What is materials informatics?
Materials Informatics is an amalgamation of the field of materials science and informatics, in order to aid the use, selection, development, and discovery of materials. 
In other words, “materials informatics”, is a novel data-driven technology which uses the fundamentals and experimental data of materials combined with the aid of advanced statistical models to predict materials of the future. 
This has been an emerging field since the early 2000s and is set to grow exponentially with higher computational capabilities and better documentation of experimental data.
Introduction to materials informatics
Materials have always been the heart of product design. It is not surprising that the tech-hub of the world, Silicon Valley, derives its name from an element which is the still the backbone of many technologies. Materials determine the form and function of products. Manufacturing superior materials will increase the possibility for new technologies, with applications disrupting industries ranging from automobiles to consumer electronics and sectors including energy and healthcare.
It took the ancient people of India centuries of persistent tinkering to get hold of the alloying elements to add to iron to prevent it from rusting. They used only their intuition, experience and creativity and just a tad bit of ‘steel science,’ which, interestingly, is a field generated through their experimentation. With technology advancing rapidly, the time for innovation is now restricted to decades. 
There are two major bottlenecks in the fast-tracking of the materials discovery. The first is the discovery of the material itself and the second is the translation of technology from a lab-scale to large-scale. The industrial capability to most quickly and efficiently develop and deploy advanced materials is therefore critical to a globally competitive manufacturing sector.
It is only in the last decade or so, thanks to the implicit and explicit acceptance of the importance of data, that “materials informatics” paradigms are rapidly becoming an essential part of the materials research portfolio.
This change in notion was brought about mainly by the Materials Genome Initiative and partly by the algorithmic developments which were capable of handling small data-sets and the ability of models to predict the future entirely based on the past.
I agree with most of the people that data-driven technologies encompass machine learning, artificial intelligence, and deep learning algorithms. However, in my opinion, even though each of these has minor differences theoretically, they drastically impact different aspects of the materials sector. All the three are essential and must work in unison to fast-track materials growth worldwide. I will bring out this aspect through the remaining articles in this blog series.
Taking a cue from the famous axiom, “Time is money”, in the first blog of the series, I will focus on the financial and technological importance of implementing data-driven technologies to hasten materials prediction.
The economic significance can be derived from the ‘Potential Impact on Risk’ of the project failing due to fundamental flaws in the material architecture, the ‘Potential Impact on Time to Market’ and the ‘Potential Impact on Relative costs’ per project per year.
The Materials Genome Initiative estimates the value of these additional benefits to be roughly 2 to 3 times the value of potential R&D efficiency impacts. Altogether, the potential economic benefit of an improved Materials Innovation Infrastructure is estimated to be worth between $123 billion and $270 billion per year. 
It is incredibly challenging to implement data-driven technology. A core component of this technology is “data”. The materials research world has always closely held onto this component, as that has usually been their USP. In today’s world, the factor of “time” holds more value. It is essential for the technology to be strong; however, if it reaches the market too late the opportunity is lost.
The best example is the perovskite solar cells which perform exceedingly well but its discovery was too late and now it fails to compete with the established silicon solar cell market.
Accepting the above and establishing more collaborative networks for data acquisition and sharing would help technology.
Additionally, establishing reliability in products has been extremely tough. This was witnessed by renowned mobile phones catching fire due to a fundamental flaw in the battery design. Scaling up and assuring strong quality control on each material product is difficult for a human but strong algorithms can spot the defects and prevent disasters.
With a strong proof of concept, machine learning approaches have recently provided a replacement for direct experimentation or computations/simulations in which fundamental equations are explicitly solved. More importantly, these methods are becoming very useful to determine materials properties that are difficult to compute or measure using traditional approaches, due to the cost, time or effort involved, but for which valid data either exists or can be generated for at least a fraction of the cost. 
I believe machine learning will not eliminate empiricism. Problems will only get more complex and additional data will be required, which in the materials domain is mostly obtained through experimentation. Materials informatics will simply aid human intuition and improve the design of experiments by eliminating the noise factors and speed up groundbreaking innovations.
With a freakishly large amount of work being done in this domain, the data being generated by this field is a growing at a pace only a machine learning algorithm can handle.
 Mulholland, Gregory J., and Sean P. Paradiso. “Perspective: Materials informatics across the product lifecycle: Selection, manufacturing, and certification.” APL Materials 4.5 (2016): 053207.
 “Materials Informatics Science and Technology of Advanced Materials.” Hurricane Katrina: A Geographical Perspective | Explore Taylor & Francis Online, 2017.
 Rahnama, Alireza, Sam Clark, and Seetharaman Sridhar. “Machine learning for predicting occurrence of interphase precipitation in HSLA steels.” Computational Materials Science 154 (2018): 169-177.
 Scott, T., Walsh, A., Anderson, B. and O’Connor, A. (2018). Economic Analysis of National Needs for Technology Infrastructure to Support the Materials Genome Initiative. [online] RTI International, pp.ES1-ES8.
 Mullin, Rick. “C&EN’s 2017’s 10 Startups to Watch.” CEN RSS, American Chemical Society, 2017.
 Singh, Seema. “Cousins of Artificial Intelligence – Towards Data Science.” Towards Data Science, Towards Data Science, 27 May 2018.
 Formatting Credit: Google Slides
 Engefaz. “RCMP – Reliability.” EGF Engefaz | Engenharia e Manutenção Preditiva, 25 Jan. 2016.
*This article is the work of the guest author shown above. The guest author is solely responsible for the accuracy and the legality of their content. The content of the article and the views expressed therein are solely those of this author and do not reflect the views of Matmatch or of any present or past employers, academic institutions, professional societies, or organizations the author is currently or was previously affiliated with.