实体-关系联合抽取指从非结构化文本中联合抽取出实体-关系三元组,是信息抽取和知识图谱构建的一项关键任务。文中提出了一种新的基于全局指针网络实体关系联合抽取方法BGPNRE(BERT-based Global Pointer Network for Named Entity-Rela...实体-关系联合抽取指从非结构化文本中联合抽取出实体-关系三元组,是信息抽取和知识图谱构建的一项关键任务。文中提出了一种新的基于全局指针网络实体关系联合抽取方法BGPNRE(BERT-based Global Pointer Network for Named Entity-Relation Joint Extraction),首先通过潜在关系预测模块预测文本中蕴含的关系,过滤掉不可能存在的关系,将实体抽取限制在预测的关系子集中;其次通过使用基于关系的全局指针网络,获取所有主客体实体的位置;最后通过全局指针网络通信模块,将主客体位置高效率地解码对齐成一个实体关系三元组。该方法避免了传统管道式方法存在的错误传播问题,同时也解决了关系冗余、实体重叠、Span提取泛化不足等问题。实验结果表明,所提方法在多关系和重叠实体抽取上表现卓越,并且在NYT和WebNLG公共数据集上达到了最先进的水平。展开更多
Aiming at the relation linking task for question answering over knowledge base,especially the multi relation linking task for complex questions,a relation linking approach based on the multi-attention recurrent neural...Aiming at the relation linking task for question answering over knowledge base,especially the multi relation linking task for complex questions,a relation linking approach based on the multi-attention recurrent neural network(RNN)model is proposed,which works for both simple and complex questions.First,the vector representations of questions are learned by the bidirectional long short-term memory(Bi-LSTM)model at the word and character levels,and named entities in questions are labeled by the conditional random field(CRF)model.Candidate entities are generated based on a dictionary,the disambiguation of candidate entities is realized based on predefined rules,and named entities mentioned in questions are linked to entities in knowledge base.Next,questions are classified into simple or complex questions by the machine learning method.Starting from the identified entities,for simple questions,one-hop relations are collected in the knowledge base as candidate relations;for complex questions,two-hop relations are collected as candidates.Finally,the multi-attention Bi-LSTM model is used to encode questions and candidate relations,compare their similarity,and return the candidate relation with the highest similarity as the result of relation linking.It is worth noting that the Bi-LSTM model with one attentions is adopted for simple questions,and the Bi-LSTM model with two attentions is adopted for complex questions.The experimental results show that,based on the effective entity linking method,the Bi-LSTM model with the attention mechanism improves the relation linking effectiveness of both simple and complex questions,which outperforms the existing relation linking methods based on graph algorithm or linguistics understanding.展开更多
Electronic Commerce or E-Commerce is the concept of making trade in electronic environment which has emerged by the use of the internet in commerce and increased after 1995. Organizations started to use electronic com...Electronic Commerce or E-Commerce is the concept of making trade in electronic environment which has emerged by the use of the internet in commerce and increased after 1995. Organizations started to use electronic commerce by internet facilities nearby their traditional marketing methods to increase their sales and profits. E-Commerce has many advantages when compared to traditional trade. This paper aims to take out these advantages for the companies which want to make use of electronic commerce and includes important advices for these companies. The paper also includes the basic version of an e-commerce business, by analysing one of the most successful firms in Turkey, yemeksepeti.com in terms of an Entity-Relationship Diagram （ERD） and Event Driven Process （EDP） to explain the organizations how an e-commerce business model should be constructed systematically. An ERD is a data modeling technique which graphically illustrates an information system＇s entities and the relationships between those entities. On the other hand, EDP is a chain for the modeling language which is used for describing business processes and workflows of an organization. This study examines yemeksepeti.com for providing an example of these two diagrams, which should be a milestone for the newcomers of an e-commerce. Organizing an e-commerce should start with the software designs and these examples of ERD and EDP which are the base for the software of an e-commerce business. The limitations of the paper include the examination of an only one company which operates only in Turkey, thus the further research should examine a global company and/or an e-commerce business that operates in several industries.展开更多
文摘实体-关系联合抽取指从非结构化文本中联合抽取出实体-关系三元组,是信息抽取和知识图谱构建的一项关键任务。文中提出了一种新的基于全局指针网络实体关系联合抽取方法BGPNRE(BERT-based Global Pointer Network for Named Entity-Relation Joint Extraction),首先通过潜在关系预测模块预测文本中蕴含的关系,过滤掉不可能存在的关系,将实体抽取限制在预测的关系子集中;其次通过使用基于关系的全局指针网络,获取所有主客体实体的位置;最后通过全局指针网络通信模块,将主客体位置高效率地解码对齐成一个实体关系三元组。该方法避免了传统管道式方法存在的错误传播问题,同时也解决了关系冗余、实体重叠、Span提取泛化不足等问题。实验结果表明,所提方法在多关系和重叠实体抽取上表现卓越,并且在NYT和WebNLG公共数据集上达到了最先进的水平。
基金The National Natural Science Foundation of China(No.61502095).
文摘Aiming at the relation linking task for question answering over knowledge base,especially the multi relation linking task for complex questions,a relation linking approach based on the multi-attention recurrent neural network(RNN)model is proposed,which works for both simple and complex questions.First,the vector representations of questions are learned by the bidirectional long short-term memory(Bi-LSTM)model at the word and character levels,and named entities in questions are labeled by the conditional random field(CRF)model.Candidate entities are generated based on a dictionary,the disambiguation of candidate entities is realized based on predefined rules,and named entities mentioned in questions are linked to entities in knowledge base.Next,questions are classified into simple or complex questions by the machine learning method.Starting from the identified entities,for simple questions,one-hop relations are collected in the knowledge base as candidate relations;for complex questions,two-hop relations are collected as candidates.Finally,the multi-attention Bi-LSTM model is used to encode questions and candidate relations,compare their similarity,and return the candidate relation with the highest similarity as the result of relation linking.It is worth noting that the Bi-LSTM model with one attentions is adopted for simple questions,and the Bi-LSTM model with two attentions is adopted for complex questions.The experimental results show that,based on the effective entity linking method,the Bi-LSTM model with the attention mechanism improves the relation linking effectiveness of both simple and complex questions,which outperforms the existing relation linking methods based on graph algorithm or linguistics understanding.
文摘Electronic Commerce or E-Commerce is the concept of making trade in electronic environment which has emerged by the use of the internet in commerce and increased after 1995. Organizations started to use electronic commerce by internet facilities nearby their traditional marketing methods to increase their sales and profits. E-Commerce has many advantages when compared to traditional trade. This paper aims to take out these advantages for the companies which want to make use of electronic commerce and includes important advices for these companies. The paper also includes the basic version of an e-commerce business, by analysing one of the most successful firms in Turkey, yemeksepeti.com in terms of an Entity-Relationship Diagram （ERD） and Event Driven Process （EDP） to explain the organizations how an e-commerce business model should be constructed systematically. An ERD is a data modeling technique which graphically illustrates an information system＇s entities and the relationships between those entities. On the other hand, EDP is a chain for the modeling language which is used for describing business processes and workflows of an organization. This study examines yemeksepeti.com for providing an example of these two diagrams, which should be a milestone for the newcomers of an e-commerce. Organizing an e-commerce should start with the software designs and these examples of ERD and EDP which are the base for the software of an e-commerce business. The limitations of the paper include the examination of an only one company which operates only in Turkey, thus the further research should examine a global company and/or an e-commerce business that operates in several industries.
