Neo4j Graph Tech Weekly
This Week in Neo4j: Workshops, GPT-3, Graph Data Science Book, Bayesian KG, and More
First, BIG news: We are happy to announce that we’re back on the?Discourse platform ---> Join our community!! The migration is done, and we’ll improve the user experience over the coming weeks. As always, let us know if there is anything we can do better!
In his trilogy of well-structured articles, Fanghua (Joshua) Yu runs demonstrations with GPT-3, creating a knowledge graph of arXiv paper metadata, doing entity and relationship extraction, and generating embeddings of paper titles. He then uses cosine similarity to find the most similar title for a certain search phrase.
Sulstice uses ChatGPT to interface with a chemical database generate Cypher queries to allow users to query molecules.??He feeds the APOC (Awesome Procedures on Cypher) generated schema, nodes, and relationships to ChatGPT, which responds with Cypher queries he then uses in an AuraDB instance.
NODES SESSION:?GraphQL Federation Using the Neo4j GraphQL Library
Darrell Warde talks about one of the latest features they are implementing in the Neo4j GraphQL Library – GraphQL Federation. GraphQL Federation enables you to define “subgraphs†and stitch them together using a gateway, allowing them to be queried as a single GraphQL schema.
NEW BOOK:?Graph Data Science With Neo4j
This latest book from Estelle Scifo covers new features of the Neo4j’s Graph Data Science library, including its handy Python client and the introduction of machine learning pipelines (classification and link prediction) directly in the library. It also contains a new chapter about the Pregel API, the way to go to extend the GDS and implement your own algorithm.
TUTORIAL:?How to Build a Bayesian Knowledge Graph
黄æ€è¡Œ explains how to integrate Bayesian inference into a knowledge graph. The result is a Bayesian knowledge graph – the knowledge graph displays a comprehensive big picture over a certain knowledge domain, while the Bayesian inference computes the conditional probabilities of a causal/correlational network.