Giacomo Balloccu

I'm a research scientist at Meta in London, where I work on inventing and building AI technologies for Online Advertisment and Targeting.

I interned two times as Applied Scientist at Amazon in Edinburgh, where I worked on Contextual Targeting and Recommendation for Online Advertisment. I did my PhD at University of Cagliari, where I was advised by Prof. Ludovico Boratto, Prof. Mirko Marras, under the research group of Prof. Gianni Fenu.

Email  /  CV  /  Bio  /  Scholar  /  Linkedin  /  Github

profile photo

Research

I'm interested in recommender systems, deep learning, learning representations and generative AI. Most of my PhD research was about creating methodologies for generating explainable paths in knowledge aware recommender systems including the use of Reinforcement Learning and Trasformer based methodologies. During my internships I mostly worked with LLM for representation learning and recommendation.

Supervisor, Mentorship and Dissemination

Supporting and developing the Sardinian ecosystem has always been central to my personal mission. Born in Nuxis, a small village, I navigated my path largely independently, driven by a natural curiosity and determination. However, I believe in fostering change by reducing friction and raising awareness to improve life satisfaction and unlock talent. Too often, potential goes untapped due to the limited opportunities in under-resourced areas like mine.

Over the years, I have actively pursued this mission by serving as a Teaching Assistant for Algorithms and Data Structures at Applied Computer Science and Data Analytics (Unica) for the past three years, teaching at the Pula Summer School (2023/2024), providing individual mentorship, and organizing Breaking Barriers in Tech Sardinia (BBT) a program of multiple events that does dissemination on scientific research and careers in tech. This year, I plan to reinvent BBT Sardinia and continue supporting mentees through dedicated mentorship and supervision.

Due to my professional commitments, I can supervise only one thesis student at a time. If you are interested in conducting research on Recommender Systems and Deep Learning, stay tuned for more details. To apply, please email the following materials with the subject line, "[Research Thesis Application - YOUR LAST NAME]."

  • CV.
  • A motivation letter explaining your research interests and why you are interested in working with me.

News

  • In November 2024 I joined Meta as Research Scientist.
  • In October 2024 I submitted my Ph.D. thesis "Explainable Decision-Making in Recommender Systems" to review!
  • Our paper "EDGE: A Conversational Interface driven by Large Language Models for Educational Knowledge Graphs Exploration" has been accepted at CIKM 2024.
  • Our paper "KGGLM: A Generative Language Model for Generalizable Knowledge Graph Representation Learning in Recommendation" has been accepted at RecSys 2024.
  • We hosted UMAP2024 in Cagliari!
  • Our paper "Learner-centered Ontology for Explainable Educational Recommendation" has been accepted at UMAP 2024.
  • In May 2024 I joined Amazon as Applied Scientist Intern for 3 months.
  • Our tutorial "Explainable Recommender Systems with Knowledge Graphs and Language Models" has been accepted and presented at ECIR 2024.
  • Our paper ‘‘Faithful Path Language Modelling for Explainable Recommendation over Knowledge Graph’’ has been made available on Arxiv.
  • In November 2022 I joined Amazon as Applied Scientist Intern for 4 months.
  • Our reproducibility paper ‘‘Knowledge is power, understanding is impact: Utility and beyond goals, explanation quality, and fairness in path reasoning recommendation’’ has been accepted at ECIR 2023
  • Our paper ‘‘Reinforcement recommendation reasoning through knowledge graphs for explanation path quality’’ has been published in the Elsevier Knowledge-Based Systems.
  • Our tutorial "Hands on explainable recommender systems with knowledge graphs" has been accepted and presented at RecSys 2022.
  • Our paper ‘‘Post processing recommender systems with knowledge graphs for recency, popularity, and diversity of explanations’’ has been accepted at SIGIR 2022.