The project will showcase the AI4EU platform with pilot experiments, in order to:
- demonstrate the exploitability of the platform for industrial applications of AI;
- foster the adoption of the platform by industrial champions, and more generally by economic sectors in different areas;
- solve technical challenges raised by advanced industrial applications;
- stimulate a vibrant partnership between Research and Industry to solve AI application problems;
The selected methodology is to include industry-strength pilot experiments with the AI4EU platform on eight distinct application domains, led by industrial partners, as shown here:
 As suggested in the Coordinated Plan on the development of Artificial Intelligence in Europe - Adopted on 15/05/2019 Reference: NT/877-EESC-2018-05386-00-00-AC-TRA
AI4Citizen will provide an AI-powered Personal Assistant for Public Services in the domain of Education. The goal is to demonstrate how AI can support all actors involved in the match-making of secondary school students required to perform an internship, with companies proposing topics for these internships. AI and the AI4EU platform will be used in the pilot experiment to enable teachers and mentors to better understand the competencies brought by external activities and to provide personalised guidance to students to choose the best suited activity.
AI4Robotics will set up a physical installation to evaluate the benefice of the AI4EU platform for predictive maintenance of mechanical pieces of industrial robots in case of premature wear. Models of predictive maintenance will be proposed to evaluate the wear of a robot arm wrist based on data from 3D accelerometers and the wear of paint pumps when they are embedded into the robot arm.
By using an experimental production facility, AI4Industry will provide scenarios for global production and manufacturing systems. These scenarios will integrate methods from symbolic AI and Machine Learning to aid and enhance manufacturing systems and make them more autonomous. Showcases on this experimental production facility will demonstrate the benefice of the AI4EU platform for improving the overall production performance.
AI4Healthcare will have the goal to show how the AI4EU platform can be used to improve the quality constancy of medical images reports. By displaying regions of interests mentioned in the report and computing KPIs about the relevance of medical acts, this pilot will show how to help radiologists to improve medical content of their reports.
AI4Media will show how the AI4EU platform can be exploited to perform 3D-generation of animated videos. As big brands have to shoot and produce locally, one goal is to shoot in one language and seamlessly repurpose to many countries by using audio dub and transfer voice actor expression on the initial video. This raises technical challenges including accurate 3D tracking of faces, audio visual-based mouth re-enactment, realistic re-render (e.g. mouth filling).
Crop quality assessment through computer vision will showcase the AI4EU platform in the context of agriculture. One goal is to help producers to reduce production costs and to obtain the maximum yield possible. In the case of vineyards, automatic counting of fruit in a plant or in the ground, evaluating their maturity and quality will allow decision-makers to select optimal harvesting time and to identify the appropriate amount of resources for recollection.
This pilot will explore the use of the AI4EU platform on air quality data captured by IoT devices, in order to improve data quality and services by combining pollution data with other information such as mobility patterns, weather forecasts, and environmental data.
AI4Cybersecurity will show how the AI4EU platform can be used to detect and prevent of false data injection attacks against parts of a sub-way signaling system meant for authorizing the most performant traffic density in a sub-way line.