Cloud-Based Real-Time Molecular Screening Platform with MolFormer

With the prospect of automating a number chemical tasks with high fidelity, language processing models are emerging at rapid speed. Here, we present cloud-based real-time platform that allows users to virtually screen molecules interest. For this purpose, molecular embeddings inferred from recently proposed large model, named MolFormer, leveraged. The currently supports three tasks: nearest neighbor retrieval, space visualization, and property prediction. Based on functionalities results obtained, believe such can play pivotal role in chemistry engineering research, as well assist drug discovery material design tasks. A demo our is provided www.ibm.biz/molecular_demo .