By: Erin Vollick
9 Oct, 2023
Canada CIFAR AI Chair Jian Tang leverages graph machine learning and deep generative models to design proteins and molecules and find new therapeutic targets in drug discovery. While Tang’s fundamental research enjoys huge impact — his 2015 paper on LINE (large scale information network embedding) has been cited more than 5,500 times, and his paper on geometric diffusion models was one of the top 50-cited AI papers in 2022 — he’s pushing further.
“The Canada CIFAR AI Chair and the Mila environment have been critical to what I have accomplished so far, helping me connect and collaborate with pharmaceutical industry partners and other AI researchers.”
Jian Tang, Canada CIFAR AI Chair, Mila
“We design molecules, but we need to be able to test those molecules in a wet lab. That’s very important for making real impact,” says Tang, who founded the AI-biotech startup BioGeometry in 2022 to move those discoveries to testing. Currently, the company is focused on designing antibodies for difficult drug targets.
Tang recently also built and released two open-source machine learning AI platforms to enable community-wide small molecules and protein design, TorchDrug and TorchProtein. Pharmaceutical giant AstraZeneca has since leveraged TorchDrug to build a second drug discovery platform for their own research.
“The Canada CIFAR AI Chair and the Mila environment have been critical to what I have accomplished so far, helping me connect and collaborate with pharmaceutical industry partners and other AI researchers,” says Tang, who came to Canada as part of the Pan-Canadian AI Strategy in 2018.