The European Lighthouse on Secure and Safe AI (ELSA), a recently established AI center of excellence composed of top European researchers, met in Barcelona for the first time for three days from October 10-12. The large and growing network aims to promote the development and deployment of cutting-edge AI solutions in the future and make Europe the global beacon of trustworthy AI. The goal of the meeting of the initial 26 partners was to advance the pressing challenges regarding the safety and security of artificial intelligence and machine learning. ELSA builds on and extends the European Laboratory for Learning and Intelligent Systems (ELLIS) - an internationally recognized AI network of excellence.
The first day of the event served to get acquainted and build solid working relations between the research and industry partners of the consortium, setting the foundations for fruitful collaboration in the years to come. In parallel, the consortium discussed the planned activities over the next three years. As ELSA project coordinator and CISPA Faculty, Prof. Dr. Mario Fritz, explained, "We want to develop trustworthy artificial intelligence. This means that in the future, we want AI to be used in a legally compliant, ethical, and secure and safe way within Europe. The design of these technologies is our responsibility, and so is the development of the positive impact on society."
On the second day, the researchers engaged in a hands-on workshop focused on the definition of the six use cases that will be explored in more detail over the course of the ELSA project. The selected use cases relate to key uses of AI and ML that will serve as a yardstick to measure progress in safe and secure AI techniques and models.
The use cases focus on:
Healthcare is a sensitive area with the highest privacy requirements. The ELSA project focuses on medical data and how AI can be trained on it in a privacy-compliant manner across different institutions.
Autonomous vehicles are safety-critical systems. Not only do they need to perform excellently, but they also need to respond appropriately to the unexpected. This could be, for example, adversarial manipulation, extreme weather conditions, or accidents. The ELSA project aims to develop test environments to evaluate the safety of such methods.
ROBOTICS - LEARNING THROUGH HUMAN INTERACTION
Enabling machines to continuously learn and perform tasks independently, with or even without human assistance, is a goal of robotics. To achieve this goal, the ML models underlying intelligent machines must be trained with large amounts of data. The purpose of the ELSA project is to ensure that data can be used effectively for the learning process without violating the privacy of individuals.
MEDIA ANALYTICS - THE FIGHT AGAINST DISINFORMATION
Thanks to the power of modern techniques, deepfakes are more easily enabled and subsequently disseminated throughout the media. A deepfake is an image or video created using artificial intelligence that appears authentic when it is not. Whereas in years past, AI-generated images contained clues to the fake, today's results are far less recognizable. This use case explores new ways to understand and detect counterfeit data.
CYBERSECURITY - MALWARE DETECTION
End-user device security is a daunting and challenging task. Many anti-malware solutions are powered by machine learning and data-driven AI algorithms to combat malware. Unfortunately, attackers can bypass these protections by attacking the AI itself. The ELSA project aims to overcome this problem by making AI-based malware detection systems more resilient to such attacks or eliminating some altogether.
Automated information extraction from documents is a crucial aspect of AI solutions for enterprises and is used for process automation. Analyzing information from documents leads to AI decision-making processes that can directly impact humans. At the same time, documents usually contain private information, which limits access to them. The ELSA project aims to develop methods for training large-scale ML models on private and widely distributed data that protect human privacy.
More information on the use cases can be found at: https://www.elsa-ai.eu/use_cases.html
On Wednesday, the closing day of the kick-off event, the participants had the opportunity to learn about the research work of the other participants in a poster session. ELSA partners presented scientific posters on artificial intelligence and machine learning. In a final discussion, the results of the last days were summarized, and the vision of ELSA - European Lighthouse on Secure and Safe AI was consolidated.
Prof. Dr. Mario Fritz concluded, "It was a great start for the ELSA project in Barcelona. We are highly motivated to continue building the virtual center of excellence for secure AI in Europe."
For more information on the project, please visit: https://www.elsa-ai.eu