Abstract
DASyR-LLM embeds a large language model inside an iterative symbolic regression algorithm to discover kinetic models. At each iteration the LLM critiques candidate rate expressions on physicochemical grounds and proposes new ones, cutting the number of iterations needed to find the true model by 42–79% compared with a state-of-the-art symbolic regression method across four case studies in catalysis and bioprocessing. Because each iteration typically corresponds to a new experiment, this could substantially reduce experimental effort.
Type
Publication
arXiv preprint
Dr Paulina Quintanilla
Assistant Professor in Process Systems Engineering
Assistant Professor in Process Systems Engineering at UCL, leading the LOOPS research group. Research interests include machine learning, optimisation and control for processes and physical systems.