Leveraging AI for Science


Scientific advances over the last several centuries have not only resulted in a greater understanding of the universe; they’ve raised the standard of living for many people across the globe. However, there are still massive challenges facing humanity, as evidenced by climate change and the COVID-19 pandemic. One of the difficulties of modern science is to make sense of the vast amount of information we’ve gathered about the world - from the Large Hadron Collider to massive genome projects—it’s impossible for any individual person to parse it all. In this lecture, I will discuss how AI techniques like machine learning can use information and data to dramatically improve solutions to challenging scientific problems, and accelerate fundamental discoveries by deepening the nature of questions researchers can ask.

About the speaker

Pushmeet Kohli Pushmeet Kohli leads the AI for Science team at Google DeepMind which aims to leverage AI and ML techniques to accelerate progress on important scientific challenges. The team conducts research in many areas of Science and has made contributions in structural biology (protein folding), quantum chemistry, genomics and pure mathematics. One of the models developed by the team, AlphaFold, was recognized as a breakthrough solution of the 50 year old challenge problem of protein structure prediction. Pushmeet also leads DeepMind’s research on Safe and Trustworthy AI which covers topics like neural networks verification, model robustness, calibration of predictive uncertainty and model interpretability. These team develops and evaluated these techniques for applications in natural language understanding, medical imaging, computer vision among others.