mathematical Statistics (MSc, 1st year, Mathematical Engineering, 16 hours)
Fundamentals of mathematics (BSc, 1st year, 24 hours)
News
February 2024: new preprint! We investigate the relation between attention-based and post-hoc explanations Â
November 30, December 1: I've been at the 2nd Nice Workshop on Interpretability
November 24: talk to the Maasai seminarÂ
November: new preprintÂ
July 2-7: I presented my work at Journées de Statistique in Bruxelles
June: I served as PC member to the KGML workshop @ECML 2023
April 25-27: I have been in Valencia for AISTATSÂ Â
January 31: I presented SMACE to the AI4media network in Florence
January 2023: our analysis of Anchors for text data got accepted to AISTATS 2023
November 17, 2022: talk at the 1st Nice Workshop on Interpretability
September 21, 2022: talk at ECML in Grenoble, presenting SMACEÂ
August 21, 2022: talk at the 2-nd Workshop on Explainable and Ethical AI – ICPR 2022 in Montreal
June 2022: SMACE got accepted @ECML, my first conference paper!
April 2022: I attended the Statlearn spring-school
November 2021: I attended the AI & Companies WeekÂ
November 19, 2021: talk at the SophIA SummitÂ
November 2021: new preprint available: we propose a new method for the explainability of composite AI systems
October 1, 2021: I started my PhD
Contact
Email: gianluigi.lopardo@inria.fr
Twitter: https://twitter.com/gigilopardo
