• 한국어정보학회:학술대회논문집
• /
• 2016.10a
• /
• pp.20-24
• /
• 2016

Sequence-to-sequence 모델은 입력열을 길이가 다른 출력열로 변환하는 모델로, 단일 신경망 구조만을 사용하는 End-to-end 방식의 모델이다. 본 논문에서는 Sequence-to-sequence 모델을 한국어 구구조 구문 분석에 적용한다. 이를 위해 구구조 구문 트리를 괄호와 구문 태그 및 어절로 이루어진 출력열의 형태로 만들고 어절들을 단일 기호 'XX'로 치환하여 출력 단어 사전의 수를 줄였다. 그리고 최근 기계번역의 성능을 높이기 위해 연구된 Attention mechanism과 Input-feeding을 적용하였다. 실험 결과, 세종말뭉치의 구구조 구문 분석 데이터에 대해 기존의 연구보다 높은 F1 89.03%의 성능을 보였다.

• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.574-579
• /
• 2018

본 논문에서는 한국어 형태소 분석 시스템을 제안하는데, 연구 목표는 오타 없는 문서를 대상으로 한 경우에도 높은 성능을 유지하면서, 동시에 오타가 있는 문서에서도 우수한 성능을 산출하는 것이다. 실험은 크게 두 종류로 나누어서 진행된다. 주 실험인 첫 번째 실험에서는, 자모 임베딩과 음절 임베딩을 결합(concatenate)한 벡터를 입력으로 Bidirectional LSTM CRFs을 수행함으로써, 세종말뭉치 대상으로 어절 정확도 97%, 그리고 1, 2, 5 어절마다 오타가 출현한 경우에서도 각각 80.09%, 87.53%, 92.49%의 높은 성능을 산출하였다. 추가 실험인 두 번째 실험에서는, 실생활에서 자주 발생하는 오타들을 집계하여 그 중에서 11가지 오타 유형을 선정 후, 각 유형에 대해 변환된 임베딩 벡터를 적용함으로써, 해당 오타를 포함한 문장에서 93.05%의 우수한 성능을 산출하였다.

• Journal of Advanced Marine Engineering and Technology
• /
• v.38 no.4
• /
• pp.444-450
• /
• 2014

Most researches on word sense disambiguation have used static sized context regardless of sentence patterns. This paper proposes to use dynamic sized context considering sentence patterns and distance between words for word sense disambiguation. We evaluated our system 12 words in 32,735sentences with Sejong POS and sense tagged corpus, and dynamic sized context showed 92.2% average accuracy for predicates, which is better than accuracy of static sized context.

• Annual Conference on Human and Language Technology
• /
• 2016.10a
• /
• pp.20-24
• /
• 2016

Sequence-to-sequence 모델은 입력열을 길이가 다른 출력열로 변환하는 모델로, 단일 신경망 구조만을 사용하는 End-to-end 방식의 모델이다. 본 논문에서는 Sequence-to-sequence 모델을 한국어 구구조 구문 분석에 적용한다. 이를 위해 구구조 구문 트리를 괄호와 구문 태그 및 어절로 이루어진 출력열의 형태로 만들고 어절들을 단일 기호 'XX'로 치환하여 출력 단어 사전의 수를 줄였다. 그리고 최근 기계번역의 성능을 높이기 위해 연구된 Attention mechanism과 Input-feeding을 적용하였다. 실험 결과, 세종말뭉치의 구구조 구문 분석 데이터에 대해 기존의 연구보다 높은 F1 89.03%의 성능을 보였다.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.12
• /
• pp.865-872
• /
• 2013

Two-level morphology analysis method is one of rule-based morphological analysis method. This approach handles morphological transformation using rules and analyzes words with morpheme connection information in a lexicon. It is independent of language and Korean Two-level system was also developed. But, it was limited in practical use, because of using very small set of lexicon built manually. And it has also a over-generation problem. In this paper, we propose an automatic construction method of Korean Two-level lexicon for PC-KIMMO from morpheme tagged corpus. We also propose a method to solve over-generation problem using lexical information and sub-tags. The experiment showed that the proposed method reduced over-generation by 68% compared with the previous method, and the performance increased from 39% to 65% in f-measure.

• The KIPS Transactions:PartB
• /
• v.11B no.4
• /
• pp.491-500
• /
• 2004

This paper presents an ontology construction method using various computational language resources, and an ontology-based word sense disambiguation method. In order to acquire a reasonably practical ontology the Kadokawa thesaurus is extended by inserting additional semantic relations into its hierarchy, which are classified as case relations and other semantic relations. To apply the ontology to disambiguate word senses, we apply the previously-secured dictionary information to select the correct senses of some ambiguous words with high precision, and then use the ontology to disambiguate the remaining ambiguous words. The mutual information between concepts in the ontology was calculated before using the ontology as knowledge for disambiguating word senses. If mutual information is regarded as a weight between ontology concepts, the ontology can be treated as a graph with weighted edges, and then we locate the weighted path from one concept to the other concept. In our practical machine translation system, our word sense disambiguation method achieved a 9% improvement over methods which do not use ontology for Korean translation.

• Korean Journal of Cognitive Science
• /
• v.17 no.4
• /
• pp.255-269
• /
• 2006

The purpose of this study is to analyze newspaper articles from corpus linguistic point of view. We used a large corpus of newspaper articles built from <21st century Sejong Project> and counted occurrences of certain expressions. A newspaper article is divided into the headline, the lead and the body. We tried to figure out how to measure the characteristics of indication and compression which are typical to headlines. Then, we focused on the differences between the headline and the lead. finally, we analyzed the sentence structure and measured the ratio of the frequency of common nouns in the body. This study verifies the existing stylistic theories of newspapers and shows new aspects of language use in newspaper articles. Texts like newspaper articles are the results of human language processing and they in turn affect the development of cognitive ability of language.

• The Journal of the Korea Contents Association
• /
• v.9 no.2
• /
• pp.18-26
• /
• 2009

A hybrid part-of-speech tagging approaches may be robust, easily extendable, and accurate because they can have the advantages of both statistical approach and rule-based approach. But conventional hybrid part-of-speech tagging systems hardly resolve some morphological ambiguities which can't be resolved by statistical information. It is because the coverage of rules is narrow. So, we define disambiguation rules for individual ambiguous word based on syntax and semantics of surround words. We select words from which the top 50% of ambiguities are occurred in Sejong corpus and build 1,814 rules for them. The accuracy of our hybrid part-of-speech tagging system using those rules is 98.28%.

• Journal of KIISE:Software and Applications
• /
• v.41 no.8
• /
• pp.580-585
• /
• 2014

This paper conducts a variety of experiments for "the extraction of Korean parallel sentences using Wikipedia data". We refer to various methods that were previously proposed for other languages. We use two approaches. The first one is to use translation probabilities that are extracted from the existing resources such as Sejong parallel corpus, and the second one is to use dictionaries such as Wiki dictionary consisting of Wikipedia titles and MRDs (machine readable dictionaries). Experimental results show that we obtained a significant improvement in system using Wikipedia data in comparison to one using only the existing resources. We finally achieve an outstanding performance, an F1-score of 57.6%. We additionally conduct experiments using a topic model. Although this experiment shows a relatively lower performance, an F1-score of 51.6%, it is expected to be worthy of further studies.

• Annual Conference on Human and Language Technology
• /
• 2017.10a
• /
• pp.305-308
• /
• 2017

한국어 형태소 분석은 많은 자연어 처리 분야에서 핵심적인 역할을 수행하고 있기 때문에 형태소를 분류하고 형태소에 맞는 알맞은 품사를 결정하는 것은 매우 중요하다. 형태소의 품사를 태깅하는 대표적인 방법은 크게 음절 단위 형태소 분석과 단어 단위 형태소 분석의 두 가지로 나눌 수 있다. 본 논문에서는 의존 파싱 분야에서 널리 활용되고 있는 전이 기반 방식을 적용하여 전이 기반 단어 단위 한국어 형태소 분석 모델을 제안하고 해당 모델을 한국어 형태소 분석 데이터인 세종 품사 부착 말뭉치 셋에 적용하여 F1 97.77 %로 기존의 성능을 더욱 향상시켰다.