Benchmarking Deep Learning for Molecule Retrosynthesis Prediction

Sunstella 2023 Summer Research Camp Submission6 Authors

Keywords: retrosynthesis; deep learning

Abstract: Retrosynthesis prediction is very important in organic chemistry and drug development. This task involves predicting the precursor reactants for synthesizing a target molecule, typically a drug. Deep learning has increasingly become indispensable in retrosynthesis prediction. Currently, numerous models focus on retrosynthesis prediction in drug discovery. However, there is a lack of unified platforms that facilitate the convenient utilization of these models. Furthermore, comparing their performance directly is challenging due to variations in datasets and evaluation metrics. Our objective is to develop a platform offering easy access to these models while benchmarking their performance under consistent experimental settings.

Submission Number: 6