Multi-objective Bayesian optimisation of a two-step synthesis of p-cymene from crude sulphate turpentine

The effect of chemical representation on active machine learning towards closed-loop optimization

“…ML-based optimizers have the potential to increase the efficiency in which chemical space is navigated, removing operator bias and ultimately reducing the total number of experiments required, thus reducing waste significantly. 7,[19][20][21] Previous reports vary in their conclusions on whether the implementation of chemical descriptors, rather than generic one-hot encoding (OHE) representations, truly boosted the predictive performance of their ML models. 19,22,23 Pdcatalyzed C(sp 3 )-H activation is a powerful reaction manifold that enables the facile introduction of functional complexity in small molecules, in addition to the late-stage functionalization of complex molecules like trimipramine (1), a tricyclic antidepressant, as recently demonstrated by one of our groups.…”

Automated pH Adjustment Driven by Robotic Workflows and Active Machine Learning

“…By applying this strategy in an iterative manner, it can be used to navigate through the parameter space, efficiently searching for desired conditions. [29][30][31] In the case of pH adjustment this refers to the amount of acid or base to achieve a target pH. Algorithm 1 illustrates the control code used for automated pH adjustment.…”

Continuous stirred-tank reactor cascade platform for self-optimization of reactions involving solids

“…[18][19][20][21][22][23][24][25] Additionally, integration of analytical techniques and controls of reactions conditions produces closed-loop systems that, when paired with an optimization protocol, enable automatic selection and evaluation of new sets of reaction conditions based on prior data. 4,[26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41] Such systems have the potential to reduce the burden of repeated manual experiments for optimization, allowing time for more creative tasks.…”