Vieri Giuliano Santucci

Main research topics (and sub-topics) with a brief description and some selected publications

1. Autonomous and developmental robotics. My main research activity focuses on the development of artificial agents capable of autonomously exploring the environment and interacting with it, acquiring new knowledge and competences without any specific assigment from human designers.

1.1 Motivational systems for autonomous artificial agents: Intrinsically motivated open-ended learning This specific line of research focuses on the analysis, development and implementation of motivational signals suited for guiding artificial agents in autonomously exploring their environment and learn new behaviours in an open-ended fashon. In particular, the research focus on the study and implementation of the biologically-inspired paradigm of intrinsic motivations, i.e. self-generated motivational signals that push agents in acquiring new knowledge just "for the sake of curiosity", through the exploration of novel states, the discovery and selection of "goals" and the learning of skills.

1. Santucci, V. G., Oudeyer, P. Y., Barto, A., & Baldassarre, G. (2020). Intrinsically motivated open-ended learning in autonomous robots. Frontiers in neurorobotics, 13, 115
2. Santucci, V. G., Baldassarre, G., & Mirolli, M. (2016). Grail: a goal-discovering robotic architecture for intrinsically-motivated learning. IEEE Transactions on Cognitive and Developmental Systems, 8(3), 214-231.
3. Santucci, V. G., Baldassarre, G., & Mirolli, M. (2013). Which is the best intrinsic motivation signal for learning multiple skills?. Frontiers in neurorobotics, 7, 22.

1.2 Cognitive Robotic Architectures

To support autonomous learning processes, artificial agents artificial agents must be endowed with an architecture that ensures the integration of mechanisms and functions for managing motivations, storing new information, and learning new behaviours. This research therefore concerns the development and implementation of such architectures, with particular attention to the relationship between their various components and their balancing.

1. Santucci, V.G., Montella, D. & Baldassarre, G. (2023). C-GRAIL: Autonomous Reinforcement Learning of Multiple and Context-Dependent Goals. IEEE Transactions on Cognitive and Developmental Systems, vol. 15, no. 1, pp. 210-222.
2. Santucci, V. G., Baldassarre, G., & Cartoni, E. (2019, August). Autonomous reinforcement learning of multiple interrelated tasks. In 2019 Joint IEEE 9th international conference on development and learning and epigenetic robotics (ICDL-EpiRob) (pp. 221-227). IEEE.
3. Santucci, V. G., Baldassarre, G., & Mirolli, M. (2016). Grail: a goal-discovering robotic architecture for intrinsically-motivated learning. IEEE Transactions on Cognitive and Developmental Systems, 8(3), 214-231.