首站-论文投稿智能助手
典型文献
Biomedical entity linking based on less labeled data
文献摘要:
1 Introduction Biomedical entity linking aims to align the natural language entity to the knowledge base concept referring to the same real-world object.Recent solutions in the biomedical field focus on the embedding-based method to jointly model the texts and the knowledge graph into a multi-dimensional entity space.However,the task of a biomedical entity linking is still facing tough challenges:1)The ambiguity of natural language descriptions,including polysemy and abbreviation,is an urgent problem to be solved.2)Biomedical texts lack detailed annotations.The weakly labeled data brings a huge challenge to entity linking.Although some heuristic methods have been used for semi-structured data analysis in the field of bioin-formatics,it is still very difficult to automatically integrate biomedical literature and structured knowledge graph with a reliable algorithm.
文献关键词:
作者姓名:
Yu HU;Derong SHEN;Tiezheng NIE;Yue KOU;Ge YU
作者机构:
School of Computer Science and Engineering,Northeastern University,Shenyang 110819,China
文献出处:
引用格式:
[1]Yu HU;Derong SHEN;Tiezheng NIE;Yue KOU;Ge YU-.Biomedical entity linking based on less labeled data)[J].计算机科学前沿,2022(03):209-211
A类:
Biomedical,polysemy,formatics
B类:
entity,linking,less,labeled,data,Introduction,aims,align,natural,language,knowledge,concept,referring,same,real,world,object,Recent,solutions,biomedical,field,focus,embedding,jointly,model,texts,graph,into,multi,dimensional,space,However,task,still,facing,tough,challenges,ambiguity,descriptions,including,abbreviation,urgent,problem,solved,lack,detailed,annotations,weakly,brings,huge,Although,some,heuristic,methods,have,been,used,semi,structured,analysis,bioin,very,difficult,automatically,integrate,literature,reliable,algorithm
AB值:
0.559402
相似文献
机标中图分类号,由域田数据科技根据网络公开资料自动分析生成,仅供学习研究参考。