Launched in late summer 2020, Trail (Trusted AI Labs) aims to accelerate the development of artificial intelligence technologies in Wallonia by bringing together the resources of the country's five French-speaking universities and four accredited research centres (Multitel, Cetic, Sirris and Cenaero).
One of the pillars of this initiative is the Trail Institute, a virtual structure led by Professor Benoît Macq of the PiLab at UCLouvain. This institute has a budget of 32 million euros.
50 PhD students, one third of them women
"The Ariac project is the largest artificial intelligence project being financed by the Walloon Region at the moment," says Professor Macq.
"It spans five years and allows the recruitment of 50 PhD students, working on cutting-edge topics related to artificial intelligence. These PhD students, a third of whom are women, are also working with accredited research centres. All this, of course, is to support Walloon companies. The main idea is to enable them to grow their businesses, using artificial intelligence (AI)."
Solution-oriented research
At the Trail Institute, we don't do artificial intelligence for the sake of doing maths. This is research stimulated by very practical questions from the field being developed. There are five priority areas: medicine, media, manufacturing, mobility and energy.
"In these fields, AI is considered a tool for improving processes and making the right decisions," he explained a few days ago in Paris, on the sidelines of Vivatech, the European meeting place for innovations and new technological trends. The scientist, together with several PhD students from the Trail Institute, was invited by the Research & Innovation department of Wallonia-Brussels International.
A tool for personalised medicine
"Let's take the example of medicine," he suggests. "Specialists can make decisions based on their experience. With the help of AI, they can also make much more reliable decisions by analysing the vast amount of data available. AI can help them predict what will happen long term."
"So we're working on artificial intelligence models that allow doctors to make much more personalised decisions for their patients and choose less aggressive options."
"One third of our PhD students do research in this field, which is particularly important in Wallonia. We want each PhD student to spend 15% of their time working on major challenges led by the accredited research centres that work with us on these issues. This takes place in workshops where we create and test solutions with companies."
AI in the sandbox
A 'Factory' has also been set up within the Trail initiative to facilitate these tool tests. It is a place where researchers come to place the software they have developed. There is also open source software developed all over the world.
"The idea of this structure is to have a sandbox, a place for experimentation and meetings where we build solutions together," says Benoît Macq. A sandbox that also includes end users.
"This prevents everyone from developing the same solution separately and allows information to circulate more quickly. For example, by quickly reporting relevant information from a major international conference attended by a Trail member."
Vector of international visibility for Wallonia
"We are developing international contacts within the framework of the Trail Institute," continues the scientist. "This is the case here at Vivatech, with the Paris ecosystem, for example. But also elsewhere. We have very advanced contacts with the Berlin ecosystem."
"We are going to become part of a European Test and experimentation facility, which includes very large hospitals, such as the Charité in Berlin, the hospitals in Lyon and the Karolinska in Sweden. Trail is a great visibility tool for our researchers."
Where will Trail be in five years? "We need to move towards becoming a real institute and step up our involvement in European research programmes. At the moment, we don't have enough of a presence there when it comes to AI."
A tool to support decisions
"My mantra is to work on collective decisions, aided by AI, and continuous learning processes. When it comes to curing or fighting global warming, the solution does not lie with one person or one laboratory. It is collective work. AI is not going to replace the doctor who uses it. But the doctor who uses it will replace the doctor who doesn't use it..."
"To do this, we must invest in collective learning. What is convincing in this context is using digital technology and artificial intelligence, not taking away the responsibility of individuals, but supporting the decision, in a collective process. There needs to be interaction between the AI and the group of humans. But of course, in the end, it is the humans who decide."
"Artificial intelligence is also a real social project. And, in this context, Trail is also important for giving young people the desire to get involved, to believe that technology is not something dark, that it is something positive. It is a tool that helps us grow. It is a tool for growth."
Source: article by Christian Du Brulle for Daily Science.