2020 LDP Conference on Linked Data

A Machine Learning Approach for Classifying Sinopia's RDF

https://ld4p.github.io/classify-rdf-2020/
BackgroundSinopia's Classifying RDF ChallengeResource TemplatesPanda's DataFrameColab Notebook DemoNext Steps

Background

Linked Data for Production (LD4P)

Original funded by the Mellon Foundation, the Linked Data for Production is currently on it's third grant round for funding the implementation of linked data systems for libraries, museums, and other cultural heritage organizations.

Sinopia, a cloud-based linked-data editor, was the primary product of the second LD4P grant cycle that started in 2018 and ended in 2020. The third LD4P grant, that runs from 2020-2022 focuses on user-interface improvements in Sinopia as well as new backend processes for integrating Sinopia with third-parties like OCLC and ShareVDE.

PyTorch and FastAI

After initial experiments using Google's TensorFlow, an open source machine learning framework originally developed by Google, we switched to using the FastAI machine learning libraries that is built on the open-source PyTorch framework. This switch was driven in part by Jeremy taking the excellent Practical Deep Learning for Coder, v3 course offered by FastAI.