Insilico presented its new molecular generation model at the 35th AAAI Conference on Artificial Intelligence
Credit: Insilico Medicine
February 11, 2021 – Insilico Medicine, a global leader in artificial intelligence (AI) for drug discovery and development, proposed a new molecular graph generative model called MolGrow at the 35th AAAI Conference on Artificial Intelligence on February 5, 2021.
The AAAI conference promotes research in AI and scientific exchange among AI researchers, practitioners, scientists, and engineers in affiliated disciplines. AAAI-21 has a diverse technical track, student abstracts, poster sessions, invited speakers. Insilico Medicine presented during one of the poster sessions.
The researchers introduced a novel model, based on the normalizing flows approach, which produces molecular structures from a single-node graph by recursively splitting every node into two. The hierarchical nature allows for precise modifications of the generated graphs using latent representations. The proposed model demonstrated improvement of distribution learning and molecular optimization metrics among graph-based models.
“We are happy to present our latest results on molecular graph modeling — a novel approach that combines advantages of invertible models and hierarchical generation. We hope our research gives new insights and ideas on how to improve graph generators,” said Max Kuznetsov, Research Scientist at Insilico Medicine.
“At Insilico Medicine, we have been developing generative models for chemistry since 2016. We developed and integrated many models into our generative chemistry platform — Chemistry42. We continue improving our generative tools, and with this paper, we explored graph-based generators,” said Daniil Polykovskiy, IT Director at Insilico Medicine. “Our new model, MolGrow, is the new state-of-the-art among node-level graph generative models on both distribution learning and molecular property optimization tasks.”
About Insilico Medicine
Insilico Medicine develops software that leverages generative models, reinforcement learning (RL), and other modern machine learning techniques for the generation of new molecular structures with specific properties. Insilico Medicine also develops software for the generation of synthetic biological data, target identification, and the prediction of clinical trials outcomes. The company integrates two business models; providing AI-powered drug discovery services and software through its Pharma.AI platform and developing its own pipeline of preclinical programs. The preclinical program is the result of pursuing novel drug targets and novel molecules discovered through its platforms. Since its inception in 2014, Insilico Medicine has raised over $52 million and received multiple industry awards. Insilico Medicine has also published over 100 peer-reviewed papers and has applied for over 25 patents.
Insilico Medicine contact information