Industry and academia are often seen as an either-or situation, but for Ari Kulmala it’s more of a both-and situation. Ari splits his time between Nokia, where he is the head of technology management and Tampere University, where he is a professor of practice in the Faculty of Information Technology and Communication Sciences. Collaboration between companies and the university is a defining feature of Ari’s work at Tampere - and of the city itself.
Ari and his colleagues at Tampere University are working on the design of specialized hardware that can accelerate AI applications like neural networks and machine learning. They’re working together with a group of industry partners to eventually fabricate a chip with optimized AI acceleration. “What we really aim to do is fill a gap,” Ari explains. “There are good people who know how machine learning and neural networks work, but they have never considered that they could design the processing hotbed themselves. There has been a lot of work to optimize the software side, but if we can also optimize the hardware side then it will create great benefits in power performance applications.” Ari’s background in both software and hardware (and his PhD in computer systems) is a huge advantage for this project. “Typically you either focus more on hardware or software, but I do both. One of the strengths of Tampere University actually is that you have the hardware and software designers in the same faculty. In most universities, this is not the case.” This unique set up facilitates the kind of interdisciplinary research that Tampere is known for.
Collaboration with industrial partners is essential to the project’s success. In his role as professor of practice, Ari has been able to draw on his wide network of national and international contacts to help facilitate some of the partnerships between the chipmakers and the university. “Due to my background, I can get, for example, the tooling we need which is not typically available to universities,” he says. They’re working with Nokia, for instance, on the chip architecture while other companies provide the neural network descriptors. Once prototype chips have been developed, some of the industry partners will also test the chips in their systems.
These links also lay the foundation for the university’s future growth. “Working with these partners allows us to get back some of the skills that once were at Tampere University, but went unfocused with the changes in mainstream industry needs during the last decade. We’re trying to get them more experience so they can master the whole production process at the university,” says Ari. He anticipates that during the next few years they will even see a few new startups emerge from this project, joining the numerous others that make Tampere Finland’s most attractive startup city.
Strong collaboration between academic research and companies has helped make Tampere one of the fastest growing cities in Finland and the country’s second largest economic centre. The city’s compact size helps facilitate cooperation between big multinationals and specialized startups and fosters an entrepreneurial culture that has helped Tampere shift from heavy industry to smart technology.
At Tampere University, Ari is driven by the potential for his work to have an impact on society. It’s what drew him to apply for the position in the first place. “When you work for one company, you work for that company's agenda and it tends to limit your viewpoint,” he says. “I’m enjoying myself at the university because I’m not limited in what I can work on. I like that you can think big here.”Continue reading