Active learning and free energy perturbation

I attended a Schrödinger webinar on active learning. I found it useful, and wanted to capture some of the interesting slides for when I need to revisit this.

Active learning is a kind of bootstrapping to build a better model without having to create a huge custom dataset

FEP is as a good as the real assay

Ultra-large screening as a service

Hit expansion: explore the space around your molecules of interest