The latest advancements in decentralized learning

AI Sweden’s AI Labs serves as a dynamic research and innovation centre, including both several focus areas and leading infrastructure enabling the rapid development of bold innovations in AI. Designed to facilitate cross-organizational collaboration, AI Labs brings together diverse partners from industry, the public sector, and academia. 

The focus is on building competence and capacity while developing innovative methods and technical solutions to effectively operationalize AI, a crucial step in driving Sweden's progress and enhancing its position in the global innovation landscape.

Recently AI Labs hosted its yearly consortium event, this time on the theme The Latest Advancements in Decentralized Learning. The event is designed to share new learnings with the partner network and allow partners to network with AI Sweden experts, other partners, and students working at AI Sweden. The agenda included both the keynote by Dr. Bryan Reimer as well as presentations of results from the different labs and projects, and a poster session and lab tour to facilitate networking and discussion:

• AI Sweden general presentation
• Keynote by Bryan Reimer: Picking The Right Screwdriver: Perspectives on The Future of Human Centered AI
• AI Labs - Advances in Decentralized Learning
• AI Security - Privacy Preserving 
• AI in Space
• Mobility: AI For Autonomous Vehicles
• Next generation infrastructure for foundation models 
• Federated Learning in Health Care 
• Talent Programs

In his keynote, Dr. Bryan Reimer, who is also a member of AI Lab’s Advisory Board, started with a foundational observation: “Moment by moment, the capabilities of AI are being reinvented.” 

This, combined with the fact that “when it comes to AI researchers, technology developers, politicians, investors, and consumers see things through different lenses” leads to challenges when striving to adopt AI in an organization. Not only is AI an ever changing technology, but we are also struggling to find a common ground for how to communicate and think about the benefits and risks of using AI in our society.

Dr. Bryan Reimers suggests that we focus on an important balance between risks and rewards. To get there, we need to think about the trust in and transparency of AI while aiming for solutions that augment humans rather than replace them.

“Consider prioritizing AI investments that aim to support or augment instead of replace human intelligence.” For instance, considering less emphasis on, among other things, new methods, building over-optimized systems, and autonomous decision-making. Instead think more about how to use existing methods, building robust systems, support decision-making, and human centered solutions.

One way to put that mindset into practice is to start working on a broader range of challenges and using AI as a tool to solve them. This is the core idea behind AI Sweden’s talent programs, which was last on the agenda.

Dr. José-Marie Griffiths, President of Dakota State University, DSU, and a member of AI Labs’ advisory board, exemplified what that can look like with the Industrial Immersion Exchange Program, where AI Sweden together with DSU offer the opportunity to gain international and industrial experience in AI and cybersecurity as part of a fully covered summer and academic credit-bearing internship in Sweden and the U.S.

“This summer we have doubled the size of the program, and we will accept 24 participants,” said Dr. José-Marie Griffiths. 12 are students at Swedish universities and 12 at US institutions.

Those who attended the event on-site had the opportunity to learn more about the talent programs in poster-sessions after the presentations.

“We already know about quite a few new ideas and new contacts that came from those discussions,” says Dr. Mats Nordlund, Head of AI Labs, who hosted the event together with Dr. Helena Theander, Head of Operations Data Factory.