Advanced Computing in the Age of AI | Monday, July 15, 2024

BSC Enhances Trustworthy Edge AI with 53-Partner European Consortium 

July 2, 2024 -- Researchers from the Computer Architecture – Operating Systems group at the Barcelona Supercomputing Center - Centro Nacional de Supercomputación (BSC-CNS) have joined EdgeAI-Trust, a three year project co-funded under the Chips Joint Undertaking (JU). The project aims to decentralize edge artificial intelligence to help transform industries, such as mobility, manufacturing and agriculture.

The project, which brings together 53 partners from across Europe, aims to decentralize AI to transform industries such as mobility, manufacturing, and agriculture.

Launched on May 1, 2024, EdgeAI-Trust brings together a 53-partner consortium from across Europe, including major suppliers, original equipment manufacturers, and leading research and academic organizations. The project focuses on distributing AI processing across multiple devices to create a trustworthy edge AI ecosystem.

This approach leverages the diverse expertise within the consortium to advance the evolution of AI, emphasizing the important role the semiconductor industry will play in the development and adoption of specialized chips, like AI accelerators, which enhance scalability and reduced latency.

BSC research will provide AI models that meet safety requirements, such as enabling explainability and traceability, and efficient platform mapping. They will leverage key results from EU-funded projects SAFEXPLAIN and REBECCA and will extend them to produce AI models that meet appropriate safety requirements and will provide software solutions to validate compliance. BSC will work with several members of the consortium to define and generate these AI models, efficiently map them onto hardware platforms, offer platform-level safety services and build comprehensive safety cases and concepts.

“EdgeAI-Trust provides BSC with the opportunity to test its AI-related technologies in the context of complex industrial use cases, hence allowing us to reach higher technology readiness level and paving the way towards exploitation of our AI-related assets,” said Jaume Abella, BSC Principal Investigator for EdgeAI-Trust and co-team leader of the Computer Architecture Operating System research group.

EdgeAI-Trust contributes to the implementation of Chips JU strategic plan, specifically integrating trustworthy edge AI technologies into complex heterogeneous components and systems. The project's objectives include:

  • Develop a trustworthy domain-independent AI architecture
  • Create reliable and collaborative large-scale edge AI solutions
  • Increase trustworthiness of edge AI solutions
  • Develop toolchains for optimizing and validating edge AI solutions
  • Large-Scale Impact and Economic Leadership through EdgeAI EDEM Platform

About EdgeAI-Trust

EdgeAI-Trust- Decentralized Edge Intelligence: Advancing Trust, Safety, and Sustainability in Europe is a 36-month EU-funded project that began on May 1, 2024. The project responds to the Chips Joint undertaking call HORIZON-KDT-JU-2023-IA-FOCUS-TOPIC-3. The project is coordinated by German partner TTTechAuto Germany and counts with the participation of 53 partners, including: ZF Friedrichshafen AG, Infineon Technologies AG, Multic Labs GMBB, ASVIN GMBH, Deutsche Zentrum fur Luft und Raumfahrt EV, Universitát Siegen, Hochschule Hamm-Lippstadt, Hochschule Offenburg, TTTEch Auto AG – Austria, TTTech Computertechnik AG, AVL List GMBH, Infineon Technologies Austria AG, Virtual Vehicle Research GMBH , Stmicroelectronics Grenoble 2 SAS – French, Stmicroelectronics Rousset SAS, Zettascale Technology SARL, VRANKEN-POMMERY Prooduction, Compagnie Vranken, Technext, Commisariat A L Energie Atomique et Aus Energies Alternatives, Universite de Reims Champagne-Ardenne, Hawai Tech, StMicroelectronics SRL – Italian, Interactive Fully Electrical Vehicles SRL, AVIMECC SPA, Exida Development SRL, Barcelona Supercomputing Center – Spain, TTTech Auto Spain, Technica Electronics Barcelona, Solver Machine Learning, Rapita Systems SL, Fent Innovative Software Solutions SSL, Clue Technologies SL, Universitat Politecnica de Valencia, Mellanox Technologies Denmark APS, Tegnology Aps, IVEX NV, SIRRIS HET Collectif Centrum Van de Technologische Industrie, Tofas Turk Otomobil Fabrikasi, AVL ARASTIRMA Ve MUHENDISLIK SANAYI VE TICARET LIMITED SIRKETI, Smart Kontrol Sistemleri ve Yazilim Anonim Sirketi, Buyutech Teknoloji Sanayi Ve Ticaret Anonim Sirketi, Charokopeio Panepistimio, Elliniko Mesogeiako, Panepistimio, Othon Tomoutzoglou, Smartsol sia, Velti Platforms and Services Limited, Iotam Internet of Things Applications and Multi Layer Development LTD, Intitut Jozef Stefan, Miromico, Zurcher Hochsule for angewandte Wissenschaften, Uyan Electronic.

Source: BSC-CNS