Teaching Nemotron Greek: Mining a Corpus, Adapting Retrieval, and Grounding Generation for Modern Greek across Specialist Domains
Building AI that can find and use Greek-language information accurately
Researchers adapted NVIDIA's Nemotron AI system to work with Modern Greek, a language previously missing from major AI tools. After training on 65,773 Greek examples, their retrieval model's performance jumped from 0.362 to 0.835 on a standard measure, and their question-answering system improved from 29.4% to 66.9% correct answers—with much better accuracy in citing sources.
Greek professionals in law, finance, energy, and medicine can now use AI systems that actually understand their language and documents instead of relying on generic multilingual tools that perform poorly on specialist Greek text. The researchers released their adapted models and a new Greek benchmark publicly, so other teams can build better Greek-language AI applications without starting from scratch.