Success for the future's deep learning

Today’s experts and future researchers in the field of deep learning, met at Luleå University of Technology during DeepLearn 2022 Autumn, the world’s largest research training event in AI. – It’s been a great experience; we had the chance to interact with some of the world’s most reputable researchers, says Marcus Liwicki, Professor of Machine Learning at Luleå University of Technology.

– Of course, we know some of them already, but them coming to Luleå gave all our AI related researchers, including PhD and Master students, a chance to interact with them. In fact, we have already found ideas for new research projects, joint co-supervision, and research exchange!

DeepLearn 2022 Autumn is a recurring event, updating participants on the most recent advances in the critical and fast developing area of deep learning. This time, 240 graduate students, postgraduate students, and industry practitioners met in Luleå, attending short courses, lectures, and face to face networking. Among the invited speaker were top researchers and representatives from companies such as Apple, Amazon, Nvidia, and Zoom.

Interesting keynote speakers

The conference keynote speakers, Elaine O. Nsoesie and Tommaso Dorigo, are both great examples of how far and diverse the research area of deep learning has come. Elaine O. Nsoesie is an Associate Professor in the Department of Global Health at the Boston University School of Public Health, focused on bias in health data and AI algorithms. According to her, there need to be structures that ensure that data and algorithms are used to redress health inequity.

– We must be aware of how we use health data. We need data for developing technology, but who will benefit from it? There are so many health applications, for example they can help us to prevent sickness and improve public health. But in the end, we need to consider that there is an ethical dilemma, says Elaine O. Nsoesie.

Tommaso Dorigo is an experimental particle physicist, working at the Italian National Institute of Nuclear Physics. At DeepLearn, he talked about the challenges and opportunities for using the application of differentiable programming to research in particle physics and related research areas.

– Historically, we physicists have designed and built instruments ourselves. We use thousands of parameters, and we want to achieve redundancy and we aim at robustness. If we want to improve this, we can use computer science and machine learning tools to find solutions for intractable problems, says Tommaso Dorigo.

Variety of topics

DeepLearn 2022 Autumn addresses a wide range of subjects and application areas.

– But the theoretical aspects are always the same, no matter what the subject is. That’s why it is so rewarding to be part of such an event, says Elaine O. Nsoesie.

Tommaso Dorigo agrees:

– It is great to interact with other research groups since you are always looking for new collaborations.

Marcus Liwicki highlights that many of the participants were fascinated and inspired by the modern infrastructure at the university.

– Some of them decided to apply for positions to move here or at least join us during internships, says Marcus Liwicki.

One of the speakers, Professor Alper Yilmaz from The Ohio State University, was pleased not only with the conference, but also with Luleå.

– This is a destination, I never thought that I would come to. But now I am sure that I will come here again!