DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Kim, Kee-Eung | - |
dc.contributor.advisor | 김기응 | - |
dc.contributor.author | Jang, Youngsoo | - |
dc.date.accessioned | 2019-09-04T02:45:56Z | - |
dc.date.available | 2019-09-04T02:45:56Z | - |
dc.date.issued | 2018 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=734083&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/267006 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 전산학부, 2018.2,[iv, 22 p. :] | - |
dc.description.abstract | The dialog state tracking, which refers to identifying the user' | - |
dc.description.abstract | s intent from the utterances, is one of the most important tasks in dialog management. In this paper, we present our dialog state tracker developed for the fifth dialog state tracking challenge (DSTC5), which focused on cross-language adaptation using machine-translated training data, which was very scarce compared to the size of the ontology. Our dialog state tracker is based on the bi-directional long short-term memory (LSTM) with a hierarchical attention mechanism in order to spot important words in user utterances. The intent prediction is done by finding the most relevant keyword in the ontology to the attention-weighted word vector. Our tracker thus has various advantages over existing approaches, such as predicting out-of-vocabulary (OOV) intent due to scarce training data or machine translation. We show that our tracker outperforms other trackers submitted to the challenge in terms of almost all performance measures. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | Dialog state tracking▼aAttention mechanism▼aHierarchical attention mechanism▼aCross-language | - |
dc.subject | 대화 상태 추적▼a어텐션 메커니즘▼a계층적 어텐션 메커니즘▼a교차 언어 | - |
dc.title | Cross-language dialog state tracking using hierarchical attention mechanism | - |
dc.title.alternative | 계층적 어텐션 메커니즘을 이용한 교차 언어 대화 상태 추적 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :전산학부, | - |
dc.contributor.alternativeauthor | 장영수 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.