Название: Accelerated Materials Discovery: How to Use Artificial Intelligence to Speed Up Development
Автор: Philip De Luna
Издательство: De Gruyter
Серия: De Gruyter Stem
Год: 2022
Страниц: 215
Язык: английский
Формат: pdf (true)
Размер: 123.4 MB

Typical timelines to go from discovery to impact in the advanced materials sector are between 10 to 30 years. Advances in robotics and Artificial Intelligence (AI) are poised to accelerate the discovery and development of new materials dramatically. This book is a primer for any materials scientist looking to future-proof their careers and get ahead of the disruption that Artificial Intelligence and robotic automation is just starting to unleash. It is meant to be an overview of how we can use these disruptive technologies to augment and supercharge our abilities to discover new materials that will solve world's biggest challenges. Written by world leading experts on accelerated materials discovery from academia (UC Berkeley, Caltech, UBC, Cornell, etc.), industry (Toyota Research Institute, Citrine Informatics) and national labs (National Research Council of Canada, Lawrence Berkeley National Labs).

There have been three critical advances in technology over the past few decades that have revolutionized the way we discover new things. First, computational power has exploded. Today we carry more computing power in the cell phones in our pockets than in the rockets that brought man to the moon. This computing power has allowed humans to transcend biological limits of brain matter to tackle ever increasingly complex and wicked problems. Second, data is everywhere. We live in a data-rich world that is only getting richer. Advances in the way we collect, analyze, and store data has led to an exponential growth of information. Data is the new gold, and there’s a rush by technology companies looking to capitalize and extract full value from this untapped resource. Third, robots are better, much better.

Right now, you can go online and purchase a robot dog from Boston Dynamics, one that can perform routine maintenance on oil rigs or explore mining caverns that are unsafe. Robotics has never been cheaper and more accessible as today, and as economies of scale mature, this is only going to drive adoption and innovation. Taken together, these enabling technologies are accelerating the way scientists discover new things – from drugs and chemicals to batteries and consumer electronics. Automation and Artificial Intelligence (AI) are being applied to augment and improve traditional discovery. These efficiency gains are critical to saving humanity’s most precious resource – time. In a world challenged by a global pandemic and a warming climate, we cannot afford the one or two decades for the traditional cycle of discovery to impact. We need to act now.

This book is a primer for any materials scientist looking to future-proof their careers and get ahead of the disruption that Artificial Intelligence and automation is just starting to unleash. It is meant to be an overview of how we can use these digital technologies to augment and supercharge our abilities to discover new materials. Our enticing offer is twofold – speed and reproducibility.

The first chapter will provide an introductory overview to accelerated materials discovery. It discusses the basic concepts and levels of automation, the role of Machine Learning and Artificial Intelligence, the importance of data, the experimental workflow, and what the laboratory of the future will look like.

The second chapter describes how artificial intelligence is being used to model materials chemistry with catalysis for clean energy conversion as the application goal. It provides basic concepts of chemical reactivity and how we can describe these properties digitally and in a machine-readable way. It then shows how applying machine learning to computational models of different catalysts can help us discover new ways to convert CO2 into fuels, make hydrogen, or transform methanol.

The third chapter tackles how we can use artificial intelligence to help us better experimentally measure, characterize, and probe the properties of materials. Spectroscopy investigates materials through electromagnetic stimulus, x-rays or light. The amount of data generated by spectroscopy measurements are massive, and the analysis is complex. By using Artificial Intelligence, we can speed up the analysis and find complex correlations in the sea of spectroscopy data. This chapter tells us how.

The fourth chapter covers the concept of self-driving laboratories. Autonomous labs that plan, conduct, and draw conclusions from experiments with minimal human input. Robotic labs have been science fiction for many years, but now the advances in robotics, AI, and data are allowing these autonomous researchers to come to life.

The fifth chapter dives deep into what makes artificial intelligence intelligent – the complex math and algorithms that give AI its power. This chapter will cover the various types of algorithms used in materials discovery including neural networks, natural language processing, and unsupervised learning. It describes the strengths and pitfalls of each and what remains to be done to improve them.

The sixth chapter is written from the perspective of how accelerated materials discovery can be used effectively in industry. Authored by a leading scientist in one of the world’s foremost materials informatics startups, this chapter ties all the concepts together and provides case studies for how and why companies should use accelerated materials discovery to remain competitive.