AI-accelerated materials innovation: From Optoelectronics to Fluorescent Biomarkers

C. Kreisbeck
Kebotix,
United States

Keywords: materials informatics, deep learning, generative AI

Summary:

The declining R&D productivity keeps imposing significant challenges in today's chemical and materials industry. Despite spending over $150B on research, on average, it takes about 10-20 years to bring a new product to market. Recent successes around virtual high-throughput screening, materials informatics, and artificial intelligence (AI) provokes rethinking innovation pipelines to enable "R&D at scale" and a paradigm shift towards a "Moore’s law for scientific discovery". This talk illustrates how Kebotix leverages its advanced digital innovation platform to accelerate the discovery of novel organic molecules for various applications. Instead of following the Edisonian approach to scientific discovery, KBX develops advanced deep learning algorithms, computational modeling pipelines, and generative AI algorithms to solve the inverse design, where the desired properties are decoded into the corresponding molecular structure. Our flexible workflow and data management system automatically collects data in a structured and machine learnable way, allowing for efficient screening of millions of compounds.