AIF-C01 Question 113
Select 2A data science team is working on a recommendation system that involves storing and querying vector embeddings generated by a machine learning model. The team needs to store these embeddings in a database optimized for vector similarity searches and ensure scalability for their growing dataset. Which AWS services should they consider for storing and querying these embeddings?
- A
Amazon OpenSearch Service
- B
Amazon Aurora
- C
Amazon Neptune
- D
Amazon DocumentDB (with MongoDB compatibility)
- E
Amazon RDS for PostgreSQL
Show answer and explanation
Correct answers: A, C
Explanation
When working with vector embeddings, it is critical to use a database optimized for vector similarity searches. Amazon OpenSearch Service is designed for full-text search and recently incorporated vector search capabilities, making it ideal for this purpose. Additionally, Amazon Neptune supports graph-based use cases and can handle embeddings effectively when combined with graph algorithms. Other services like Aurora, DocumentDB, and RDS for PostgreSQL are not optimized for vector search and thus are not suitable for this specific requirement.
- A. Correct.
Amazon OpenSearch Service supports vector search capabilities, making it suitable for storing and querying embeddings efficiently.
- B. Incorrect.
Amazon Aurora is a relational database that is not optimized for vector similarity searches. It is better suited for structured data and transactional use cases.
- C. Correct.
Amazon Neptune is a graph database that supports graph-based queries and vector embeddings, making it a valid option for storing and querying embeddings.
- D. Incorrect.
Amazon DocumentDB is a document database compatible with MongoDB but does not natively support vector similarity search, making it less ideal for this use case.
- E. Incorrect.
Amazon RDS for PostgreSQL is a relational database that does not have native support for vector similarity searches, making it unsuitable for this scenario.