Materials Informatics Study Note

This account aims to share my study about materials informatics. My challenge is reading three papers and tweet their summaries everyday! Let’s study MI!!

[Materials Informatics Study Note @mat_info_study]

A big data study on #MI! Billions of material data was used to train a model. The balance of atoms in the composition are considered to remove a bias. Big data driven new material search may come soon!!

https://t.co/t9eaGngRA9

[Materials Informatics Study Note @mat_info_study]

#Matminer is a great tool for the data mining for #MI.

https://t.co/avCS4vPYvf

[Materials Informatics Study Note @mat_info_study]

A great #review for the #MI studies in #catalyst. The discussion is very primitive but extremely fruitful. This review spent three pages to explain what is “#data”. All the researchers to study catalyst with MI should be started from this great review.

https://t.co/yB9vF0IYAe

[Materials Informatics Study Note @mat_info_study]

The introduction for #ChemML, a library for machine learning on #chemistry. I found the functionality of the ChemML is rich while explanation about how to use it is poor, unfortunately. Or maybe just I could not find it.

https://t.co/QVDeW31ocA

https://t.co/l1KiotZOM8

[Materials Informatics Study Note @mat_info_study]

This is a brief #review of #MI mainly discuss the material screening for Li-ion battery. The force-field (FF) approach was introduced as a screening technology of materials with higher efficiency than DFT. However, the accuracy of DFT is better than FF. https://t.co/H8ThaYalRR

[Materials Informatics Study Note @mat_info_study]

A study to screen #materials for #solarcells. Various materials, including #perovskites, was investigated by first principle calculations. The theoretical evaluation was checked by experiments. The study was well done and comprehensive.

https://t.co/f60NLRKLNg

[Materials Informatics Study Note @mat_info_study]

The comprehensive review on #materialsinformatics! I read this review since it is one of the most cited paper in #MI. I found the reason on its quality. I can recommend to all the new comer in #MI like me.

https://t.co/n3wmRU4zLX

[Materials Informatics Study Note @mat_info_study]

A great study applying #MI procedure to spintronics. The easy determination of the magnetic properties of materials is one of dreams for #spintronics researchers, which is realized by #ML technology as demonstrated. I expect the #ML model is opened.

https://t.co/YtimIPqBKZ

[Materials Informatics Study Note @mat_info_study]

Today, I read the #ML part of this review paper. Again, it is well written and comprehensive! I knew the basic process, especially how to write codes, of the #ML but did not know its mathematical features. This paper gives their intuitive understanding :)

https://t.co/n3wmRU4zLX

[Materials Informatics Study Note @mat_info_study]

An #MI study considering #topology on halide #perovskites materials. The #ML model considering #persistenthomology and Ricci curvature gives better accuracy compared to conventional models. This study may open new field on #MI in #topologicalmaterials.

https://t.co/beOMNFjGXS

[Materials Informatics Study Note @mat_info_study]

An excellent #review about #MI! I have read the chapters about #DFT and I believe this is one of the best review!! DFT is very complicated and difficult to understand. But this review well summarize at least the points of DFT clearly.
#AI #OpenAccess

https://t.co/n3wmRU4zLX

[Materials Informatics Study Note @mat_info_study]

A study predicting #bandgap of materials by #ML. The result indicates that the bandgap is the most strongly depends on the magnetic moment… Why? This is very strange result for me. Various #ML models were used for the analysis.
#AI #MI

https://t.co/zL0VTeAsXV

[Materials Informatics Study Note @mat_info_study]

In this video, we can feel the history of the research in #MI through the 20 years experience of the senior researcher. The most impressive is the relatively slow progress in computer material science and the break through by the material big database.

https://t.co/7qArbEWuYo

[Materials Informatics Study Note @mat_info_study]

This study changes my basic understanding of #ML. I thought ML is useful for interpolation of data, while this study clearly reveals the great potential of ML by for finding new and great materials by finding anomalies with #autoencoder :)

https://t.co/RZbWFnidof

[Materials Informatics Study Note @mat_info_study]

This online lecture tells the development of catalyst with first principle calculation and #ML. Moreover, the dynamics of molecules near a surface is discussed and the open catalyst project is introduced.

https://t.co/fF6ITsdu0l

[Materials Informatics Study Note @mat_info_study]

The first chapter of this book tells the brief history of the development of material database. It was realized thanks to the great collaborations of physicists and engineers, and was the first step for the current excitations of #MI.

https://t.co/0x0SU6ICVG

[Materials Informatics Study Note @mat_info_study]

An advanced study of #MI using graphical neural network (#GNN) with self-#attention scheme. For me, very beginner of MI, quite tough to understand, but I could feel the great potential of the #Finder, introduced in this study.
#MAchineLearning
https://t.co/BWqzDtuVcj

[Materials Informatics Study Note @mat_info_study]

A great study about the thermal properties of inorganic oxide materials. 30,000 of data, obtained by text-mining of published papers, was analyzed systematically with #ML. #tamman’s rule was reevaluated and arranged with big data.

https://t.co/qLJWsLzY4M

[Materials Informatics Study Note @mat_info_study]

A great lecture by #materialsproject about #MI for the development of #catalyst. This lecture gives a wide perspective in the current stage of the field. The #opencatalystproject will be the great opportunity to study the cutting edge of the technology.

https://t.co/fF6ITsdu0l

[Materials Informatics Study Note @mat_info_study]

I wonder if we use box for desktop PC, the system can be much cheaper.

[Materials Informatics Study Note @mat_info_study]

Development of an #opensource hardware for low-cost, high-precision, and high-throughput experiments is reported! For #MI, uniform standardization of experiments are critical. This may be started by making uniform standard of experimental equipments.

https://t.co/YBIFTuLGUl