Powered by Zoomin Software. For more details please contactZoomin

The Semaphore Fact Extraction Framework (FACTS)

Complex fact strategies

  • Last Updated: May 29, 2026
  • 2 minute read
    • Semaphore
    • Documentation

A complex fact strategy is used when a document fact returns more than one fact name, arranged in a grouping or hierarchy.

Key Considerations for Fact Extraction

  1. Is the fact locatable using a zone?
    Is it a phrase range, defined by either a zoner (such as a date, person, etc.), a text pattern, or a matching label (including stemming, casing, etc.)?
  2. Are there several instances of such zones in the content?
    There may be zones that are not considered facts.
  3. Are there several instances of the fact to return?
    There may be more than one fact to be returned.

Decision Tree

It’s a complex fact—there is more than one value being returned.

Can that complex fact be identified in and of itself—that is, the facts that make up the complex fact, and how they present in content, are enough to identify the complex fact in the content?           

    YES: Might there be multiple complex facts?     

        YES: For each complex fact, is each in its own context instance?    

            YES: Then we can use any fact context and any anchor context

            NO: Then we can use any fact context we want but only anchor contexts that provide a window

        NO: Then we can use any context we want for the fact context and for the anchor context 

    NO: Might there be multiple complex facts?      

        YES: For each complex fact, is there just one that can appear in its own context instance?  

            YES: Then we can use any fact context and any anchor context

            NO: Then we can use any fact context we want but only anchor contexts that provide a window

        NO: Then we can use any context we want
TitleResults for “How to create a CRG?”Also Available inAlert