Bayer Taps High-Powered Google TPUs for Drug Discovery
Drug discovery is a notoriously expensive, tedious and -- despite the relatively recent inclusion of quantum computing -- time-consuming process. Pharmaceutical giant Bayer intends to alleviate the burden with a new partnership with Google.
In a collaboration announced this week, Bayer will use Google's machine learning-specific Tensor Processing Units (TPUs) to speed up the quantum computing calculations that Bayer uses in its drug discovery process.
"By combining Google Cloud's computing power with Bayer's leading expertise in drug discovery we intend to unleash the potential of large-scale quantum chemistry," said Marianne De Backer, the head of strategy and business development at Bayer, in a prepared statement.
According to Google, its TPUs are a boon for accurately training vast and complex workloads, the likes of which are fundamental in modern in drug discovery:
Cloud TPU resources accelerate the performance of linear algebra computation, which is used heavily in machine learning applications. TPUs minimize the time-to-accuracy when you train large, complex neural network models. Models that previously took weeks to train on other hardware platforms can converge in hours on TPUs.
Bayer will use Google TPUs to speed up and scale the quantum chemistry calculations it already uses for drug discovery, as well as to "demonstrate fully quantum mechanical modeling of protein-ligand interactions."
According to Bayer, "The results will determine the scientific and economic viability of large-scale density functional theory calculations for practical applications."
The company argues that increasing the efficiency of the drug discovery process in this way is critical for meeting the needs of patients, with Google Cloud CEO Thomas Kurian calling faster drug discovery "one of the most important applications for AI and high-performance computing."