• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Expert systems: The international journal of knowledge engineering >Label informed hierarchical transformers for sequential sentence classification in scientific abstracts
【24h】

Label informed hierarchical transformers for sequential sentence classification in scientific abstracts

机译：Label informed hierarchical transformers for sequential sentence classification in scientific abstracts

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相关主题

摘要

Segmenting scientific abstracts into discourse categories like background, objective,method, result, and conclusion is useful in many downstream tasks like search, recommendationand summarization. This task of classifying each sentence in theabstract into one of a given set of discourse categories is called sequential sentenceclassification. Existing machine learning-based approaches to this problem considerthe content of only the abstract to obtain the neural representation of each sentence,which is then labelled with a discourse category. But this ignores the semantic informationoffered by the discourse labels themselves. In this paper, we propose LIHT,Label Informed Hierarchical Transformers – a method for sequential sentence classificationthat explicitly and hierarchically exploits the semantic information in thelabels to learn label-aware neural sentence representations. The hierarchical modelhelps to capture not only the fine-grained interactions between the discourse labelsand the words in the abstract at the sentence level but also the potential dependenciesthat may exist in the label sequence. Thus, LIHT generates label-aware contextualsentence representations that are then labelled with a conditional random field.We evaluate LIHT on three publicly available datasets, namely, PUBMED-RCT,NICTA-PIBOSO and CSAbstract. The incremental gain in F1-score in all the threecases over the respective state-of-the-art approaches is around 1%. Though thegains are modest, LIHT establishes a new performance benchmark for this task and isa novel technique of independent interest. We also perform an ablation study toidentify the contribution of each component of LIHT in the observed performance,and a case study to visualize the roles of the different components of our model.

著录项

获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号