http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Fine-grained attention mechanism for neural machine translation
Choi, Heeyoul,Cho, Kyunghyun,Bengio, Yoshua Elsevier 2018 Neurocomputing Vol.284 No.-
<P><B>Abstract</B></P> <P>Neural machine translation (NMT) has been a new paradigm in machine translation, and the attention mechanism has become the dominant approach with the state-of-the-art records in many language pairs. While there are variants of the attention mechanism, all of them use only temporal attention where one scalar value is assigned to one context vector corresponding to a source word. In this paper, we propose a fine-grained (or 2D) attention mechanism where each dimension of a context vector will receive a separate attention score. In experiments with the task of En-De and En-Fi translation, the fine-grained attention method improves the translation quality in terms of BLEU score. In addition, our alignment analysis reveals how the fine-grained attention mechanism exploits the internal structure of context vectors.</P>
Context-dependent word representation for neural machine translation
Choi, Heeyoul,Cho, Kyunghyun,Bengio, Yoshua Elsevier 2017 Computer speech & language Vol.45 No.-
<P><B>Abstract</B></P> <P>We first observe a potential weakness of continuous vector representations of symbols in neural machine translation. That is, the continuous vector representation, or a word embedding vector, of a symbol encodes multiple dimensions of similarity, equivalent to encoding more than one meaning of the word. This has the consequence that the encoder and decoder recurrent networks in neural machine translation need to spend substantial amount of their capacity in disambiguating source and target words based on the context which is defined by a source sentence. Based on this observation, in this paper we propose to contextualize the word embedding vectors using a nonlinear bag-of-words representation of the source sentence. Additionally, we propose to represent special tokens (such as numbers, proper nouns and acronyms) with typed symbols to facilitate translating those words that are not well-suited to be translated via continuous vectors. The experiments on En–Fr and En–De reveal that the proposed approaches of contextualization and symbolization improves the translation quality of neural machine translation systems significantly.</P>
Pet Shop Recommendation System based on Implicit Feedback
Heeyoul Choi(최희열),Yunhee Kang(강윤희),Myungju Kang(강명주) 한국디지털콘텐츠학회 2017 한국디지털콘텐츠학회논문지 Vol.18 No.8
Due to the advances in machine learning and artificial intelligence technologies, many new services have become available. Among such services, recommendation systems have already been successfully applied to commercial services and made profits as in online shopping malls. Most recommendation algorithms in commercial services are based on content analysis or explicit feedback rates as in movie recommendations. However, many online shopping malls have difficulties in content analysis or are lacking explicit feedbacks on their items, which results in no recommendation system for their items. Even for such service systems, user log data is easily available, and if recommendations are possible with such log data, the quality of their service can be improved. In this paper, we extract implicit feedback like click information for items from log data and provide a recommendation system based on the implicit feedback. The proposed system is applied to a real in-service online shopping mall.
Parameter learning for alpha integration.
Choi, Heeyoul,Choi, Seungjin,Choe, Yoonsuck MIT Press 2013 Neural computation Vol.25 No.6
<P>In pattern recognition, data integration is an important issue, and when properly done, it can lead to improved performance. Also, data integration can be used to help model and understand multimodal processing in the brain. Amari proposed α-integration as a principled way of blending multiple positive measures (e.g., stochastic models in the form of probability distributions), enabling an optimal integration in the sense of minimizing the α-divergence. It also encompasses existing integration methods as its special case, for example, a weighted average and an exponential mixture. The parameter α determines integration characteristics, and the weight vector w assigns the degree of importance to each measure. In most work, however, α and w are given in advance rather than learned. In this letter, we present a parameter learning algorithm for learning α and ω from data when multiple integrated target values are available. Numerical experiments on synthetic as well as real-world data demonstrate the effectiveness of the proposed method.</P>
Estimating Resident Registration Numbers of Individuals in Korea: Revisited
( Heeyoul Kim ),( Ki-woong Park ),( Daeseon Choi ),( Younho Lee ) 한국인터넷정보학회 2018 KSII Transactions on Internet and Information Syst Vol.12 No.6
Choi et al's work [1] in 2015 demonstrated that the resident registration numbers (RRNs) of individuals could be conveniently estimated through their personal information that is ordinarily disclosed in social network services. As a follow-up to the study, we introduce the status of the RRN system in Korea in terms of its use in the online environment, particularly focusing on their secure use. We demonstrate that it is still vulnerable against a straightforward attack. We establish that we can determine the RRNs of the current president Moon Jae-In and the world-class singer PSY.
Heeyoul Choi,Yunhee Kang 보안공학연구지원센터 2016 International Journal of u- and e- Service, Scienc Vol.9 No.1
Most analyses in pedagogy have been based on surveys, while in many other research areas like cognitive science and psychology, data-driven research has made significant progress based on large-scale data automatically generated and archived. Recently in pedagogy, learning achievement data has been archived, and EduData is one of such data sets provided by Korean ministry of education. Many data driven analysis algorithms can be applied to such data. As a first data-driven analysis to EduData, we applied the linear regression model to check which factors are effective to Korean student’s learning achievement. Finally, we proposed a model to predict degree of achievement. Experimental results show the performance of our models.