Testbeds and competence to boost AI development in health

Digital health at RISE is one of 52 partners collaborating in a large international EU project - TEF Health.  The project is coordinated from the Medical University of Berlin, Charité, and the Swedish node is led by Karolinska Institutet in collaboration with RISE, Karolinska University Hospital and SciLifeLab.

"There is a clear goal to develop Artificial Intelligence (AI) in the EU, especially in health and healthcare. We will do this by supporting companies and organizations," says Alireza Salehi, senior researcher at Digital Health and the Centre for Applied AI at RISE.

The goal is high, by 2027 the EU wants 75% of all companies to have some form of AI solution. For this to happen, it is absolutely necessary to have a collaboration between several actors with expertise in digital systems and health. A clear organization is also needed, which is why the project is divided into ten work areas called work packages, one of which is led by Sweden. This area focuses on "Use Cases and Demonstrators" by using X-ray analysis (image analysis), numerical data (biological lab data in omics) and AI. RISE, Sweden's largest research institute, is responsible for the AI part. At the same time, RISE will collaborate with other EU partners to ensure high quality AI solutions when certifying AI systems in health. Test beds are very important for innovation opportunities in the EU. The TEF project has access to many test beds in the EU and it is particularly important that AI in healthcare is developed together with as many partners as possible. The aim is a good interaction between the technical solutions and the complex regulations where data can be utilized through AI. One of the biggest challenges is standardization where AI solutions are dependent on many different parts such as data collection, modeling or validation. Therefore, the Sweden node (RISE and KI, KS and SciLifeLab) for AI and health plays an important role by using data from "Use Cases and Demonstrators".

"Without the development of AI, processes would be slower, cost more and companies will not be able to keep up with societal developments now that data can be produced much more efficiently than before. "Artificial intelligence is also an important support for decision-makers," Alireza explains.

How will this be done?

In addition to small and medium-sizedbusinesses, researchers, hospitals, clinics, health centres and organisations will be able to use the project's resources. RISE provides expertise in data modelling and AI, and owns around 30 test beds. It also has expertise in quality, certification, business models, ethics and regulations.  SciLifeLab, responsible for advanced instruments applied with ten different platforms and expertise in the analysis of biological samples at the atomic level, known as molecular bioscience. Karolinska Institutet and the University Hospital provide medical expertise, patient data and the Center for Imaging Research (CIR) for image analysis. Together, all parties become a strong engine in the development of new AI solutions. Planning and capturing customer needs is also key. By guiding companies, they will have access to the right test bed and expertise based on their needs. One example is to use Cyber Range, a test bed run by RISE, which can build a virtual environment that is fully controlled, and offer tools for creating stable and secure IT systems in combination with AI solutions for health and healthcare.

Not a solo voyage

The AI journey is not a solo voyage. Many platforms and disciplines in biology, engineering and regulation need to communicate and work together.

"It's a challenge, but we are very good at teamwork and really respect each other as partners," Alireza points out.

It is also important to take small steps forward, test, evaluate and perhaps certify before implementing new AI solutions. Ultimately, you want AI to add value, at the individual level and for the benefit of society.

"One example is that today we often receive the same dose of medicine even though we do not need the same medication.  AI could calculate and personalize the medication. Reduced use of medicine would in turn benefit the environment."

AI can provide happier endings

AI is based on mathematical models, known as algorithms. The models can infer, predict and favor a better end to different scenarios, such as diseases or societal phenomena.

"AI is not magic, it is a calculation on data that once happened. If we can use it to model what is going to happen in five years, the better the ending. Some examples are predicting prostate cancer and breast cancer, or using data from a pandemic to get a happier ending to a virus spread in society," says Alireza.

Enablers of the project

Without funding from the EU and Vinnova, the project could not be carried out, €10 million in funding.  In addition, the knowledge-intensive exchange between the participating countries is extremely important. In addition to Sweden, Portugal, Italy, Spain, Belgium, France, Germany, Finland and Slovenia are participating in the project.

Do you want help from TEF Health?

TEF Health services are free of charge for small and medium-sized companies developing high-quality, ethical and legal digital solutions for health. The project provides access to expertise, test beds and experimental facilities for AI and robotics in health (TEF). Welcome to contact RISE!

Read more about the offer on the project page:TEF Health - supporting businesses to use AI solutions in health | RISE
Read more about RISE testbed for cyber security: Cyber range

RISE explains

• TEF

Acronym for Testing Experimentation Facility for Health AI and Robotics, i.e. testing and experimentation facilities for AI and robotics in health.

• Omics
  Branches of science in biological disciplines, such as genetics and cell mechanics.  A concrete example is lab data from biological samples, such as blood.