DSpace Repository

Semantically Readable Distributed Representation Learning and Its Expandability Using a Word Semantic Vector Dictionary

Show simple item record

dc.contributor.author KESHI, Ikuo en
dc.contributor.author SUZUKI, Yu en
dc.contributor.author YOSHINO, Koichiro en
dc.contributor.author NAKAMURA, Satoshi en
dc.date.accessioned 2018-04-11T01:48:50Z en
dc.date.available 2018-04-11T01:48:50Z en
dc.date.issued 2018-04-01 en
dc.identifier.issn 0916-8532 en
dc.identifier.uri http://hdl.handle.net/10061/12334 en
dc.description.abstract The problem with distributed representations generated by neural networks is that the meaning of the features is difficult to understand. We propose a new method that gives a specific meaning to each node of a hidden layer by introducing a manually created word semantic vector dictionary into the initial weights and by using paragraph vector models. We conducted experiments to test the hypotheses using a single domain benchmark for Japanese Twitter sentiment analysis and then evaluated the expandability of the method using a diverse and large-scale benchmark. Moreover, we tested the domain-independence of the method using a Wikipedia corpus. Our experimental results demonstrated that the learned vector is better than the performance of the existing paragraph vector in the evaluation of the Twitter sentiment analysis task using the single domain benchmark. Also, we determined the readability of document embeddings, which means distributed representations of documents, in a user test. The definition of readability in this paper is that people can understand the meaning of large weighted features of distributed representations. A total of 52.4% of the top five weighted hidden nodes were related to tweets where one of the paragraph vector models learned the document embeddings. For the expandability evaluation of the method, we improved the dictionary based on the results of the hypothesis test and examined the relationship of the readability of learned word vectors and the task accuracy of Twitter sentiment analysis using the diverse and large-scale benchmark. We also conducted a word similarity task using the Wikipedia corpus to test the domain-independence of the method. We found the expandability results of the method are better than or comparable to the performance of the paragraph vector. Also, the objective and subjective evaluation support each hidden node maintaining a specific meaning. Thus, the proposed method succeeded in improving readability. en
dc.language.iso en en
dc.publisher The Institute of Electronics, Information and Communication Engineers en
dc.rights Copyright c 2018 The Institute of Electronics, Information and Communication Engineers en
dc.subject distributed representation en
dc.subject word semantic vector dictionary en
dc.subject paragraph vector en
dc.subject word2vec en
dc.subject Twitter en
dc.subject sentiment analysis en
dc.title Semantically Readable Distributed Representation Learning and Its Expandability Using a Word Semantic Vector Dictionary en
dc.type.nii Journal Article en
dc.contributor.transcription ケシ, イクオ ja
dc.contributor.transcription スズキ, ユウ ja
dc.contributor.transcription ヨシノ, コウイチロウ ja
dc.contributor.transcription ナカムラ, サトシ ja
dc.contributor.alternative 芥子, 育雄 ja
dc.contributor.alternative 鈴木, 優 ja
dc.contributor.alternative 吉野 , 幸一郎 ja
dc.contributor.alternative 中村, 哲 ja
dc.textversion publisher en
dc.identifier.jtitle IEICE Transactions on Information and Systems en
dc.identifier.volume E101.D en
dc.identifier.issue 4 en
dc.identifier.spage 1066 en
dc.identifier.epage 1078 en
dc.relation.doi 10.1587/transinf.2017DAP0019 en
dc.identifier.NAIST-ID 84369982 en
dc.identifier.NAIST-ID 74651399 en
dc.identifier.NAIST-ID 74651712 en
dc.identifier.NAIST-ID 73296626 en


Files in this item

This item appears in the following Collection(s)

Show simple item record

Search DSpace


Advanced Search

Browse

My Account