Here we summarize recent progress in machine learning for the chemical sciences. We outline machine-learning techniques that are suitable for addressing research questions in this domain, as well as future directions for the field. We envisage a future in which the design, synthesis, characterization and application of molecules and materials is accelerated by artificial intelligence.

/pdf/machine-learning-for-molecular-and-materials-science-2c34wzvcqw.pdf

Machine learning for molecular and materials science.

This presentation is a case study taken from the travel and holiday industry. Paxport/Multicom, based in UK and Sweden, have recently adopted a recommendation system for holiday accommodation bookings. Machine learning techniques such as Collaborative Filtering have been applied using Python (3.5.1), with Jupyter (4.0.6) as the main framework. Data scale and sparsity present significant challenges in the case study, and so the effectiveness of various techniques are described as well as the performance of Python-based libraries such as Python Data Analysis Library (Pandas), and Scikit-learn (built on NumPy, SciPy and matplotlib). The presentation is suitable for all levels of programmers.

Machine learning with Python

Electrochemical carbon dioxide recycling provides an attractive approach to synthesizing fuels and chemical feedstocks using renewable energy. On the path to deploying this technology, basic and applied scientific hurdles remain. Integrating catalytic design with mechanistic understanding yields scientific insights and progresses the technology towards industrial relevance. Catalysts must be able to generate valuable carbon-based products with better selectivity, lower overpotentials and improved current densities with extended operation. Here, we describe progress and identify mechanistic questions and performance metrics for catalysts that can enable carbon-neutral renewable energy storage and utilization. Electrochemical carbon dioxide reduction is an attractive approach for obtaining fuels and chemical feedstocks using renewable energy. In this Review, the authors describe progress so far, identify mechanistic questions and performance metrics, and discuss design principles for improved activity and selectivity.

/pdf/designing-materials-for-electrochemical-carbon-dioxide-39fig3zw4f.pdf

Designing materials for electrochemical carbon dioxide recycling

Technologies such as batteries, biomaterials and heterogeneous catalysts have functions that are defined by mixtures of molecular and mesoscale components. As yet, this multi-length-scale complexity cannot be fully captured by atomistic simulations, and the design of such materials from first principles is still rare1–5. Likewise, experimental complexity scales exponentially with the number of variables, restricting most searches to narrow areas of materials space. Robots can assist in experimental searches6–14 but their widespread adoption in materials research is challenging because of the diversity of sample types, operations, instruments and measurements required. Here we use a mobile robot to search for improved photocatalysts for hydrogen production from water15. The robot operated autonomously over eight days, performing 688 experiments within a ten-variable experimental space, driven by a batched Bayesian search algorithm16–18. This autonomous search identified photocatalyst mixtures that were six times more active than the initial formulations, selecting beneficial components and deselecting negative ones. Our strategy uses a dexterous19,20 free-roaming robot21–24, automating the researcher rather than the instruments. This modular approach could be deployed in conventional laboratories for a range of research problems beyond photocatalysis. A mobile robot autonomously operates analytical instruments in a wet chemistry laboratory, performing a photocatalyst optimization task much faster than a human would be able to.

/pdf/a-mobile-robotic-chemist-204cvrj0qk.pdf

A mobile robotic chemist

The discovery and development of novel materials in the field of energy are essential to accelerate the transition to a low-carbon economy. Bringing recent technological innovations in automation, robotics and computer science together with current approaches in chemistry, materials synthesis and characterization will act as a catalyst for revolutionizing traditional research and development in both industry and academia. This Perspective provides a vision for an integrated artificial intelligence approach towards autonomous materials discovery, which, in our opinion, will emerge within the next 5 to 10 years. The approach we discuss requires the integration of the following tools, which have already seen substantial development to date: high-throughput virtual screening, automated synthesis planning, automated laboratories and machine learning algorithms. In addition to reducing the time to deployment of new materials by an order of magnitude, this integrated approach is expected to lower the cost associated with the initial discovery. Thus, the price of the final products (for example, solar panels, batteries and electric vehicles) will also decrease. This in turn will enable industries and governments to meet more ambitious targets in terms of reducing greenhouse gas emissions at a faster pace. The discovery and development of advanced materials are imperative for the clean energy sector. We envision that a closed-loop approach, which combines high-throughput computation, artificial intelligence and advanced robotics, will sizeably reduce the time to deployment and the costs associated with materials development.

/pdf/accelerating-the-discovery-of-materials-for-clean-energy-in-2ejogam5fn.pdf

Accelerating the discovery of materials for clean energy in the era of smart automation

The discovery of chemical reactions is an inherently unpredictable and time-consuming process1. An attractive alternative is to predict reactivity, although relevant approaches, such as computer-aided reaction design, are still in their infancy2. Reaction prediction based on high-level quantum chemical methods is complex3, even for simple molecules. Although machine learning is powerful for data analysis4,5, its applications in chemistry are still being developed6. Inspired by strategies based on chemists’ intuition7, we propose that a reaction system controlled by a machine learning algorithm may be able to explore the space of chemical reactions quickly, especially if trained by an expert8. Here we present an organic synthesis robot that can perform chemical reactions and analysis faster than they can be performed manually, as well as predict the reactivity of possible reagent combinations after conducting a small number of experiments, thus effectively navigating chemical reaction space. By using machine learning for decision making, enabled by binary encoding of the chemical inputs, the reactions can be assessed in real time using nuclear magnetic resonance and infrared spectroscopy. The machine learning system was able to predict the reactivity of about 1,000 reaction combinations with accuracy greater than 80 per cent after considering the outcomes of slightly over 10 per cent of the dataset. This approach was also used to calculate the reactivity of published datasets. Further, by using real-time data from our robot, these predictions were followed up manually by a chemist, leading to the discovery of four reactions.

/pdf/controlling-an-organic-synthesis-robot-with-machine-learning-4squxv5lvc.pdf

Controlling an organic synthesis robot with machine learning to search for new reactivity

The synthesis of complex organic compounds is largely a manual process that is often incompletely documented. To address these shortcomings, we developed an abstraction that maps commonly reported methodological instructions into discrete steps amenable to automation. These unit operations were implemented in a modular robotic platform using a chemical programming language which formalizes and controls the assembly of the molecules. We validated the concept by directing the automated system to synthesize three pharmaceutical compounds, Nytol, rufinamide, and sildenafil, without any human intervention. Yields and purities of products and intermediates were comparable to or better than those achieved manually. The syntheses are captured as digital code that can be published, versioned, and transferred flexibly between platforms with no modification, thereby greatly enhancing reproducibility and reliable access to complex molecules.

/pdf/organic-synthesis-in-a-modular-robotic-system-driven-by-a-55znp5yxug.pdf

Organic synthesis in a modular robotic system driven by a chemical programming language

Although extending the reactivity of a given class of molecules is relatively straightforward, the discovery of genuinely new reactivity and the molecules that result is a wholly more challenging problem. If new reactions can be considered unpredictable using current chemical knowledge, then we suggest that they are not merely new but also novel. Such a classification, however, requires an expert judge to have access to all current chemical knowledge or risks a lack of information being interpreted as unpredictability. Here, we describe how searching chemical space using automation and algorithms improves the probability of discovery. The former enables routine chemical tasks to be performed more quickly and consistently, while the latter uses algorithms to facilitate the searching of chemical knowledge databases. Experimental systems can also be developed to discover novel molecules, reactions and mechanisms by augmenting the intuition of the human expert. In order to find new chemical laws, we must seek to question current assumptions and biases. Accomplishing that involves using two areas of algorithmic approaches: algorithms to perform searches, and more general machine learning and statistical modelling algorithms to predict the chemistry under investigation. We propose that such a chemical intelligence approach is already being used and that, in the not-too-distant future, the automated chemical reactor systems controlled by these algorithms and monitored by a sensor array will be capable of navigating and searching chemical space more quickly, efficiently and, importantly, without bias. This approach promises to yield not only new molecules but also unpredictable and thus novel reactivity. Automation can help in performing routine tasks quickly and consistently. Algorithms facilitate the searching of current knowledge. Combining the two could lead to a chemically intelligent approach to the discovery of not only new molecules but also novel and unpredictable reactivity.

How to explore chemical space using algorithms and automation

There is a growing drive in the chemistry community to exploit rapidly growing robotic technologies along with artificial intelligence-based approaches. Applying this to chemistry requires a holistic approach to chemical synthesis design and execution. Here, we outline a universal approach to this problem beginning with an abstract representation of the practice of chemical synthesis that then informs the programming and automation required for its practical realization. Using this foundation to construct closed-loop robotic chemical search engines, we can generate new discoveries that may be verified, optimized, and repeated entirely automatically. These robots can perform chemical reactions and analyses much faster than can be done manually. As such, this leads to a road map whereby molecules can be discovered, optimized, and made on demand from a digital code.

Universal chemical synthesis and discovery with 'The Chemputer'

The exploration of chemical space for new reactivity, reactions and molecules is limited by the need for separate work-up-separation steps searching for molecules rather than reactivity. Herein we present a system that can autonomously evaluate chemical reactivity within a network of 64 possible reaction combinations and aims for new reactivity, rather than a predefined set of targets. The robotic system combines chemical handling, in-line spectroscopy and real-time feedback and analysis with an algorithm that is able to distinguish and select the most reactive pathways, generating a reaction selection index (RSI) without need for separate work-up or purification steps. This allows the automatic navigation of a chemical network, leading to previously unreported molecules while needing only to do a fraction of the total possible reactions without any prior knowledge of the chemistry. We show the RSI correlates with reactivity and is able to search chemical space using the most reactive pathways.

/pdf/an-autonomous-organic-reaction-search-engine-for-chemical-2qas23i5f6.pdf

Jarosław M. Granda

Papers

Controlling an organic synthesis robot with machine learning to search for new reactivity

Organic synthesis in a modular robotic system driven by a chemical programming language

How to explore chemical space using algorithms and automation

Universal chemical synthesis and discovery with 'The Chemputer'

An autonomous organic reaction search engine for chemical reactivity.