Is Domain Knowledge Necessary for Machine Learning Materials Properties?
| Autor: | Ryan J. Murdock, Taylor D. Sparks, Anthony Yu-Tung Wang, Steven K. Kauwe |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Artificial neural network
Computer science business.industry Materials informatics Machine learning computer.software_genre Industrial and Manufacturing Engineering Field (computer science) Simple (abstract algebra) Informatics Encoding (memory) Metallic materials Domain knowledge General Materials Science Artificial intelligence business computer |
| Zdroj: | Integrating Materials and Manufacturing Innovation. 9:221-227 |
| ISSN: | 2193-9772 2193-9764 |
| Popis: | New featurization schemes for describing materials as composition vectors in order to predict their properties using machine learning are common in the field of Materials Informatics. However, little is known about the comparative efficacy of these methods. This work sets out to make clear which featurization methods should be used across various circumstances. Our findings include, surprisingly, that simple fractional and random-noise representations of elements can be as effective as traditional and new descriptors when using large amounts of data. However, in the absence of large datasets or for data that is not fully representative, we show that the integration of domain knowledge offers advantages in predictive ability. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink