From Data Science to AI via Graph Analytics
A Talk by George Cushen , Paco Nathan , Axel-Cyrille Ngonga Ngomo and Brad Rees
About this Talk
Until recently, few people had much experience putting graphs to work — not beyond university homework based on Dijkstra’s algorithm or calculating centrality. Imagination filled gaps where direct experience was rare.
Today, however, graph applications are becoming more commonplace. They don’t require the enormous scale of social networks. Many interesting graph use cases fit within memory, obviating the need for databases.
While SPARQL has good applications, using graphs doesn’t imply the entirety of the Semantic Web. With deep learning, graph embedding models go further toward abstraction, recognizing patterns in context without the original graph data anywhere nearby.
In this panel, we’ll embark on a journey from data science to AI via graph analytics. Some of the questions we will address:
How can metadata help with data governance and lineage? Where does one look for metadata standards, and how can they be used?
What are some common types of graph analytics and algorithms, and what are they good for?
What are graph embeddings, and what are they good for? How does one get started using them in the real world?
Panel moderated by Paco Nathan
About The Speakers
Founder / Author, Derwen.ai / O'Reilly
Machine learning and knowledge graphs are a dynamic duo, and few people have more experience in this than Paco Nathan. Paco explores the best ways to mix Data Science and Machine Learning with Knowledge Graphs.
Professor of Data Science, Head of DICE Research Group, Paderborn University
Axel leads the DICE research group at the university of Paderborn. His research is focused on machine learning, knowledge graphs, and big data.