Presenter: Enrico Palumbo
Monday, January 15th, 2024 5:30 PM
Location: Sala Piccola Covivio, Corso Ferrucci 112
Allowing users to discover new entities such as books, music, and movies is an important goal for online platforms. Exploratory queries such as “new indie rock releases” or “acoustic guitar covers to learn” can help users explore the catalog and find their next favorite song, and allow platforms to better leverage their catalog by surfacing more niche content. However, the popularity of an entity has a strong influence on which entities are presented by the search system, especially if a machine-learned model for retrieval was trained on user interactions. Also, if users are not aware that they can find new content through exploratory queries, they will rely mostly on navigational queries such as the title of a familiar song or of a famous podcast show.
In this talk, we show how recent advancements in the field of language modeling and graph learning can be used to reduce bias in search systems and provide exploratory query suggestions, helping users discover new content.
Enrico Palumbo is an applied AI researcher, currently working at Spotify as a senior research scientist. His area of expertise is the application of Language Technologies to Search and Recommendations, with a recent focus on generative models. He has contributed to the research and development of AI products and features currently in use by millions of people across the world (e.g. Spotify Search, Amazon Alexa). He is the author of more than 20 publications in top-tier AI conferences and journals. He holds a PhD in Computer Engineering, he was supervised by Links Foundation, Eurecom and Politecnico di Torino.