The Large AI Grand Challenge

Beneficiary & submission

The Large AI Grand Challenge is open from 16 November 2023, 12h CET to 13 of February 2024, 17:00.00 Brussels time. Applicants are invited to register, create a user profile, and complete the full application form, prepare the required supporting documents and submit them via the F6S platform and accessible through AI-BOOST website: https://aiboost-project.eu/large-ai-grand-challenge/. All proposals must be fully completed and submitted via F6S platform before the deadline: Main steps required: Registration via F6S platform Dully complete the application from (all mandatory fields must be completed) Attach the Declaration of Honour dully filled in and signed by the legal representative. Attach the Attachment 2 – Tables and Images Submit the application form before the deadline. See guidelines for applicants: https://aiboost-project.eu/large-ai-grand-challenge/ The evaluation process shall take approximately two (2) months starting from the closing date of the call. After this period, applicants will be informed about the result. Please refer to the guidelines for applicants to gather all information regarding the eligibility criteria, selection criteria, and the application process: https://aiboost-project.eu/large-ai-grand-challenge/

Further information

Large AI Grand Challenge - General context Large-scale AI models refer to a new generation of general-purpose AI systems that can adapt to various domains and tasks without significant modification. Thanks to their adaptability, these models hold immense potential to revolutionise various industries. Notable examples of large-scale AI models include OpenAI’s GPT-4 and Meta’s LLaMA 2. It is of strategic importance for European sovereignty to ensure European companies master this field. Large AI Grand Challenge - Scope The purpose of the Large AI Grand Challenge is to foster the development of large-scale AI models in Europe and to substantially increase the visibility of Europe's activity in this field. The competition will reward innovative start-ups for devising ambitious strategies and making commitments to develop large-scale AI models that will provide a competitive edge for Europe. The expected outcome of the Large AI Grand Challenge is the selection of up to 4 proposals to create innovative foundational language models that will outperform state-of-the-art systems in a number of relevant tasks. The development of these models should necessarily involve the use of High-Performance Computing (HPC). Large AI Grand Challenge – Overview Participants in the Grand Challenge are invited to submit a proposal for the development of a language foundation model, utilising one of the EuroHPC JU targeted facilities (i.e. LUMI or Leonardo supercomputers). The model must be trained from scratch, possess a minimum of 30 billion parameters, and be trained following state-of-the-art optimal scaling laws for computing and training data size. Applicants are required to provide a detailed project scope and plan for the development of the large-scale AI model. They must also offer solid justification for the planned model's relevance, the justification for the use of HPC, and demonstrate their team's expertise in training foundation models using HPC systems, as well as the efficient use of the target supercomputers (i.e. LUMI or Leonardo), along with an appropriate plan of the use of the computing time. An SME can present a proposal targeting the LUMI supercomputer and an independent proposal targeting the Leonardo supercomputer. Only one proposal can be submitted by a SME to each of the specific supercomputer. However, it's important to note that an individual SME is not eligible to receive both prizes. A critical component of the proposal will be the inclusion of a substantial differentiating factor compared to existing large language models. This could be accomplished either by introducing innovative enhancements or by devising novel models that effectively address the limitations of current ones. Additionally, the proposal should provide a rationale for the utilization of HPC facilities, accompanied by a well-defined plan. Large AI Grand Challenge – Specific Conditions Applicants should consider the specific architecture and development framework that best suits their models, and the software requirements should be present in the project proposal. Projects are expected to use the facilities of one of these two EuroHPC facilities, LUMI and Leonardo supercomputers with a maximum amount of computing hours of 2 million of GPU hours per system in total (including pre-processing, paralelization, experimenting, benchmarking and training). A description of data to be used should be also provided, as well as a data management plan. All projects should comply with the GDPR and other applicable privacy protection and non-discrimination rules. General benchmarking will be the criteria to evaluate the success for all projects, the benchmarks are available at the evaluation guidelines. An impact and ethical assessment will be required for the proposals Multimodal models are not expected during this challenge. Available Budget - Monetary Prizes Subject to conditions describer below, up to four (4) prizes may be awarded by the panel of the Large AI Grand Challenge as follows: LUMI Winner: Up to two prizes of Euros 250,000 and an allocation of 2 million GPU hours on the LUMI facility per project. This allocation will be used to develop the large-scale AI model described in the proposals in the 12 months following the prize awards. If there is ultimately only one winner, the total prize will be 500,000 Euros, along with an allocation of 4 million GPU hours on the LUMI facility. Leonardo Winner: Up to two prizes of Euros 250,000 and an allocation of 2 million GPU hours on the Leonardo facility per project. This allocation will be used to develop the large-scale AI model described in the proposals in the 12 months following the prize awards. If there is ultimately only one winner, the total prize will be 500,000 Euros, along with an allocation of 4 million GPU hours on the Leonardo facility. The allocations are granted for a period of one (1) year. Access to EuroHPC systems implies adherence to the rules and conditions set in the EuroHPC Access Policy. The winners will be listed by Team name on the official Challenge website and promoted by AI-BOOST. An individual SME cannot receive both prizes. In case of termination of the Challenge or if the requirements are not met, the prizes may not be awarded upon decision of the Panel.