Uploaded image for project: 'Compass '
  1. Compass
  2. COMPASS-7606

Compass: Vector Search Hybrid Scoring

    • Type: Icon: Story Story
    • Resolution: Unresolved
    • Priority: Icon: Major - P3 Major - P3
    • None
    • Affects Version/s: None
    • Component/s: None
    • Developer Tools

      Epic Summary

      Summary

      Support result scoring based on combined results obtained from multiple search, vector-search and possibly non-search pipelines. 

      Allow users to be able to use well known functions to combine scores (like listed here) or to define custom score composition functions.

      Motivation

      Users that have multiple search indexes (e.g. a regular search index and vector search index) may want to retrieve documents based on a score that depends on both vector distance and match quality to a regular search query.

      We want to provide a reasonably efficient and convenient way of obtaining a combining results from multiple pipelines that return documents with their respective scores. 

      Documentation

      Product Description
      Scope
      Technical Design
      Docs Update

            Assignee:
            Unassigned Unassigned
            Reporter:
            sara.akram@mongodb.com Sara Akram (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

              Created:
              Updated: