Talks @ GAIPS: “Counterfactual Explanations and Algorithmic Recourse: Enhancing Interpretability and Transparency in Machine Learning Models” by Samuele Tonati
On September 26, GAIPS (Group of AI for People and Society) from INESC-ID, will host a talk titled “Counterfactual Explanations and Algorithmic Recourse: Enhancing Interpretability and Transparency in Machine Learning Models” presented by Samuele Tonati, a visiting PhD student at GAIPS from Scuola Normale Superiore Pisa (Italy) who is collaborating with INESC-ID researcher Rui Prada for his PhD research.
Date & Time: September 26, 11h00
Where: Room 1-44, Técnico – Oeiras, TagusPark | Online: here
Abstract: Samuele Tonati’s research focuses on using counterfactual explanations and recommendations to explain black-box machine learning models, emphasising properties like minimality, actionability, and diversity. It aims at assessing the robustness of these explanations to data shifts and examining the relationship between counterfactual distance and user effort. Additionally, his research explores trade-offs in designing counterfactuals and introduces customised metrics to evaluate feature change attainability, tailored to the needs of stakeholders like customers, practitioners, and compliance officers.
Image | © GAIPS