Bayer imaging FM classifies drug targets using Amazon SageMaker HyperPod

Bayer Pharmaceuticals used Amazon SageMaker HyperPod to revolutionize their drug discovery process by training foundation models (FMs) for analyzing biomedical imaging data.

• Trained self-supervised imaging foundation models like DINO, MAE, and SimCLR
• Processed 50 TB of cell culture and histopathological images
• Used a cluster of four ml.p4de.24xlarge EC2 instances with 8 NVIDIA A100 GPUs each
• Can now analyze data from 100,000 compounds in high-content screening experiments
• Accelerates drug candidate identification through advanced AI image analysis

The solution enables Bayer to quickly process complex imaging data, identify promising therapeutic candidates, and streamline their drug discovery workflow using cutting-edge AI technologies.