HyFactor: Hydrogen-count labelled graph-based defactorization Autoencoder

Tagir Akhmetshin; Arkadii Lin; Daniyar Mazitov; Evgenii Ziaikin; Timur Madzhidov; Alexandre Varnek
ChemRxiv preprint and has not been peer-reviewed.
AI/ML Published: (Dec/2021)
DOI: https://chemrxiv.org/engage/chemrxiv/article-details/61aa38576d4e8f3bdba8aead
Abstract:

Graph-based architectures are becoming increasingly popular as a tool for structure generation. Here, we introduce a novel open-source architecture HyFactor which is inspired by previously reported DEFactor architecture and based on the hydrogen labeled graphs. Since the original DEFactor code was not available, its new implementation (ReFactor) was prepared in this work for the benchmarking purpose. HyFactor demonstrates its high performance on the ZINC 250K MOSES and ChEMBL data set and in molecular generation tasks, it is considerably more effective than ReFactor. The code of HyFactor and all models obtained in this study are publicly available from our GitHub repository: https://github.com/Laboratoire-de- Chemoinformatique/hyfactor