Databricks Certified Generative AI Engineer - Associate
Question: 11
A company wants to build a system where users can input natural language questions, and the system retrieves relevant information from a document repository, then generates a natural language answer. The system should use a retriever component to search the document repository and a generator component to produce answers in natural language based on the retrieved documents. Which combination of components would best fit this requirement?
Real Exam
Community
LLeetQuiz
Explanation:
Explanation
B. Dense passage retriever followed by a generative model (e.g., GPT-based)
This combination is ideal for building a system that retrieves relevant information from a document repository and generates natural language answers. The dense passage retriever uses dense embeddings to search and retrieve the most relevant documents, and the generative model produces fluent, context-aware natural language responses based on those documents.
Why this works best:
- Dense passage retriever: Uses semantic similarity through embeddings to find documents that are contextually relevant to the user's question, not just keyword matches
- Generative model (e.g., GPT-based): Can synthesize information from multiple retrieved documents and generate coherent, natural language answers that directly address the user's query
Why other options are less suitable:
- A: NER followed by summarization doesn't provide effective document retrieval and lacks the generative capabilities for question-answering
- C: Text classification followed by generation doesn't handle the document retrieval aspect effectively
- D: Keyword-based search is less effective than semantic search for understanding natural language questions