alandria/rubq

A Russian data set for question answering over Wikidata

RuBQ: A Russian Knowledge Base Question Answering and Machine Reading Comprehension Data Set

RuBQ 1.0

In May 2020 we presented the first version of RuBQ (pronounced [`rubik]) -- Russian Knowledge Base Questions, a KBQA dataset that consists of 1,500 Russian questions of varying complexity along with their English machine translations, corresponding SPARQL queries, answers, as well as a subset of Wikidata covering entities with Russian labels. To the best of our knowledge, this is the first Russian KBQA and semantic parsing dataset. The dataset is thought to be used as a development and test sets in cross-lingual transfer, few-shot learning, or learning with synthetic data scenarios.

RuBQ 2.0

In December 2020 we built the second version of RuBQ. The dataset extension is based on questions obtained through search engine query suggestion services. The dataset doubled in size: RuBQ 2.0 contains 2,910 questions along with the answers and SPARQL queries. We also expanded the dataset with machine reading comprehension capabilities: RuBQ 2.0 incorporates answer-bearing paragraphs from Wikipedia for the majority of questions. Thus, the dataset is now not only suitable for the evaluation of KBQA, but also can be used to evaluate machine reading comprehension, paragraph retrieval, and end-to-end open-domain question answering. The dataset can be also used for experiments in hybrid QA, where KBQA and text-based QA can enrich and complement each other.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.