RT Journal Article
SR 00
ID 10.1021/acs.jcim.9b00304
A1 Duros, Vasilios
A1 Grizou, Jonathan
A1 Sharma, Abhishek
A1 Mehr, S. Hessam M.
A1 Bubliauskas, Andrius
A1 Frei, Przemyslaw
A1 Miras, Haralampos N.
A1 Cronin, Leroy
T1 Intuition-enabled machine learning beats the competition when joint human-robot teams perform inorganic chemical experiments
JF Journal of Chemical Information and Modeling
YR 2019
FD 2019-06-24
VO 59
IS 6
SP 2664
OP 2671
AB Traditionally, chemists have relied on years of training and accumulated experience in order to discov-er new molecules. But the space of possible molecules so vast, only a limited exploration with the tra-ditional methods can be ever possible. This means that many opportunities for the discovery of inter-esting 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 towards the de-velopment 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 compar-ing collaboration between human experimenters with an algorithm-based search, against their perfor-mance 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 beat robots or humans working alone.
PB American Chemical Society
SN 1549-9596
LK https://eprints.gla.ac.uk/185445/