Finding the most potent compounds using active learning on molecular pairs

• Zachary Fralish and
• Daniel Reker

Beilstein J. Org. Chem. 2024, 20, 2152–2162, doi:10.3762/bjoc.20.185

• regimes by enabling faster and more accurate identification of more diverse molecular hits against critical drug targets. Keywords: active learning; drug design; machine learning; molecular optimization; potency predictions; Introduction Active learning is a powerful concept in molecular machine

• ][4] and steer molecular optimization for drug discovery [5][6][7][8]. Active learning is particularly powerful during early project stages. However, one major downside is that, at these early project stages, only a very small amount of training data is available to learn from [9] which can be

• , model exploitation can lead to analog identification, which can limit the acquired knowledge and the scaffold diversity of selected hits [1]. We previously showed that leveraging pairwise molecular representations as training data can support molecular optimization by directly training on and predicting