2021/01/05 by Vineeth Venugopal, Venugopal, Vineeth, Sourav Sahoo +9 · 4 citations
Materials Science · Computer Science · #Machine Learning in Materials Science #X-ray Diffraction in Crystallography #Geochemistry and Geologic Mapping
paper · pdf · doi:10.48550/arxiv.2101.01508
Most of the knowledge in materials science literature is in the form of\nunstructured data such as text and images. Here, we present a framework\nemploying natural language processing, which automates text and image\ncomprehension and precision knowledge extraction from inorganic glasses'\nliterature. The abstracts are automatically categorized using latent Dirichlet\nallocation (LDA), providing a way to classify and search semantically linked\npublications. Similarly, a comprehensive summary of images and plots are\npresented using the 'Caption Cluster Plot' (CCP), which provides direct access\nto the images buried in the papers. Finally, we combine the LDA and CCP with\nthe chemical elements occurring in the manuscript to present an 'Elemental\nmap', a topical and image-wise distribution of chemical elements in the\nliterature. Overall, the framework presented here can be a generic and powerful\ntool to extract and disseminate material-specific information on\ncomposition-structure-processing-property dataspaces, allowing insights into\nfundamental problems relevant to the materials science community and\naccelerated materials discovery.\n