Active learning using coverage score

A computing device for computational drug design, comprising:

• an input component configured to receive data indicative of a population of a plurality of compounds, each compound having one or more molecular properties, and to receive data indicative of a training set of compounds from the population for which one or more biological properties are known;
• a processor configured to:
• select a subset of one or more compounds from the population that are not in the training set;
• determine, for each compound in the selected subset, a compound score based on at least one molecular property of a compound, wherein determining the compound score comprises comparing a frequency of a molecular property in a sampled set comprising the training set and the selected subset to a frequency of the molecular property in the population;
• determine a subset score of the selected subset based on the compound score of each compound in the selected subset;
• generate, based on the determined subset score, an updated training set of compounds comprising the one or more compounds of the selected subset;
• and
• train a machine learning model using the updated training set of compounds.