Therefore, doing classical machine learning on molecular data requires some form of feature engineering. While quantum computers might provide a comprehensive solution to this challenge in the future, it is nevertheless possible to use quantum mechanical and quantum computing ideas to guide both the choice of molecular representations and the design of machine learning models. Vid will describe both the theory behind these ideas as well as potential practical applications, focusing on problems in chemistry, and specifically in the pharmaceutical industry.
YOU MAY ALSO LIKE:
- Quantum Computing Startup Panel Discussion (SkillsCast recorded in March 2018)
- ML Pipelines from the Get Go (Without Tears) (SkillsCast recorded in August 2022)
- Keynote — Provably correct, asymptotically efficient, higher-order reverse-mode automatic differentiation (SkillsCast recorded in November 2021)
Machine Learning Models for Molecular Data
Vid Stojevic
CTO and co-founder of a new start-up called GTN - Generative Tensorial Networks ( http://gtn.ai ) - who are using “advanced cutting-edge quantum physics and machine learning methods to enable the next 150 years of drug discovery” and are currently at the stage of actively hiring 10 engineers/physicists + others to start their company.