top of page
Machine Learning based prediction methods
Proficient property prediction
Is it possible to go from the structure of a molecule to its properties? If reactivity is understood perfectly, it should always be predictable. If toxicity cannot be related to structure, what else could we need to know? Machine learning can take us from data to classification and analysis. How proficient can computers be at understanding molecular properties? This lecture will look at applications of machine learning in suggesting the properties of substances from molecular structure.
Evolving the Cambridge Structural Database to derive new insights
The Cambridge Structural Database is a database of organic and metal-organic crystal structures that has been curated since the 1960s. Recently the number of structures has passed 1.25M. This presentation will introduce the CSD, show some of the recent ways in which we can build models based on the data within the database, how others can build models based on the database, and then speak to how CCDC is working to improve the repository going forwards to make the resource more comprehensive with additional structural meta-data.
bottom of page