Active learning–driven design space reduction for material discovery via Bayesian optimization : a COF case study
Μάθηση με ενεργή συμμετοχή για τη μείωση του χώρου σχεδιασμού στην ανακάλυψη υλικών μέσω βελτιστοποίησης Bayes : μια μελέτη περίπτωσης COF

View/ Open
Keywords
Active learning ; Bayesian optimizationAbstract
This thesis explores the applications of Active Learning and Bayesian Optimization in material discovery, with the objective of identifying optimal material configurations in a cost-efficient manner. The work focuses on leveraging machine learning models to explore, prune and optimize vast design spaces with the fewest resources possible.
Findings indicate that a combination of machine learning models and algorithms provide robust tools when exploring uncharted design spaces and doing so in an inexpensive manner. The setup used is not tightly chained to a specific dataset or material class, making it broadly applicable to abstract optimization problems.


