Periodic Reporting for period 1 - AI-BOOST (Artificial intelligence for better opportunities and scientific progress towards trustworthy and human-centric digital environment)
Periodo di rendicontazione: 2023-09-01 al 2024-08-31
Artificial Intelligence and Robotics hold tremendous potential for enhancing living standards, ensuring safer mobility, improving healthcare, creating new jobs, and personalising public services. However, challenges such as trust and reliability hinder their widespread adoption. Despite the EU's strong AI research community, it faces a smaller talent pool and less private investment compared to the U.S. and Asia.
To enhance Europe’s competitiveness, there is an urgent need for a participatory methodology that brings together industry, academia, and society. Open innovation competitions have shown success in driving innovation, attracting talent, and securing funding, yet the European landscape suffers from fragmentation, isolated initiatives, unsustainable funding models, and inadequate impact measurement.
The overall objective of the project is to create and run a highly replicable AI open innovation competition that attracts the best teams all over EU and Associated Countries to drive scientific progress in the major AI Areas. The project will foster collaboration between the key stakeholders in AI community to define attractive AI challenges with the potential to lead trustworthy and human-centric real-world solutions
To enhance Europe’s competitiveness, there is an urgent need for a participatory methodology that brings together industry, academia, and society. Open innovation competitions have shown success in driving innovation, attracting talent, and securing funding, yet the European landscape suffers from fragmentation, isolated initiatives, unsustainable funding models, and inadequate impact measurement.
The overall objective of the project is to create and run a highly replicable AI open innovation competition that attracts the best teams all over EU and Associated Countries to drive scientific progress in the major AI Areas. The project will foster collaboration between the key stakeholders in AI community to define attractive AI challenges with the potential to lead trustworthy and human-centric real-world solutions
During the PR1, the consortium achieved significant progress towards the project’s Specific Objectives (SOs):
In pursuit of SO1 and SO2, the consortium achieved a significant milestone by completing the 1st and largest of the 7 planned challenges, the Large AI Grand Challenge. This generative AI competition utilised half of the project’s total prize budget and required extensive time and resources, yet it laid a solid foundation for the subsequent AI Open Innovation Challenges. A comprehensive methodology was created to structure impactful AI challenges, and preliminary planning has begun for the next 3 Open AI Challenges, supported by vital networks to ensure smooth implementation.
The Large AI Grand Challenge, aimed at accelerating the development of Large Language Models (LLMs) was a flagship initiative, reinforcing Europe’s position in advanced AI. Through this initiative, we encouraged pioneering startups and SMEs to propose ambitious strategies and commit to innovations that bolster Europe’s competitive position in the AI landscape. Support from key entities, including EC and EUROHPC, has been instrumental. The planning for subsequent Open AI Challenges is now underway and will require support from further organizations.
For SO3, while primary tasks are planned for future phases, necessary datasets and infrastructure access are being prepared to ensure equal opportunities for all participants, including those from under-resourced organisations. This groundwork aims to create an inclusive research environment, vital for meaningful AI research during future competition stages.
SO4 has seen substantial progress through Large AI Grand Challenge participation, which attracted 280 registrants, with 94 completing applications by the deadline. The applicant pool comprised diverse, high-calibre startups and SMEs, of which 4 winners were selected, sharing a prize pool of €1 million and receiving access to 2 of Europe’s top supercomputing facilities, LUMI and LEONARDO, for a combined total of 8 million hours. To maximise outreach and engagement, targeted communications and marketing strategies were developed and implemented as outlined in the GA.
In line with SO5, the Large AI Grand Challenge competition was supported by EC and EUROHPC, with the latter providing an additional prize of 8 million GPU hours. These resources further incentivised participation and enhanced computational access for high-impact projects. PR1 also saw the initiation of discussions to define sponsorship packages.
Progress on SO6 involved reaching out to key sectoral organisations, capitalising on the consortium’s established network. Connections were made with selected projects under CL4-2023-HUMAN-01-01 and CL4-2023-HUMAN-01-03 topics, launched in early 2024. These initial contacts have set a solid foundation for ongoing collaboration with EU-supported initiatives, and we look forward to building on them in the PR2.
Finally, initial efforts under SO7 resulted in the drafting of a preliminary exploitation plan, focusing on strategic steps for post-EU funding sustainability.
In pursuit of SO1 and SO2, the consortium achieved a significant milestone by completing the 1st and largest of the 7 planned challenges, the Large AI Grand Challenge. This generative AI competition utilised half of the project’s total prize budget and required extensive time and resources, yet it laid a solid foundation for the subsequent AI Open Innovation Challenges. A comprehensive methodology was created to structure impactful AI challenges, and preliminary planning has begun for the next 3 Open AI Challenges, supported by vital networks to ensure smooth implementation.
The Large AI Grand Challenge, aimed at accelerating the development of Large Language Models (LLMs) was a flagship initiative, reinforcing Europe’s position in advanced AI. Through this initiative, we encouraged pioneering startups and SMEs to propose ambitious strategies and commit to innovations that bolster Europe’s competitive position in the AI landscape. Support from key entities, including EC and EUROHPC, has been instrumental. The planning for subsequent Open AI Challenges is now underway and will require support from further organizations.
For SO3, while primary tasks are planned for future phases, necessary datasets and infrastructure access are being prepared to ensure equal opportunities for all participants, including those from under-resourced organisations. This groundwork aims to create an inclusive research environment, vital for meaningful AI research during future competition stages.
SO4 has seen substantial progress through Large AI Grand Challenge participation, which attracted 280 registrants, with 94 completing applications by the deadline. The applicant pool comprised diverse, high-calibre startups and SMEs, of which 4 winners were selected, sharing a prize pool of €1 million and receiving access to 2 of Europe’s top supercomputing facilities, LUMI and LEONARDO, for a combined total of 8 million hours. To maximise outreach and engagement, targeted communications and marketing strategies were developed and implemented as outlined in the GA.
In line with SO5, the Large AI Grand Challenge competition was supported by EC and EUROHPC, with the latter providing an additional prize of 8 million GPU hours. These resources further incentivised participation and enhanced computational access for high-impact projects. PR1 also saw the initiation of discussions to define sponsorship packages.
Progress on SO6 involved reaching out to key sectoral organisations, capitalising on the consortium’s established network. Connections were made with selected projects under CL4-2023-HUMAN-01-01 and CL4-2023-HUMAN-01-03 topics, launched in early 2024. These initial contacts have set a solid foundation for ongoing collaboration with EU-supported initiatives, and we look forward to building on them in the PR2.
Finally, initial efforts under SO7 resulted in the drafting of a preliminary exploitation plan, focusing on strategic steps for post-EU funding sustainability.
It is crucial to emphasise that this project does not focus on research and development. Therefore, it is not aiming at developing solutions or technologies that would result in a technological leap. Instead, the project intends to organise AI Open Innovation competitions and award prizes to organisations that effectively tackle specific technological challenges defined in each competition, in line with established standards and criteria. In other words, it seeks to achieve indirect outcomes that could represent significant advancements in the state of the art. A prime example of this initiative is the Large AI Grand Challenge, which aims to catalyse substantial progress in AI across Europe. Currently, we are supporting four start-ups by providing them with valuable resources and opportunities to develop large language models that aspire to outperform existing state-of-the-art systems in various relevant tasks.