arXiv · 2022-10-18
Making a MIRACL: Multilingual Information Retrieval Across a Continuum of Languages
Открыть запись доказательств
Вопрос исследования
How can a retrieval benchmark cover languages with very different resource levels while using native queries and relevance judgments?
Метод
An 18-language, same-language Wikipedia retrieval benchmark built with native-speaker queries and judgments; lexical, dense and hybrid baselines were compared on the released pools.
Что поддерживают данные
Evaluate each language explicitly, document corpus and judgment construction, and keep native review central. The reported hybrid result is a historical benchmark observation.
Ограничения и переносимость
This is same-language information retrieval, not translation assessment or web SEO. Heuristic segmentation, candidate-pool gaps and unfinished test labels limit transferability; it says nothing about hreflang or Google rankings.
Классификация
- Поддержано
- Историческое
- Только академическое
Влияние на NAVINES: Лаборатория сравнения многоязычных страниц