Nov 25, 2021

Let’s Discuss Graph Neural Networks

Written by

We publish dozens of new articles every week on TDS, covering a dizzying range of topics. We’re here to help you avoid decision paralysis: in this week’s Variable, we focus on GNNs (graph neural networks) and invite you to explore this exciting subdomain of machine learning with three standout articles. (If GNNs aren’t your thing,…

We publish dozens of recent articles each week on TDS, protecting a dizzying vary of matters. We’re right here that can assist you keep away from resolution paralysis: on this week’s Variable, we give attention to GNNs (graph neural networks) and invite you to discover this thrilling subdomain of machine studying with three standout articles. (If GNNs aren’t your factor, scroll down for the remainder of our weekly highlights.)What can differential geometry and algebraic topology inform us about GNNs? A brand new collection from Michael Bronstein is all the time a trigger for celebration, and his new one isn’t any exception. Within the first installment, Michael lays the groundwork for the work to come back, addresses a few of GNNs’ frequent plights, and means that approaching them by way of the lens of differential geometry and algebraic topology can “deliver a brand new perspective to necessary and difficult issues in graph machine studying.”Should you’re new to GNNs and will use an accessible explainer on the subject earlier than diving into the nitty-gritty of Michael’s submit, no worries! Begin with Shanon Hong’s mild introduction to graph concept, which additionally covers what GNNs can and can’t do. Subsequent, head proper over to Aishwarya Jadhav’s perennial favourite, the place she discusses graph neural networks’ real-world functions.Photograph by Al Soot on UnsplashWe’ve been busy on a number of different fronts this previous week—should you’d wish to make amends for a few of our greatest current articles, let the click and bookmarking start:How does information science form law- and policy-making on the German Ministry of Well being? Learn Elliot Gunn’s insightful Q&A with Lars Roemheld, Director of AI & Information on the well being innovation hub, the place they talk about the confluence of information science and authorities innovation.Must you or shouldn’t you monitor ML fashions? It’s a seemingly easy query, however Devanshi Verma (and coauthors Matthew Fligiel, Rupa Ghosh, and Dr. Arnab Bose) discover its stakes (and associated complexities) at depth.For a hands-on method to studying a brand new ability, take a look at Parul Pandey’s newest, the place she exhibits the right way to create interactive maps with GeoPandas.How do you win a Kaggle competitors and assist your group come out first amongst 2650 different entrants? Gilles Vandewiele walks us by way of the ups, downs, and breakthrough moments of his (and his teammates’) current success.The intersection of AI, fairness, and healthcare is presently on the heart of a number of essential conversations, and Angela Wilkins’s current submit fleshes out the real-world dangers of bias creep and unequal information.Staying near AI’s results on societies and communities, Jeremie Harris’s TDS Podcast episode with Gillian Hadfield coated the present challenges of regulators and legislators to maintain up with technological progress, and pointed to the advantages of AI that isn’t merely explainable, but additionally justifiable.Thanks for becoming a member of us on one other week of studying and exploration, and for participating with the work we publish.Till the subsequent Variable,TDS Editors

Article Tags:
Article Categories:
Graph · Let's

Leave a Reply

Your email address will not be published. Required fields are marked *