Skip to content

Gianluca Baldassarre

  • Istituto di Scienze e Tecnologie della Cognizione Consiglio Nazionale delle Ricerche
  • 8Followers
  • 1Following

About

I coordinate the Research Group called “Laboratory of Computational Embodied Neuroscience”. Learning and embodiment are two central pillars of our research. We use computational models, tested in simulated organisms bodies when possible, to understand how brain learns and produces behaviour, in particular goal-directed and habitual behaviour. Our models are based on leaky- or spiking-neuron neural networks, and are constrained through empirical data on the anatomy of brain (especially macro- and meso-anatomy), and its physiology and plasticity. Our models capture the system-level organisation of brain to be able to address the target behaviours as studied in empirical experiments. In synergy with this, our group also works within Developmental Robotics with the aim to understand how children and robots can undergo open-ended learning: this leads to the autonomous acquisition of increasingly complex sensorimotor skills based on intrinsic motivations (curiosity, novelty, surprise, competence improvement, etc.) and goal self-generation. To this purpose we build neural-network architectures formed by various components learning with unsupervised, supervised and reinforcement-learning algorithms. We test the architectures with both simulated and real humanoid robots, with a focus on scenarios involving the interaction (e.g., reaching and manipulation) of objects in the environment. We build these architectures either with the technological purpose to produce useful robots or with the aim to investigate the behaviour of children and babies as studied in developmental psychology.

Professional experience

National Research Council Institute of Cognitive Sciences and Technologies

Followers (8)

Explore network

Following (1)

Other IDs