Carbon Figures of Merit Knowledge Creation with a Hybrid Solution and Carbon Tables API
This work addresses the need for faster knowledge extraction in materials science for carbon capture, but it appears incremental as it builds on existing methods for data processing and APIs.
The paper tackles the problem of accelerating the discovery of materials for CO2 absorption/adsorption by developing an asynchronous REST API called Carbon Tables, which creates knowledge from tables in scientific PDFs and stores it in knowledge graphs using a hybrid heuristic and machine learning approach, enabling sophisticated search and further knowledge creation.
Nowadays there are algorithms, methods, and platforms that are being created to accelerate the discovery of materials that are able to absorb or adsorb $CO_2$ molecules that are in the atmosphere or during the combustion in power plants, for instance. In this work an asynchronous REST API is described to accelerate the creation of Carbon figures of merit knowledge, called Carbon Tables, because the knowledge is created from tables in scientific PDF documents and stored in knowledge graphs. The figures of merit knowledge creation solution uses a hybrid approach, in which heuristics and machine learning are part of. As a result, one can search the knowledge with mature and sophisticated cognitive tools, and create more with regards to Carbon figures of merit.