arXiv · 2022-10-18
Making a MIRACL: Multilingual Information Retrieval Across a Continuum of Languages
Abrir ficha de evidencia
Pregunta de investigación
How can a retrieval benchmark cover languages with very different resource levels while using native queries and relevance judgments?
Método
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.
Qué respalda la evidencia
Evaluate each language explicitly, document corpus and judgment construction, and keep native review central. The reported hybrid result is a historical benchmark observation.
Límites y transferibilidad
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.
Clasificación
- Respaldado
- Histórico
- Solo académico
Influencia en NAVINES: Laboratorio de comparación de páginas multilingües