By: Justine Brooks
16 Dec, 2024
Last week, the thirty-eighth annual Conference on Neural Information Processing Systems (NeurIPS) took place in Vancouver. One of the world’s preeminent conferences on machine learning, the event showcased work from across machine learning disciplines, including information theory, computer vision and linguistics.
NeurIPS first began in 1987 and is now a multi-track, interdisciplinary annual meeting featuring talks, demonstrations, workshops and oral and poster presentations, among other activities. This marks the first time the conference has been hosted in Canada since 2019.
Canada’s three National AI Institutes were well represented this year, with nearly 200 papers accepted from institute-affiliated researchers.
The Vector Institute in Toronto represented 98 accepted papers from Vector Faculty, Affiliates and post-doctoral students, including 51 of those papers co-authored by Canada CIFAR AI Chairs at Vector. Mila, located in Montréal, had over 60 papers accepted and hosted 80 workshops. Amii, based in Edmonton, represented 25 accepted papers, 17 authored by Canada CIFAR AI Chairs at Amii.
Canadian research was also recognized by the annual NeurIPS Test of Time Paper Award, which recognizes papers published ten years ago at NeurIPS 2014 that significantly shaped the research field. This year’s award made an exception to be presented to two papers, one of which was co-authored by Canada CIFAR AI Chairs at Mila, Yoshua Bengio and Aaron Courville among others. Their 2014 paper, ‘Generative Adversarial Nets,’ has been cited more than 85,000 times and has been foundational for generative modeling research.
This year at NeurIPS, CIFAR hosted a reception and townhall with our research community to provide updates on the recently announced CAISI Research Program at CIFAR and to share and discuss insights on potential priority research areas for the program. Over 60 Canada CIFAR AI Chairs and Fellows in the CIFAR Learning in Machines and Brains program attended.
CIFAR also partnered with the UK AI Safety Institute to host a roundtable discussion on AI Safety, bringing together leading experts in the field from Canada and the UK. Participants identified opportunities for collaboration between the two organizations and key priorities for researchers in AI safety. Outcomes and further activities regarding the CAISI Research Program at CIFAR will be announced in the new year.
Next year’s NeurIPS conference will be held in San Diego.
All photo credits on this page: Lisa King/TwinLens Photography