Autonomous discovery in the chemical sciences part II: Outlook

Abstract

This two-part review examines how automation has contributed to different aspects of discovery in the chemical sciences. In this second part, we reflect on a selection of exemplary studies. It is increasingly important to articulate what the role of automation and computation has been in the scientific process and how that has or has not accelerated discovery. One can argue that even the best automated systems have yet to ``discover'' despite being incredibly useful as laboratory assistants. We must carefully consider how they have been and can be applied to future problems of chemical discovery in order to effectively design and interact with future autonomous platforms. The majority of this article defines a large set of open research directions, including improving our ability to work with complex data, build empirical models, automate both physical and computational experiments for validation, select experiments, and evaluate whether we are making progress toward the ultimate goal of autonomous discovery. Addressing these practical and methodological challenges will greatly advance the extent to which autonomous systems can make meaningful discoveries.

arXiv commentRevised version available at 10.1002/anie.201909989

References in corpus (38)

• Automatic chemical design using a data-driven continuous representation of molecules
• On representing chemical environments
• Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning
• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• Molecular Graph Convolutions: Moving Beyond Fingerprints
• AFLOW: An automatic framework for high-throughput materials discovery
• Explainable Machine Learning for Scientific Insights and Discoveries
• Big Data of Materials Science - Critical Role of the Descriptor
• GuacaMol: Benchmarking Models for De Novo Molecular Design
• Junction Tree Variational Autoencoder for Molecular Graph Generation
• Protein-Ligand Scoring with Convolutional Neural Networks
• Informed Machine Learning -- A Taxonomy and Survey of Integrating Knowledge into Learning Systems
• Deep learning for molecular design - a review of the state of the art
• Data-driven materials science: status, challenges and perspectives
• Machine learning prediction errors better than DFT accuracy
• Machine Learning of Molecular Electronic Properties in Chemical Compound Space
• VAMPnets: Deep learning of molecular kinetics
• SchNet: A continuous-filter convolutional neural network for modeling quantum interactions
• MolGAN: An implicit generative model for small molecular graphs
• Learning Deep Generative Models of Graphs
• Massively Multitask Networks for Drug Discovery
• Deep Learning: A Critical Appraisal
• Accelerating Materials Development via Automation, Machine Learning, and High-Performance Computing
• Resolving transition metal chemical space: feature selection for machine learning and structure-property relationships
• Electronic Spectra from TDDFT and Machine Learning in Chemical Space
• Atom2Vec: learning atoms for materials discovery
• Improving Chemical Autoencoder Latent Space and Molecular De novo Generation Diversity with Heteroencoders
• Chemception: A Deep Neural Network with Minimal Chemistry Knowledge Matches the Performance of Expert-developed QSAR/QSPR Models
• Grammar Variational Autoencoder
• Deep Confidence: A Computationally Efficient Framework for Calculating Reliable Errors for Deep Neural Networks
• Generative Models for Automatic Chemical Design
• Graph convolutional neural networks as "general-purpose" property predictors: the universality and limits of applicability
• Error-Controlled Exploration of Chemical Reaction Networks with Gaussian Processes
• SMILES2Vec: An Interpretable General-Purpose Deep Neural Network for Predicting Chemical Properties
• Latent Molecular Optimization for Targeted Therapeutic Design
• A Model to Search for Synthesizable Molecules
• Fully Integrated On-FPGA Molecular Dynamics Simulations
• A Coupon-Collector Model of Machine-Aided Discovery

Cited by in corpus (19)

• Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems
• Autonomous discovery in the chemical sciences part I: Progress
• Gryffin: An algorithm for Bayesian optimization of categorical variables informed by expert knowledge
• Bayesian optimization with known experimental and design constraints for chemistry applications
• Inverse methods for design of soft materials
• Optimizing Molecules using Efficient Queries from Property Evaluations
• Quantum Machine Learning Framework for Virtual Screening in Drug Discovery: a Prospective Quantum Advantage
• Olympus: a benchmarking framework for noisy optimization and experiment planning
• We Should at Least Be Able to Design Molecules That Dock Well
• Learning To Navigate The Synthetically Accessible Chemical Space Using Reinforcement Learning
• Golem: An algorithm for robust experiment and process optimization
• Predicting Gas-Particle Partitioning Coefficients of Atmospheric Molecules with Machine Learning
• Amortized Tree Generation for Bottom-up Synthesis Planning and Synthesizable Molecular Design
• Spiers Memorial Lecture: How to do impactful research in artificial intelligence for chemistry and materials science
• Deep Learning in Protein Structural Modeling and Design
• Autonomous Drug Design with Multi-Armed Bandits
• An In-depth Summary of Recent Artificial Intelligence Applications in Drug Design
• Gemini: Dynamic Bias Correction for Autonomous Experimentation and Molecular Simulation
• Toward Building Science Discovery Machines