Reference
Foreign Language Translation AI for Intel glossary
Terms used in Foreign Language Translation AI for Intel and its documentation.
Applies to Foreign Language Translation AI for Intel as it works today. Moraph updates the platform continuously, and this page is reviewed when it changes.
Terms used in Foreign Language Translation AI for Intel and its documentation.
| Term | Meaning |
|---|---|
| Machine output | A transcript or translation produced by a model and not yet reviewed by a linguist. |
| Gisted | A translation a linguist has checked for overall meaning but not reviewed line by line. |
| Reviewed | A translation a linguist has corrected and checked in full. |
| Certified | A translation a senior linguist has approved for use in reporting. |
| Watchlist | A list of keywords, names, places and topics maintained by analysts and matched in both languages. |
| Transliteration variant | Another spelling of a name or term in a different script or romanization. |
| Tone indicator | A hint that a passage has notable tone or sentiment, never a finding on its own. |
| Mission glossary | Preferred translations of recurring terms, maintained by senior linguists. |
| Confidence score | The model's own estimate of how reliable a transcript segment is. |
| Handling rules | Rules that control who can see and export items based on markings and access attributes. |
Was this page helpful?
Report a problem with this page
Need help with your own setup rather than this page? Raise a support ticket.