AOAI Search - Identifying Anomalies in Search Results

Manojkumar Ganesamoorthy 0 Reputation points Microsoft Employee
2024-06-26T15:50:27.83+00:00

Created a Hybrid Index (Vector + Keyword) and uploaded a few documents in AI Studio.

Using AI Studio to chat (Q&A, Summarize etc.) with the documents.

Ask : What is MSFT official guidance/recommendation to surface anomalies or inconsistencies in content/text that’s returned from the search results.

Example

Document 1 – Containing a reference to storage of 10GB

Document 2 – Containing a reference to storage of 15GB

Azure AI Search
Azure AI Search
An Azure search service with built-in artificial intelligence capabilities that enrich information to help identify and explore relevant content at scale.
1,049 questions
Azure OpenAI Service
Azure OpenAI Service
An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
3,188 questions
{count} votes

1 answer

Sort by: Most helpful
  1. Azar 22,870 Reputation points MVP
    2024-06-26T17:37:50.3733333+00:00

    Hi there Manojkumar Ganesamoorthy

    Thats a good question and thanks for using QandA platform

    for this i guess yyou can use AI and machine learning techniques to analyze and compare the content of returned documents. In your case, after creating a Hybrid Index (Vector + Keyword) and uploading documents in AI Studio, you can use vector embeddings to find semantically similar documents and highlight differences in specific attributes (e.g., storage size).

    or Use keyword-based queries to identify documents containing specific terms and compare these terms across documents.

    you can also write custom scripts to parse and compare the content of documents, focusing on attributes like storage size.

    Example:

    • Document 1 mentions 10GB of storage.
    • Document 2 mentions 15GB of storage.

    By running these analyses, you can surface inconsistencies and flag them for review.

    If this helps kindly accept the answer thanks much.


Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.