knowledge-graph-learning
knowledge-graph-learning copied to clipboard
NAACL-2019-FreebaseQA: A New Factoid QA Data Set Matching Trivia-Style Question-Answer Pairs with Freebase
Summary:
这是篇介绍新数据集的论文。数据集叫FreebaseQA,是基于Freebase knowledge graph的开放式QA数据集。不适合用于某个特定domain。
Resource:
Paper information:
- Author: Department of Electrical Engineering and Computer Science York University
- Dataset:
- keywords:
Notes:
数据集的生成方式:利用Freebase里的subject-predicate-object triples来生成QA pair。然后对问题中的每个entities标注tag,然后在Freebase里寻找和标注的tag相关的predicates。最后使用人工去除triples中的false positives。
但看是github里issue的留言,这个数据集还有很多不足的地方。
Model Graph:
Result::
Thoughts:
Next Reading: