Hi ,
I’m working on a multilingual chatbot using Vertex AI Agent Builder and its Search data store feature.
Does the Agent Builder data store support cross-lingual semantic search (e.g., English queries matching German documents)?
What are the best practices for building a multilingual semantic search agent in Agent Builder?
Hi @avineet ,
Welcome to Google Cloud Community!
Currently, Vertex AI Agent Builder’s Search data store does not directly support cross-lingual semantic search.
Building cross-lingual semantic search capabilities requires clear text preparation, choosing the right model, training the data, testing, and tuning for accurate results. You can try implementing or integrating the Cloud Translation API as a translation layer through keyword matching, either by translating the user query or the documents into a common language before performing the semantic search. However, please note that this is a direct translation and does not directly provide a measure of similarity between texts in different languages. Alternatively, you can also consider using robust multilingual embeddings, such as Text Embedding, which creates embeddings of text that capture its semantic meaning. These embeddings are designed to represent the meaning of a piece of text, allowing you to find texts that are semantically similar to your query, even if they do not share the same keywords.
Was this helpful? If so, please accept this answer as “Solution”. If you need additional assistance, reply here within 2 business days and I’ll be happy to help.