• Phonetics and Speech Sciences
• /
• v.7 no.4
• /
• pp.41-47
• /
• 2015

This paper examines whether the Global Vector model is applicable to Korean data as a universal learning algorithm. The main purpose of this study is to compare the global vector model (GloVe) with the word2vec models such as a continuous bag-of-words (CBOW) model and a skip-gram (SG) model. For this purpose, we conducted an experiment by employing an evaluation corpus consisting of 70 target words and 819 pairs of Korean words for word similarities and analogies, respectively. Results of the word similarity task indicated that the Pearson correlation coefficients of 0.3133 as compared with the human judgement in GloVe, 0.2637 in CBOW and 0.2177 in SG. The word analogy task showed that the overall accuracy rate of 67% in semantic and syntactic relations was obtained in GloVe, 66% in CBOW and 57% in SG.

• Annual Conference on Human and Language Technology
• /
• 2015.10a
• /
• pp.165-168
• /
• 2015

기계학습 기반의 자연어처리 모듈에서 중요한 단계 중 하나는 모듈의 입력으로 단어를 표현하는 것이다. 벡터의 사이즈가 크고, 단어 간의 유사성의 개념이 존재하지 않는 One-hot 형태와 대조적으로 유사성을 표현하기 위해서 단어를 벡터로 표현하는 단어 표현 (word representation/embedding) 생성 작업은 자연어 처리 작업의 기계학습 모델의 성능을 개선하고, 몇몇 자연어 처리 분야의 모델에서 성능 향상을 보여 주어 많은 관심을 받고 있다. 본 논문에서는 Word2Vec, CCA, 그리고 GloVe를 사용하여 106,552개의 PubMed의 바이오메디컬 논문의 요약으로 구축된 말뭉치 카테고리의 각 단어 표현 모델의 카테고리 분류 능력을 확인한다. 세부적으로 나눈 카테고리에는 질병의 이름, 질병 증상, 그리고 난소암 마커가 있다. 분류 능력을 확인하기 위해 t-SNE를 이용하여 2차원으로 단어 표현 결과를 맵핑하여 가시화 한다.

• Annual Conference on Human and Language Technology
• /
• 2017.10a
• /
• pp.71-75
• /
• 2017

외래어란 외국어로부터 들어와 한국어에 동화되고 한국어로서 사용되는 언어이다. 나날이 우리의 언어사용 문화에서 외래어의 사용 비율은 높아져가는 추세로, 전문분야에서는 특히 두드러진다. 그러므로 더 효율적이고 효과적인 자연언어처리를 위해서 문서 내 외래어 인식은 중요한 전처리 과정이다. 따라서 본 논문에서는 bidirectional LSTM(이하 bi-LSTM)-CRF 모형의 심층학습을 이용한 음절태깅 기반의 외래어 인식 시스템을 제안한다. 제안하는 시스템의 외래어 인식 학습 과정은 다음과 같다. 첫째, 학습용 말뭉치 자료의 한글 음절들과 공백, 마침표(.)를 토대로 word2vec을 통해 학습용 피쳐(feature) 자료를 생성한다. 둘째, 학습용 말뭉치 자료와 학습용 피쳐 자료를 결합하여 bi-LSTM 모형 학습 자료를 구축한다. 셋째, bi-LSTM 모형을 거쳐 학습된 결과물을 CRF 모형에서 로그 가능도(log likelyhood)와 비터비(Viterbi) 알고리즘을 통해 학습 결과물을 내놓는다. 넷째, 학습용 말뭉치 자료의 정답과 비교한 뒤 모형 내부의 수치들을 조정한다. 다섯째, 학습을 마칠 때까지 반복한다. 본 논문에서 제안하는 시스템을 이용하여 자체적인 뉴스 수집 자료에 대해서 높은 정확도와 재현율을 기록하였다.

• 한국어정보학회:학술대회논문집
• /
• 2017.10a
• /
• pp.71-75
• /
• 2017

외래어란 외국어로부터 들어와 한국어에 동화되고 한국어로서 사용되는 언어이다. 나날이 우리의 언어사용 문화에서 외래어의 사용 비율은 높아져가는 추세로, 전문분야에서는 특히 두드러진다. 그러므로 더 효율적이고 효과적인 자연언어처리를 위해서 문서 내 외래어 인식은 중요한 전처리 과정이다. 따라서 본 논문에서는 bidirectional LSTM(이하 bi-LSTM)-CRF 모형의 심층학습을 이용한 음절태깅 기반의 외래어 인식 시스템을 제안한다. 제안하는 시스템의 외래어 인식 학습 과정은 다음과 같다. 첫째, 학습용 말뭉치 자료의 한글 음절들과 공백, 마침표(.)를 토대로 word2vec을 통해 학습용 피쳐(feature) 자료를 생성한다. 둘째, 학습용 말뭉치 자료와 학습용 피쳐 자료를 결합하여 bi-LSTM 모형 학습 자료를 구축한다. 셋째, bi-LSTM 모형을 거쳐 학습된 결과물을 CRF 모형에서 로그 가능도(log likelyhood)와 비터비(Viterbi) 알고리즘을 통해 학습 결과물을 내놓는다. 넷째, 학습용 말뭉치 자료의 정답과 비교한 뒤 모형 내부의 수치들을 조정한다. 다섯째, 학습을 마칠 때까지 반복한다. 본 논문에서 제안하는 시스템을 이용하여 자체적인 뉴스 수집 자료에 대해서 높은 정확도와 재현율을 기록하였다.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.15 no.4
• /
• pp.103-110
• /
• 2019

A patent contains detailed information of the developed technology and is published to the public. Thus, patents can be used to overcome the limitations of traditional technology trend research and prediction techniques. Recently, due to the advantages of patented analytical methodology, IP R&D is carried out worldwide. The patent is big data and has a huge amount, various domains, and structured and unstructured data characteristics. For this reason, there are many difficulties in collecting and researching patent information. Patent research generally writes the Search formula to collect patent documents from DB. The collected patent documents contain some noise patents that are irrelevant to the purpose of analysis, so they are removed. However, eliminating noise patents is a manual task of reading and classifying technology, which is time consuming and expensive. In this study, we propose a model that automatically classifies The Noise patent for efficient patent information research. The proposed method performs Patent Embedding using Word2Vec and generates Noise seed label. In addition, noise patent classification is performed using the Random forest. The experimental data is published and registered with the USPTO among the patents related to Ocean Surveillance & Tracking Network technology. As a result of experimenting with the proposed model, it showed 73% accuracy with the label actually given by experts.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.441-443
• /
• 2021

Sentence or text similarity is a measure of the degree of similarity between two sentences. Techniques for measuring text similarity include Jacquard similarity, cosine similarity, Euclidean similarity, and Manhattan similarity. Currently, the cosine similarity technique is most often used, but since this is an analysis according to the occurrence or frequency of a word in a sentence, the analysis on the semantic relationship is insufficient. Therefore, we try to improve the efficiency of analysis on the similarity of sentences by giving relations between words using ontology and including semantic similarity when extracting words that are commonly included in two sentences.

• The Journal of Information Systems
• /
• v.27 no.4
• /
• pp.71-96
• /
• 2018

Purpose Recently, social interest in the convergence research is at its peak. However, contrary to the keen interest in convergence research, an infrastructure that makes it easier to recruit researchers from other fields is not yet well established, which is why researchers are having considerable difficulty in carrying out real convergence research. In this study, we implemented a researcher recommendation system that helps researchers who want to collaborate easily recruit researchers from other fields, and we expect it to serve as a springboard for growth in the convergence research field. Design/methodology/approach In this study, we implemented a system that recommends proper researchers when users enter keyword in the field of research that they want to collaborate using word embedding techniques, word2vec. In addition, we also implemented function of keyword suggestions by using keywords drawn from LDA Topicmodeling Algorithm. Finally, the UI of the researcher recommendation system was completed by utilizing the collaborative messenger Slack to facilitate immediate exchange of information with the recommended researchers and to accommodate various applications for collaboration. Findings In this study, we validated the completed researcher recommendation system by ensuring that the list of researchers recommended by entering a specific keyword is accurate and that words learned as a similar word with a particular researcher match the researcher's field of research. The results showed 85.89% accuracy in the former, and in the latter case, mostly, the words drawn as similar words were found to match the researcher's field of research, leading to excellent performance of the researcher recommendation system.

• Journal of Korea Multimedia Society
• /
• v.21 no.9
• /
• pp.1035-1043
• /
• 2018

There is a case where the core content of another person's work is decorated as though it is his own thoughts by changing own thoughts without showing the source. Plagiarism test of copykiller free service used in plagiarism check is performed by comparing plagiarism more than 6th word. However, it is not enough to judge it as a plagiarism with a six - word match if it is replaced with a similar word. Therefore, in this paper, we construct word clusters by using DBSCAN algorithm, find synonyms, convert the words in the clusters into representative synonyms, and construct L-R tables through L-R parsing. We then propose a method for determining the similarity of documents by applying weights to the thesaurus and weights for each paragraph of the thesis.

• Journal of The Korean Association of Information Education
• /
• v.25 no.5
• /
• pp.691-704
• /
• 2021

The development of AI technology has brought about a big change in our lives. The importance of AI and data education is also growing as AI's influence from life to society to the economy grows. In response, the OECD Education Research Report and various domestic information and curriculum studies deal with data literacy and present it as an essential competency. However, the definition of data literacy and the content and scope of the components vary among researchers. Thus, we analyze the semantic similarity of words through Word2Vec deep learning natural language processing methods along with the definitions of key data literacy studies and analysis of word frequency utilized in components, to present objective and comprehensive definition and components. It was revised and supplemented by expert review, and we defined data literacy as the 'basic ability of knowledge construction and communication to collect, analyze, and use data and process it as information for problem solving'. Furthermore we propose the components of each category of knowledge, skills, values and attitudes. We hope that the definition and components of data literacy derived from this study will serve as a good foundation for the systematization and education research of AI education related to students' future competency.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.1
• /
• pp.254-276
• /
• 2019

Since the amount of information on the internet is growing rapidly, it is not easy for a user to find relevant information for his/her query. To tackle this issue, the researchers are paying much attention to Document Summarization. The key point in any successful document summarizer is a good document representation. The traditional approaches based on word overlapping mostly fail to produce that kind of representation. Word embedding has shown good performance allowing words to match on a semantic level. Naively concatenating word embeddings makes common words dominant which in turn diminish the representation quality. In this paper, we employ word embeddings to improve the weighting schemes for calculating the Latent Semantic Analysis input matrix. Two embedding-based weighting schemes are proposed and then combined to calculate the values of this matrix. They are modified versions of the augment weight and the entropy frequency that combine the strength of traditional weighting schemes and word embedding. The proposed approach is evaluated on three English datasets, DUC 2002, DUC 2004 and Multilingual 2015 Single-document Summarization. Experimental results on the three datasets show that the proposed model achieved competitive performance compared to the state-of-the-art leading to a conclusion that it provides a better document representation and a better document summary as a result.