NQiI (Natural Questions In Icelandic)

Natural Questions in Icelandic (NQiI) is a valuable dataset designed for extractive question answering (QA) in the Icelandic language. Let me provide you with some details about this dataset:

1. Purpose and Importance:
2. The NQiI dataset was created to facilitate the development and evaluation of Icelandic QA systems.

3. It also supports the development of QA methods that need to work across a wide range of morphologically and grammatically diverse languages in a multilingual context¹.

4. Dataset Creation:

5. Contributors were asked to come up with questions they were genuinely interested in knowing the answers to.
6. Later, they were tasked with finding answers to each other's questions using a previously published methodology.

7. The questions in NQiI are "natural" in the sense that they arise from genuine curiosity¹.

8. Dataset Details:

9. The complete NQiI dataset contains 18,000 labeled entries.

10. Among these, 5,568 entries are directly suitable for training an extractive QA system specifically for Icelandic¹.

11. Resource and Evaluation:

12. NQiI serves as a valuable resource for Icelandic language research.
13. Researchers have used it to create and evaluate systems capable of extractive QA in Icelandic¹.

(1) Natural Questions in Icelandic - ACL Anthology. https://aclanthology.org/2022.lrec-1.477/. (2) NQiI - Natural Questions In Icelandic - v1.0 - CLARIN. https://repository.clarin.is/repository/xmlui/handle/20.500.12537/143. (3) Natural Questions in Icelandic - ACL Anthology. https://aclanthology.org/2022.lrec-1.477/. (4) Natural Questions in Icelandic - ACL Anthology. https://aclanthology.org/2022.lrec-1.477.pdf. (5) undefined. https://aclanthology.org/2022.lrec-1.477.