NEWS · MODELS · #1291
Google Translate pairs Gemini models with human linguists to improve conversational translations
Google says it is pairing its latest Gemini models with human linguistic experts on the Google Translate team to prioritize conversational meaning over literal phrasing and to better support under-represented, data-scarce languages; the article profiles three team members (Janneke, Isaac, Harsh) and notes Translate now supports nearly 250 languages.
KEY POINTS
- Google says it is pairing its latest Gemini models with human linguistic experts on the Google Translate team to prioritize conversational meaning over literal phrasing and to better support under-represented, data-scarce languages; the article profiles three team members (Janneke, Isaac, Harsh) and notes Translate now supports nearly 250 languages.
- Combining large language models with human expertise indicates Google's approach to make AI translations more context-aware and to include data-scarce languages.
- Celebrating International Translation Day: Meet 3 linguistic experts behind Google Translate
WHY IT MATTERS
Combining large language models with human expertise indicates Google's approach to make AI translations more context-aware and to include data-scarce languages.