Welcome to the new version of CaltechAUTHORS. Login is currently restricted to library staff. If you notice any issues, please email coda@library.caltech.edu
Published November 5, 2020 | public
Journal Article

The Journal of Physical Chemistry A/B/C Virtual Special Issue on Machine Learning in Physical Chemistry

Abstract

Physical chemistry stands today at an exciting transition state where the integration of machine learning and data science tools into all corners of the field stands poised to do nothing short of revolutionizing the discipline. These powerful techniques—when appropriately combined with domain knowledge, tools, and expertise—have led to new physical insights, better understanding, accelerated discovery, rational design, and inverse engineering that transcend traditional approaches to materials, molecular, and chemical science and engineering. The primary driver of this trend has been the impressive advances enabled by machine learning, artificial intelligence, and data science tools, ranging from the discovery of novel electronic and optical materials by high-throughput virtual screening, to the massive acceleration of molecular simulations using learned classical force fields with quantum accuracy, to the powering of "self-driving laboratories" for automated chemical discovery. The 2011 White House Materials Genome Initiative (MGI), the 2017 NSF Data-Driven Discovery Science in Chemistry (D3SC) initiative, and the 2019 NSF Big Idea Harnessing the Data Revolution are some of the US federal programs that have provided incentive, attention, momentum, and support to power these advances and help drive the field forward. Necessity is also the mother of invention, and the prevalence of large data sets routinely generated by high-throughput virtual screening or automated experimentation have spurred the need for scalable data science and machine learning techniques to parse, explore, and harness the full power of these voluminous data streams. It bears remembering that physical chemistry is no stranger to machine learning, most visibly in the cheminformatics and quantitative structure property relation (QSPR) work that emerged in the 1980s. Some of the techniques being implemented today are, to some degree, reinventions of these ideas, but others are fundamentally new concepts that have been adopted and adapted from diverse fields including computer vision, manifold learning, and deep learning. This Virtual Special Issue on Machine Learning in Physical Chemistry covering all sections of The Journal of Physical Chemistry A/B/C pays tribute to this development, and the relevance and popularity of this topic is reflected in the depth and breadth of excellent articles in this exciting collection.

Additional Information

© 2020 American Chemical Society. Published as part of The Journal of Physical Chemistry virtual special issue "Machine Learning in Physical Chemistry". This Preface is published jointly in The Journal of Physical Chemistry A/B/C.

Additional details

Created:
August 20, 2023
Modified:
October 20, 2023