Porto leads €8 million European project to overcome “Achilles heel” of artificial intelligence
A European consortium coordinated by Porto-based Fraunhofer Portugal AICOS has launched ACHILLES, a project dedicated to transforming the future of artificial intelligence.
The Fraunhofer Portugal AICOS Research Centre (FhP-AICOS) was founded in Porto in 2009, following a partnership between the German Fraunhofer Society, the Foundation for Science and Technology (FCT), and the University of Porto. The Centre currently employs around 100 people - 80 of whom are highly qualified researchers -, possessing a portfolio of clients in areas such as health, agriculture, retail, and energy.
Earlier this month, FhP-AICOS officially launched ACHILLES, deemed by the centre as an "innovative project dedicated to transforming the future of artificial intelligence, making it lighter, more transparent and safer.” Borrowing its name from the Greek warrior Achilles, the project wants to overcome AI's weaknesses, namely efficiency and reliability, with a modern approach.
"As AI becomes more and more integrated into our daily lives, concerns about transparency, ethics and energy consumption are growing. ACHILLES addresses these challenges by introducing innovative approaches and supporting the development of AI systems that are 'lighter' (sustainable and efficient), 'transparent' (easy to understand and document) and 'safe' (robust and compliant with ethical and legal standards),” explained FhP-AICOS.
ACHILLES was launched by a European consortium of 16 organizations from 10 countries, funded with more than eight million euros under the Horizon Europe framework programme and led by FhP-AICOS as the coordinating partner.
"At the heart of ACHILLES is an interactive development cycle inspired by the positive aspects of clinical trials,” says André Carreiro, senior researcher at FhP-AICOS and the project's coordinator. This cycle "comprises four main modules: human-centric methodologies, data-centric operations, model-centric strategies and implementation-centric optimizations. The cycle concludes with a return to human-centered approaches, ensuring that explainability and transparency are incorporated into the system,” explains André Carreiro.
The project coordinator also highlights the project's strategic impact on both technical-scientific advances and policies: "With ACHILLES, we seek to promote advances in AI that are in line with EU ethical standards, legal requirements and a commitment to transparency,” Carreiro adds.
