What is it about?
Advances in hardware and data-driven methods have led to new materials being discovered with just a computer. One such development is the release of the online polymer informatics platform called Polymer Genome. Polymer Genome predicts the properties of a polymer based on its chemical formula or a drawing of the polymer. The tool uses machine learning to first learn the relationship between known polymers and their properties. It then extends this correlation to predict the properties of different structures. This paper explains how the tool works. This includes an explanation of how data is generated and read by the computer models, the details of the algorithms, and instances of how the tool can be used to test and develop polymers.
Photo by FLY:D on Unsplash
Why is it important?
Polymers are a diverse group of materials that range from simple molecules to complex long chains. The wide range of properties makes it difficult to find the right polymers for specific applications. In this regard, the tool provides a good starting point to test the suitability of polymers. As a machine-learning-based method, it is faster than conventional simulations. The platform can be accessed by simply visiting a website and it can be used to compare and test the properties of polymers almost instantly. KEY TAKEAWAY: Polymer Genome is a convenient tool to predict the properties of polymers. This paper provides an in-depth description of how the tool works. Its ease of access and its quick predictions make it a useful tool for the development of polymers.
Read the Original
This page is a summary of: Machine-learning predictions of polymer properties with Polymer Genome, Journal of Applied Physics, November 2020, American Institute of Physics, DOI: 10.1063/5.0023759.
You can read the full text:
Be the first to contribute to this page