Evaluating proposals in response to the EIC Accelerator programme.The EIC Accelerator supports companies (principally SMEs, including start-ups) to scale up high impact innovations with the potential to create new markets or disrupt existing ones. The EIC Accelerator provides a unique combination of funding from EUR 0.5 to EUR 17.5 million and Business Acceleration Services (see Section V of the EIC Work programme 2024).The EIC Accelerator focuses in particular on innovations building on scientific discovery or technological breakthroughs (‘deep tech’) and where significant funding is needed over a long timeframe before returns can be generated (‘patient capital’). Such innovations often struggle to attract financing because the risks and time period involved are too high. Funding and support from the EIC Accelerator is designed to enable such innovators to attract the full investment amounts needed for scale up in a shorter timeframe.The EIC Accelerator supports the later stages of technology development as well as scale up. The technology component of your innovation must therefore have been tested and validated in a laboratory and other relevant environment (e.g.,at least Technology Readiness Level 5). The EIC Accelerator looks to support companies where the EIC support will act as a catalyst to crowd in other investors necessary for the scale up of the innovation.

Research Scientist in the HRI team of Pal Robotics. With a Marie Curie fellowship cofounded by the H2020 and Acciò, I'm working on the PRO-CARED project that aims at investigating proactive robot behaviours and mental models in HRI.

As part of my Marie Curie Fellowship, I spend my first year at the Projects of Intelligent Robotics and Advanced Cognitive Systems (PRISCA) Laboratory of the University of Naples Federico II under the supervision of Prof. Silvia Rossi investigating proactive robot behaviour and mental models in hri.

TITLE: PERSONALISING ROBOT ASSISTANCE FOR COGNITIVE TRAINING THERAPYThe overall goal of this work was to demonstrate how a Socially Assistive Robot (SAR) can be employed for providing in situ multi-modal personalised assistance to Persons with Dementia (PwDs). The implications of this thesis are both scientific and societal. Regarding the scientific ones, we propose a social robot that can be personalised by therapists and that can provide in situ tailored assistance to the individual’s unique needs, addressing the current research gaps found in the state of the art: designing (i) an autonomous robotic system that can (ii) learn and adapt to the users’ preferences from a few interactions, during (iii) cognitive training therapies. Concerning social implications, the deployment of such robots would help to bridge the current societal gap between a growing ageing population in need of more healthcare support and a shortage of caregivers. By providing therapists with such a powerful tool, we aim at increasing their effectiveness, allowing them to delegate repetitive tasks, such as cognitive training exercises, to the machines and, therefore, to concentrate on other more relevant activities.

Worked on developing a set of machine learning algorithms aiming to go beyond the current limitations. The algorithms had been used for gaze tracking technology with RGB cameras as well as for image and video compression.The platform allowed for the quick creation of models to capture and follow specific sparse changes of information in the frequency domain.Outcome:- Designed, developed and deployed new machine learning algorithms. - Applied and validated the developed algorithms to real-world scenarios.- Gained knowledge in frequency-domain algorithms.- Worked in a dynamic and fast-paced team with top experts.

Worked on the first gaze tracking solution for an F1 simulator at Ferrari. Furthermore, developed a silhouette detection pipeline for a kart simulator for a Ferrari Driver Academy pilot. Outcome:- Gained and strengthened knowledge in using OpenCV library as well as Matlab Simulink. - Developed face and eye detection algorithms based on Viola-Jones method as well as eye shapes based on Active Shape Model. - Prototyped a 3D gaze tracking solution using stereo vision to estimate the position of eyes on the screen.- Worked in close collaboration with the programming and electronics leads.

Analyzed and evaluated the opportunity to introduce artificial intelligence into the automotive market.Outcome:- Offered strategic support on the next generation of smart sensors.- Produced reports, with a strong analysis of machine learning and data mining state of the art algorithms.