In 2019, Marco Minici earned his Master’s Degree in Data Science from Sapienza University of Rome, presenting a thesis titled “The Effect of People Recommenders on Opinion Dynamics.”

His research revolves around the topic of Reliable and Secure Social Media. He is actively exploring two main research directions: (1) creating agent-based models to understand subtle, long-term algorithmic effects on users, and (2) using probabilistic machine learning methods to uncover latent features shaping online discussions. He also devotes part of his time to industrial research working on automatically tagging digital products with descriptive keywords using Multimodal ML and Extreme Multi-Label classification frameworks. During a six-month internship as an Applied Scientist with the Amazon Music Personalization team in Berlin, Marco initiated a research line addressing Catastrophic Forgetting in the ranking algorithm. This involved scoping the problem, implementing an auditing pipeline, offering insights, and proposing as well as measuring the impact of mitigation strategies.

List of collaborators (not exhaustive):

• Giuseppe Manco, ICAR-CNR.
• Francesco Bonchi, CENTAI and EURECAT.
• Federico Cinus, CENTAI.
• Massimo Guarascio, ICAR-CNR.
• Francesco Fabbri, Spotify.
• Corrado Monti, CENTAI.