Copyright (c) 2026 MindMesh Academy. All rights reserved. This content is proprietary and may not be reproduced or distributed without permission.

5.3.5. Synonyms Configuration

  • Concept: Define equivalent terms for better matching
  • Purpose: Handle vocabulary variations
  • Benefit: Improved recall without duplicate entries
When adding synonyms via API:
  • ✅ Can be added in any order
  • ✅ Requires at least one QA pair first
  • ❌ Special characters not allowed (#, @, ~, etc.)
// This will cause an error - special characters not allowed
{
  "synonyms": [
    {
      "alterations": [
        "fix problems",
        "troubleshoot",
        "#diagnostic"  // ❌ Will cause error
      ]
    }
  ]
}

Solution: Remove special characters from the API call.

Key Trade-Offs:
  • CLU vs. Question Answering: CLU handles complex intents but requires training; QA handles FAQs with minimal setup
  • Precision vs. Recall: Threshold adjustments trade one for the other

Reflection Question: Your CLU model has high precision (95%) but low recall (70%) for one intent. What might cause this, and how would you address it?