What is EMPAIA?
In early 2019, the Federal Ministry for Economic Affairs and Energy (BMWi) launched the competition “Artificial intelligence as a driver for economically relevant ecosystems” (a.k.a. "AI Innovation Competition") - a specific step towards implementing the Federal Government's strategy on artificial intelligence. Out of 130 submissions from all areas of the economy, 35 projects were selected for funding in a 6-month competition phase. One of the successful proposals was EMPAIA, submitted by a consortium of Charité, Fraunhofer MEVIS and MindPeak GmbH.
In the competition phase from April to October 2019, the EMPAIA concept was further developed, extensive public relations work took place and a congress exhibit was held. Many new partners were added from clinical-diagnostic institutions, professional associations and the private sector. Together with clinical partners, the concept of "reference centers" was established, in which innovative AI products will be tested. Numerous discussions were also held with future private industry partners who will contribute considerable resources to the project. This work formed the basis for the application for funding in the subsequent implementation phase, which was submitted in September 2019. EMPAIA emerged as the winner of the competition phase together with 15 other large collaborative projects and is now receiving funding of around € 11M over three years for the implementation phase. In addition, EMPAIA was able to raise more than € 6.2M in additional funding from private industry partners. The EMPAIA project officially started on January 1, 2020.
The consortium was reorganized for the implementation phase. Participating partners are now: Charité (Institute of Pathology), Fraunhofer Institute for Digital Medicine MEVIS, TU Berlin (DAI-Labor), vitagroup AG and the QuIP (Quality Assurance Initiative Pathology GmbH) .
After EMPAIA got off to a full start on July 1, 2020, the next goal is the minimally viable product (MVP). To this end, several teams are working on goals that are connected in various ways.
The tasks of EMPAIA also include the development of competencies, i.e. the transfer of knowledge, particularly in the fields of pathology (for IT specialists) and IT / AI (for pathologists). The EMPAIA Academy was developed for this purpose - a series of events that first took place in March 2021. Due to the extremely positive responses, the Academy format will be further developed in order to provide the contents to an international audience. Thanks to the Academy, EMPAIA was able to attract many additional interested parties in pathology, computer science and the private sector.
The five committees also started their regular meetings in the first quarter of 2021. The committees work on the following topics:
"Standardization", "Profitability and Billing", "Testing of Algorithms & Systems", "Legal Issues" and "Studies".
The cooperation with the 10 national and three international reference centers is also moving forward quickly. Since 2020 there have been extensive discussions and virtual conferences, as well as assessments of the existing infrastructure and joint project planning. In addition, an intensive cooperation in the field of AI explainability began, which the Austrian sister project "EMPAIA-A" at the Medical University of Graz is focusing on. The newly formed public relations team is dedicated to intensifying the social media activities and news areas.
The platform provides standardized access to data for the development and validation of AI algorithms as well as tools for their creation and use via a repository.
EMPAIA provides a secure marketplace for clinical diagnostics and research support services. The focus of activities is on seamless integration into existing workflows.
In addition to validation, the certification of algorithms for diagnostic use and for the purpose of billing has an important role. The platform supports the orchestration of developers, reference institutes and certifiers to provide quality-assured solutions.
What does EMPAIA offer for ...