ChemOS: An orchestration software to democratize autonomous discovery

Abstract

The current Edisonian approach to discovery requires up to two decades of fundamental and applied research for materials technologies to reach the market. Such a slow and capital-intensive turnaround calls for disruptive strategies to expedite innovation. Self-driving laboratories have the potential to provide the means to revolutionize experimentation by empowering automation with artificial intelligence to enable autonomous discovery. However, the lack of adequate software solutions significantly impedes the development of selfdriving laboratories. In this paper, we make progress towards addressing this challenge, and we propose and develop an implementation of ChemOS; a portable, modular and versatile software package which supplies the structured layers necessary for the deployment and operation of self-driving laboratories. ChemOS facilitates the integration of automated equipment, and it enables remote control of automated laboratories. ChemOS can operate at various degrees of autonomy; from fully unsupervised experimentation to actively including inputs and feedbacks from researchers into the experimentation loop. The flexibility of ChemOS provides a broad range of functionality as demonstrated on five applications, which were executed on different automated equipment, highlighting various aspects of the software package.

Accelerating organic solar cell material's discovery: high-throughput screening and big data

“…93 Accordingly, they demonstrated a substantial reduction on the number of data points required to complete the screening study (from 500-1000 to 30 samples only) by making use of the ChemOS orchestrating software package. 108,109 The raw material requirements were also downscaled proportionally: from 10-15 mg following conventionally-automated high-throughput experimentation to less than 1 mg per compound in the best self-driven scenario. While very promising, their highthroughput experimental screening procedure is thus far limited to drop cast films only, hence the assessment of the photovoltaic performance itself is not encompassed in their study and it is restricted to stability evaluations on bare films.…”

Deep learning and generative methods in cheminformatics and chemical biology: navigating small molecule space intelligently

“…This clearly applies to problems in chemistry and chemical biology that involve navigating large search spaces of molecules, and intelligent automation has been an important subset of activities here (e.g. [172][173][174][175][176][177][178][179][180][181][182]). 'Active learning' describes the kinds of methods that use knowledge of existing data to determine where best to explore next, and is normally used to balance exploration (looking for promising regions of the search space) with exploitation (a promising local search) [183].…”

Models of necessity

“…Currently, several research groups are working on the area of automated syntheses so that not only the reactions and workup are automated, but also the conditions of the reaction closely monitored, controlled, varied and automatically recorded. For examples, see [46][47][48][49]. A convergence of technologies, such as robotics, telemetrics, analytics, in addition to using ML/AI techniques, could allow an infrastructure to develop that enables data-driven discovery [50] and transforms chemistry from an empirical to a predictive science [51].…”