In this lab, you will learn about semantic search and how a knowledge graph in Neo4j, combined with text embeddings from Amazon Bedrock, can be leveraged for fast and effective knowledge retrieval. We will add text documents from 10-K filings to our graph to demonstrate these capabilities.
Step through these two notebooks in order to complete the lab. Run them in Azure ML as we did before.
- 01-text-embedding.ipynb - Use Bedrock Titan text embedding service on 10-K filings.
- 02-semantic-search.ipynb - Demonstrates knowledge retrieval with connected data (multiple hops from documents) using vector search and graph traversals in Cypher.