| About this Abstract |
| Meeting |
MS&T24: Materials Science & Technology
|
| Symposium
|
Standards for Data Science in Additive Manufacturing
|
| Presentation Title |
Scientific Data FAIRification and Dynamic Knowledge Infrastructure to Drive AI |
| Author(s) |
Balashanmuga Priyan Rajamohan, Alexander Harding Bradley, Thomas Ciardi, Arafath Nihar, Laura Bruckman, Yinghui Wu, Erika Barcelos, Roger French |
| On-Site Speaker (Planned) |
Balashanmuga Priyan Rajamohan |
| Abstract Scope |
There is a significant disparity between the petabyte-sized experimental data generated in scientific laboratories and the potential to enhance this data for reproducibility and interoperability. This gap hinders the full utilization of data-centric AI initiatives, slowing scientific innovation as more time is dedicated to understanding, organizing, and maintaining the data. Additionally, it's not uncommon to see metadata discarded, even when it could be essential for providing context and semantic understanding. We introduce a Python package designed to simplify the conversion of CSV data into Findable, Accessible, Interoperable, and Reproducible(FAIR) Ontologies and JSON-LDs. By leveraging these ontologies, data from heterogeneous sources can be seamlessly integrated, creating an ever-evolving knowledge base. Furthermore, we present a robust Knowledge Infrastructure to support data-intensive graphical computation, interactive visualizations, and advanced deep learning methods, such as st-GNNs. |