Thank you for participating in NLDL 2024!

By: Petter Bjørklund, Communications Advisor, Visual Intelligence

Thank you to our participants who travelled to Tromsø to attend this year’s Northern Lights Deep Learning Conference (NLDL) from the 9th to 11th of January! It has been a joyous experience meeting all of you, and we hope you have made several new acquaintances and colleagues during your stay.

250 researchers from 28 different countries gathered at the conference to exchange ideas and experiences within the fields of artificial intelligence (AI) and deep learning.

‍You can also read the current article in Norwegian at uit.no.

High-quality scientific contributions and keynotes

The conference comprised of insightful and diverse poster sessions from our participants, as well as thought-provoking keynotes and discussions on relevant and emerging topics within the mentioned fields.

Keynote topics included the trustworthiness of AI (Aasa Feragen), chances and challenges of Reliable AI (Gitta Kutyniok), fair self-supervised learning in multiple modalities with applications to medicine (Narges Razavian), and The Statistical Finite Element method (Mark Girolami).

Participants also presented their research in a series of enriching and informative oral sessions throughout the NLDL week.

Their presentations involved research from a multitude of AI-related subtopics, such as deep reinforcement learning, dual convolutional neural network pipelines, and neural machine translation.

‍

The level of scientific contributions that is presented at NLDL is of high quality and continues to increase by each year. We want to thank everyone who participated and contributed to what has been an exciting and interesting week, says associate professor Michael Kampffmeyer, the general co-chair of the NLDL committee.

Sneak peek of AI exhibition for children

Several participants also partook in the NLDL Winter School, a five-day course consisting of various tutorials by experts in the AI field.

Invited speakers from The University of Cambridge (Mihaela van der Schaar and Robert Davis) and Norwegian AI Cloud (Sabry Razick, Hicham Agueny, and Magnar Bjørgve) gave tutorials on specific topics on deep learning such as Synthetic Data, Generative Models, and Explainability.

- It has been very satisfying and delightful to interact with diverse students from all over the world throughout the entire week. It is great to see that so many decided to participate in the Winter School, says van der Schaar.

The participants were given a surprise sneak peek of the new AI exhibition for children at Nordnorsk vitensenter, a project which is spearheaded by Samuel Kuttner and Benjamin Ricaud of the UiT Machine Learning Group.

The exhibition aims to foster youths’ interest and understanding of AI-related technologies, as well as their ability to think critically about AI tools’ relevance to their everyday lives and society in general.

- It further aims to recruit future researchers to the field and will demonstrate how AI can be used in self-driving cars, to manipulate pictures, to recognize and characterize features such as age, facial expressions, gender and mood, as well as to assist in identifying abnormalities and diseases inside our bodies, says Kuttner.

Kuttner and Ricaud have been hard at work to ensure that the exhibition is as robust as possible. The duo gave participants the opportunity to beta test the different AI-assisted interfaces, ranging from drawing tools to facial recognition tools, before its impending launch later this spring.

Diversity, inclusion, and industry

NLDL aims to be an arena for promoting discussion and debate on relevant topics within the AI field, one of which is the disparities in AI which was addressed in the side event “Diversity in AI”.

The event gathered representatives from industry, public administration and research to discuss initiatives to ensure inclusivity and diversity in AI for the future.

The event consisted of a series of keynotes and a moderated panel discussion between the keynote speakers, which included Narges Razavian (New York University Medical Center), Alvaro Fernandez (Stavanger University Hospital), and Elisavet Kozyri (University of Tromsø). The panel moderator was Elisabeth Wetzer (University of Tromsø).

- There is a big question in AI concerning how we can make sure that these models do not mimic old biases, that they perform well across all dimensions, and learn good features for all subpopulations. If they do not and are put into practice, the gap gets even worse, says Razavian.

The position of AI in industry was the overarching topic of “The Industry Event”, where invited experts discussed why AI is important for industry and how it can provide additional value to different industrial domains.

- The event also aimed to address the problems AI can solve, and how academia and business should collaborate on facilitating the safe and sustainable use of AI, says Klas Pettersen, the CEO of NORA.ai.

The audience asked stimulating questions to the panel speakers, providing a sound and reflective discussion on the applications of AI in industry.

See you all next year!

The NLDL conference returns in 2025 on the 7th to the 9th of January alongside the NLDL Winter School from 6th-10th of January. Be sure to mark your calendars, and we hope to see you all again next January!

The NLDL conference is held annually by the research center Visual Intelligence and UiT The Arctic University of Norway, with NORA.ai as a core collaborative partner.