The laboratory studies Artificial Intelligence applied to new technologies and solutions to develop systems that support prediction and decision-making processes.
Applied Artificial Intelligence
Research Unit
Common Artificial Intelligence
The research unit develops new methods and approaches for Artificial Intelligence applications that are already available. The aim is to create fully customisable ‘proprietary’ models that have a high degree of complexity (in terms of parameters) and that can be trained by exploiting the computational power of the davinci-1 supercomputer.
Researchers: Federico Betti, Fabio Valerio Ferrari, Matteo Guidi
Trustworthy Artificial Intelligence
The research unit develops the scientific and technological background needed to integrate and certify Artificial Intelligence into Leonardo’s products and services.
Researchers: Roberto Morelli,Rosario Di Carlo
Continual, Autonomous and Special Learning
The research unit develops methods and approaches for the continuous development of predictive models that allow networks to adapt to different operating contexts, to train autonomously, including with only limited data or if there are restrictions imposed by compliance with GDPR (General Data Protection Regulation).
Researchers: Bruce McIntosh, Matteo Guidi
Efficient (Deep) Neural Network
The research unit develops methods and approaches to reduce the complexity of networks and models, to transfer them onto small-scale devices or those with limited power. The research is based on models and infrastructure hardware dedicated to local data calculations.