Intuition-Enabled Machine Learning Beats the Competition When Joint Human-Robot Teams Perform Inorganic Chemical Experiments

Abstract

Traditionally, chemists have relied on years of training and accumulated experience in order to discover new molecules. But the space of possible molecules is so vast that only a limited exploration with the traditional methods can be ever possible. This means that many opportunities for the discovery of interesting phenomena have been missed, and in addition, the inherent variability of these phenomena can make them difficult to control and understand. The current state-of-the-art is moving toward the development of automated and eventually fully autonomous systems coupled with in-line analytics and decision-making algorithms. Yet even these, despite the substantial progress achieved recently, still cannot easily tackle large combinatorial spaces, as they are limited by the lack of high-quality data. Herein, we explore the utility of active learning methods for exploring the chemical space by comparing the collaboration between human experimenters with an algorithm-based search against their performance individually to probe the self-assembly and crystallization of the polyoxometalate cluster Na6[Mo120Ce6O366H12(H2O)78]·200H2O (1). We show that the robot-human teams are able to increase the prediction accuracy to 75.6 ± 1.8%, from 71.8 ± 0.3% with the algorithm alone and 66.3 ± 1.8% from only the human experimenters demonstrating that human-robot teams can beat robots or humans working alone.

Autonome Entdeckung in den chemischen Wissenschaften, Teil I: Fortschritt

“…[296] ¾hnliche Closed-Loop-Optimierungen werden auch für Materialanwendungen durchgeführt. Tr otz unterschiedlicher Analytik fürd ie verschiedenen interessierenden Eigenschaften ist der Gesamtablauf derselbe.O ptimierungsziele waren hier unter anderem die Emissionsintensitätv on Quantenpunkten, [297] die Umwandlung und Partikelgrçße bei einer Copolymerisation, [290] die Identifizierung von Kristallisationsbedingungen fürP olyoxometallate, [298,299] die Herstellung von Bose-Einstein-Kondensaten [300] und die Synthese einer Metall-organischen Gerüstverbindung (MOF) mit großer Oberfläche. [301] Besonders bemerkenswert ist die Optimie-…”

“…Trotz unterschiedlicher Analytik für die verschiedenen interessierenden Eigenschaften ist der Gesamtablauf derselbe. Optimierungsziele waren hier unter anderem die Emissionsintensität von Quantenpunkten, [297] die Umwandlung und Partikelgröße bei einer Copolymerisation, [290] die Identifizierung von Kristallisationsbedingungen für Polyoxometallate, [298, 299] die Herstellung von Bose‐Einstein‐Kondensaten [300] und die Synthese einer Metall‐organischen Gerüstverbindung (MOF) mit großer Oberfläche [301] . Besonders bemerkenswert ist die Optimierung der MOF‐Synthese von Moosavi et al., da hier die relative Bedeutung der Syntheseparameter aus Daten aus früheren MOF‐Synthesen abgeschätzt wurde, um maximal unterschiedliche experimentelle Startdesigns zu ermöglichen, wodurch direkt mit der empirischen iterativen Optimierung begonnen werden konnte [301] …”

Autonomous Discovery in the Chemical Sciences Part I: Progress

“…[292,293] Multi-step reactions are particularly challenging to optimize because the effects of changing one parameter can propagate through downstream process steps.T hey are typically broken up into individual synthetic steps to improve the tractability of the problem [294,295] or optimized approximately through screening, rather than through true closed-loop feedback. [296] Similar closed-loop optimizations have been demonstrated for materials-focused applications.D ifferent properties of interest necessitate different analytical endpoints,b ut the overall workflow is the same.O ptimization goals have included the emission intensity of quantum dots, [297] the conversion and particle size resulting from ac opolymerization, [290] the identification of crystallization conditions for polyoxometalates, [298,299] the production of Bose-Einstein condensates, [300] and the realization of ametal-organic framework (MOF) with high surface area. [301] TheM OF synthesis optimization by Moosavi et al is particularly noteworthy in that prior data on syntheses of other MOFs were used to estimate the relative importance of synthetic parameters to enable am aximally diverse initial design of experiments, jump-starting the phase of iterative empirical optimization.…”

“…Different properties of interest necessitate different analytical endpoints, but the overall workflow is the same. Optimization goals have included the emission intensity of quantum dots, [297] the conversion and particle size resulting from a copolymerization, [290] the identification of crystallization conditions for polyoxometalates, [298, 299] the production of Bose–Einstein condensates, [300] and the realization of a metal–organic framework (MOF) with high surface area [301] . The MOF synthesis optimization by Moosavi et al.…”

Universal Chemical Synthesis and Discovery with ‘The Chemputer’

“…In addition, the algorithm explored a wider range of space that would need to be performed either by human or purely random search. Recently, the same researchers observed that collaboration between smart robotics and humans may be even more efficient than either alone [22]. Grizou and colleagues described a chemical robotic discovery assistant equipped with a curiosity algorithm that can efficiently explore a complex chemical system in search of complex emergent phenomena exhibited by proto-cell droplets [23].…”