Paper 'HiERO: understanding the hierarchy of human behavior enhances reasoning on egocentric videos' accepted at ICCV 2025 and will be presented at the conference. Another paper 'Hier-EgoPack: Hierarchical Egocentric Video Understanding with Diverse Task Perspectives' has also been published.
Research Experience
Currently conducting research at the VANDAL lab at Politecnico di Torino, focusing on artificial intelligence and video understanding. In 2025, he had a research visit at the LTS4 lab at EPFL.
Education
PhD student at Politecnico di Torino, supervised by Prof. Giuseppe Averta. In summer 2025, he visited the LTS4 lab at EPFL (Lausanne, Switzerland), under the supervision of Prof. Pascal Frossard. He is also part of the ELLIS PhD & Postdoc Program.
Background
Research interests include learning artificial intelligence by watching videos of human activities, building models that can understand the hierarchical and compositional nature of human activities. During his PhD, he is exploring how to build video understanding models that can capture these hierarchies and learn abstract concepts from video that are interpretable and reusable.