Rapid annotation through human-machine collaboration
HLT '02 Proceedings of the second international conference on Human Language Technology Research(2002)
摘要
This paper addresses the problem of efficiently obtaining training data for a new entity type or relation. We describe a methodology for rapidly obtaining annotation by using seed examples and human feedback, and we show that this method allows annotation to be performed approximately 20 times faster than manual annotation alone, with small degradation in annotation accuracy.
更多查看译文
关键词
training data,small degradation,human feedback,human-machine collaboration,new entity type,annotation accuracy,seed example,rapid annotation,manual annotation
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络