README.md

CKBQA

Update June 2nd, 2021: add BibTex.

Update May 6th, 2021: our paper will appear at NAACL 2021 System Demonstration track, please check it out for more details on data annotations and evaluations.

Introduction

This repository contains a Chinese KBQA dataset expanded from CCKS CKBQA Competition Dataset. We also provide a novel framework based on these additional annotations in a paper, an online demo of such KBQA system is available, have fun checking it out!

Citation & Trouble Shooting

Please contact us via {zhangminhao, ry_zhang}@pku.edu.cn if you have any questions or kindly cite the following paper if you find this resource helpful.

@inproceedings{zhang-etal-2021-namer,
    title = "{NAMER}: A Node-Based Multitasking Framework for Multi-Hop Knowledge Base Question Answering",
    author = "Zhang, Minhao  and
      Zhang, Ruoyu  and
      Zou, Lei  and
      Lin, Yinnian  and
      Hu, Sen",
    booktitle = "Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies: Demonstrations",
    month = jun,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/2021.naacl-demos.3",
    pages = "18--25",
}

Data Format

The validation dataset and training dataset are provided in this repository in json format. Each entry contains a question, its question id, its corresponding gold SPARQL, and the struct of the SPARQL. We define variable, entity, literal and type in SPARQL as node. The struct shows triples' relations, the head and tail nodes of triples and their mentions, including start and end offsets. Besides, filter information is also contained in the struct.

Here are two examples:

SPARQL: select ?x where { <微软> <主要软件产品> ?x. }

Struct: { "selected_variable": { "node": "?x", "type": "variable", "mention": { "start_offset": 7, "end_offset": 11, "natural_language": "软件产品" } }, "triple": [ { "head": { "node": "<微软>", "type": "entity", "mention": { "start_offset": 0, "end_offset": 4, "natural_language": "微软公司" } }, "relation": "<主要软件产品>", "tail": { "node": "?x", "type": "variable", "mention": { "start_offset": 7, "end_offset": 11, "natural_language": "软件产品" } } } ], "filter": [] }

SPARQL: select ?y where { <最后的晚餐_（达·芬奇画作）> <作者> ?x. ?x <职业> ?z. ?y <职业> ?z. filter(?y!=?x). }

Struct: { "selected_variable": { "node": "?y", "type": "variable", "mention": { "start_offset": 17, "end_offset": 18, "natural_language": "人" } }, "triple": [ { "head": { "node": "<最后的晚餐_（达·芬奇画作）>", "type": "entity", "mention": { "start_offset": 1, "end_offset": 8, "natural_language": "《最后的晚餐》" } }, "relation": "<作者>", "tail": { "node": "?x", "type": "variable", "mention": { "start_offset": 9, "end_offset": 11, "natural_language": "作者" } } }, { "head": { "node": "?x", "type": "variable", "mention": { "start_offset": 9, "end_offset": 11, "natural_language": "作者" } }, "relation": "<职业>", "tail": { "node": "?z", "type": "variable", "mention": { "start_offset": 14, "end_offset": 16, "natural_language": "职业" } } }, { "head": { "node": "?y", "type": "variable", "mention": { "start_offset": 17, "end_offset": 18, "natural_language": "人" } }, "relation": "<职业>", "tail": { "node": "?z", "type": "variable", "mention": { "start_offset": 14, "end_offset": 16, "natural_language": "职业" } } } ], "filter": [ "filter(?y!=?x)" ] }