graphs in vdb

README.md

@@ -430,7 +430,7 @@ Recommended Viewing:
 Traditional databases are optimized for storing and querying structured data, such as text, numbers, and dates. They are not designed to handle vector data efficiently. Vector databases, on the other hand, are specifically designed to store and query vector data. They use specialized indexing techniques and algorithms to enable fast and accurate similarity search such as quantization and clustering of vectors.
 
 ### How does a vector database work?
-A vector database usually contains indexes of vectors, these indexes contain matrices of vector embeddings, ordered in such a way that they can be queried efficiently. When a query is made, either text or a vector embedding is provided as input, in the case of text, it is embedded, and the vector database will query the appropriate index to retrieve the most similar vectors based on distance metrics. Usually, the vectors are compared using metrics such as cosine similarity, dot product, or Euclidean distance. Vectors also relate to a dictionary of metadata that could contain information such as the document ID, the document title, the corresponding text and more.
+A vector database usually contains indexes of vectors, these indexes contain matrices of vector embeddings, often a graph data structure is used as well, ordered in such a way that they can be queried efficiently. When a query is made, either text or a vector embedding is provided as input, in the case of text, it is embedded, and the vector database will query the appropriate index to retrieve the most similar vectors based on distance metrics. Usually, the vectors are compared using metrics such as cosine similarity, dot product, or Euclidean distance. Vectors also relate to a dictionary of metadata that could contain information such as the document ID, the document title, the corresponding text and more.
 
 ### What are the different strategies for searching in a vector database or index?
 Search strategies in vector databases include: