Curriculum
Computer Science for Societal Challenges and Innovation, XXXIX series
Grant sponsor
FBK
Supervisor

Luciano Serafini, Lamberto Ballan
Co-supervisor
s
Gianluca Campana
Contact
filippo.ziliotto@studenti.unipd.it

Project description
The field of artificial intelligence (AI) has witnessed a transformative shift in recent years, evolving from the era of "internet AI" to the emergent paradigm of "embodied AI." In this new landscape, AI algorithms and agents have departed from the conventional practice of relying solely on internet-derived datasets, comprising images, videos, and textual data. Instead, they have transitioned towards actively acquiring knowledge through direct interactions with their physical surroundings. This transition has gained a substantial growth and innovation, driven by the escalating demand for sophisticated embodied AI simulators.
My research investigates the integration of reinforcement learning agents within indoor environments, with a specific emphasis on incorporating neuro-symbolic reasoning techniques. It also represents a collaborative effort in partnership with the Fondazione Bruno Kessler (FBK) research center. The objective is to advance our understanding of how these agents can effectively operate in real-world indoor scenes. Ultimately translating these findings into practical implementations for the utilization of robotic systems available at FBK.