• Proceedings of the Korean Information Science Society Conference
• /
• 2004.04b
• /
• pp.937-939
• /
• 2004

전문용어 조어법 분석은 기존의 전문용어들의 어휘의 구성과 구조를 파악하여 전문용어 생성의 원리를 밝혀 여러 응용시스템에 이용하기 위한 기초 작업에다. 조어법 정보를 이용한 전문용어 번역 시스템은 조어법 분석 결과의 조어단위 정렬과 색인을 통하여, 새로운 영어 용어에 대한 한국어 대역이 후보 집합을 생성한다. 생성된 후보들은 언어 모델의 정보량의 차이를 이용한 가중치에 의하여 순서화된다. 본 논문에서 제안하는 가중치 방법을 이용하여 조어법 분석 결과에 포함되지 않은 용어들을 대상으로 성능을 평가했을 때, 영-한 조어단위 번역의 n-best 정확률에서 1순위 정확률은 약 61%, 10순위 정확률은 97%의 성능을 보였다.

• Annual Conference on Human and Language Technology
• /
• 2010.10a
• /
• pp.86-93
• /
• 2010

개인 휴대용 단말기의 보급률이 높아짐에 따라, SMS 메시지가 또 하나의 새로운 의사소통 수단으로 발전하였다. 특히 통화보다 가격이 저렴하고, 통화 후 따로 적어두지 않아도 자동으로 저장되는 특징으로 인해 약속 등을 정할 때 많은 도움이 된다. 본 논문은 일반적인 정보추출 방법을 적용하여 이러한 SMS 메시지에서 자동으로 약속 시간과 장소를 추출한다. 기계학습 기법으로는 CRF를 이용하였으며, 비속어나 신조어가 많고 줄임말이 많은 SMS 메시지의 특징상 토큰분리나 품사 부착 등의 전처리 언어엔진을 사용하지 않았으며, 대신 Bi-Gram 언어모델을 사용하였으며, 학습 시 사전이나 어휘 등의 다양한 자질들을 적용하여 시스템의 정확도를 높였다.

• The KIPS Transactions:PartB
• /
• v.15B no.4
• /
• pp.365-374
• /
• 2008

In this paper, we propose a model to enrich an ontology by incrementally extending the relations through variations of patterns. In order to generalize initial patterns, combinations of features are considered as candidate patterns. The candidate patterns are used to extract relations from Wikipedia, which are sorted out according to reliability based on corpus frequency. Selected patterns then are used to extract relations, while extracted relations are again used to extend the patterns of the relation. Through making variations of patterns in incremental enrichment process, the range of pattern selection is broaden and refined, which can increase coverage and accuracy of relations extracted. In the experiments with single-feature based pattern models, we observe that the features of lexical, headword, and hypernym provide reliable information, while POS and syntactic features provide general information that is useful for enrichment of relations. Based on observations on the feature types that are appropriate for each syntactic unit type, we propose a pattern model based on the composition of features as our ongoing work.

• Journal of KIISE:Software and Applications
• /
• v.37 no.7
• /
• pp.548-557
• /
• 2010

In this paper, we propose a semi-automatic modeling approach of ontology to annotate VOD to realize the IPTV's intelligent searching. The ontology is made by combining partial tree that extracts hypernym, hyponym, and synonym of keywords related to a service domain from WordNet. Further, we add to the partial tree new keywords that are undefined in WordNet, such as foreign words and words written in Chinese characters. The ontology consists of two parts: generic hierarchy and specific hierarchy. The former is the semantic model of vocabularies such as keywords and contents of keywords. They are defined as classes including property restrictions in the ontology. The latter is generated using the reasoning technique by inferring contents of keywords based on the generic hierarchy. An annotation generates metadata (i.e., contents and genre) of VOD based on the specific hierarchy. The generic hierarchy can be applied to other domains, and the specific hierarchy helps modeling the ontology to fit the service domain. This approach is proved as good to generate metadata independent of any specific domain. As a result, the proposed method produced around 82% precision with 2,400 VOD annotation test data.

• The Journal of the Acoustical Society of Korea
• /
• v.18 no.1
• /
• pp.53-60
• /
• 1999

This paper is a study on the korean broadcasting speech recognition. Here we present the methods for the large vocabuary continuous speech recognition. Our main concerns are the language modeling and the search algorithm. The used acoustic model is the uni-phone semi-continuous hidden markov model and the used linguistic model is the N-gram model. The search algorithm consist of three phases in order to utilize all available acoustic and linguistic information. First, we use the forward Viterbi beam search to find word end frames and to estimate related scores. Second, we use the backword Viterbi beam search to find word begin frames and to estimate related scores. Finally, we use A/sup */ search to combine the above two results with the N-grams language model and to get recognition results. Using these methods maximum 96.0% word recognition rate and 99.2% syllable recognition rate are achieved for the speaker-independent continuous speech recognition problem with about 12,000 vocabulary size.

• The Journal of the Acoustical Society of Korea
• /
• v.16 no.6
• /
• pp.106-109
• /
• 1997

For the practical isolated word recognition system, the ability to reject the out-of -vocabulary(OOV) is required. In this paper, we present a rejection method which uses the clustered phoneme modeling combined with postprocessing by likelihood ratio scoring. Our baseline speech recognition system was based on the whole-word continuous HMM. And 6 clustered phoneme models were generated using statistical method from the 45 context independent phoneme models, which were trained using the phonetically balanced speech database. The test of the rejection performance for speaker independent isolated words recogntion task on the 22 section names shows that our method is superior to the conventional postprocessing method, performing the rejection according to the likelihood difference between the first and second candidates. Furthermore, this clustered phoneme models do not require retraining for the other isolated word recognition system with different vocabulary sets.

• Journal of KIISE
• /
• v.44 no.7
• /
• pp.674-679
• /
• 2017

In big-data environments wherein large amounts of text documents are produced daily, titles are very important clues that enable a prompt catching of the key ideas in documents; however, titles are absent for numerous document types such as blog articles and social-media messages. In this paper, a title-generation model for which sequence-to-sequence RNNs with attention and copying mechanisms are employed is proposed. For the proposed model, input sentences are encoded based on bi-directional GRU (gated recurrent unit) networks, and the title words are generated through a decoding of the encoded sentences with keywords that are automatically selected from the input sentences. Regarding the experiments with 93631 training-data documents and 500 test-data documents, the attention-mechanism performances are more effective (ROUGE-1: 0.1935, ROUGE-2: 0.0364, ROUGE-L: 0.1555) than those of the copying mechanism; in addition, the qualitative-evaluation radiative performance of the former is higher.

• Journal of KIISE:Software and Applications
• /
• v.30 no.11
• /
• pp.1112-1123
• /
• 2003

WSD(word sense disambiguation) is one of the most difficult problems in Korean information processing. The Bayesian model that used semantic information, extracted from definition corpus(1 million POS-tagged eojeol, Korean dictionary definitions), resulted in accuracy of 72.08% (nouns 78.12%, verbs 62.45%). This paper proposes the statistical WSD model using NPH(New Prior Probability of Homonym sense) and distance weights. We select 46 homonyms(30 nouns, 16 verbs) occurred high frequency in definition corpus, and then we experiment the model on 47,977 contexts from ‘21C Sejong Corpus’(3.5 million POS-tagged eojeol). The WSD model using NPH improves on accuracy to average 1.70% and the one using NPH and distance weights improves to 2.01%.

• Journal of KIISE:Software and Applications
• /
• v.35 no.5
• /
• pp.313-324
• /
• 2008

A sentence reduction is the information compression process which removes extraneous words and phrases and retains basic meaning of the original sentence. Most researches in the sentence reduction have required a large number of lexical and syntactic resources and focused on extracting or removing extraneous constituents such as words, phrases and clauses of the sentence via the complicated parsing process. However, these researches have some problems. First, the lexical resource which can be obtained in loaming data is very limited. Second, it is difficult to reduce the sentence to languages that have no method for reliable syntactic parsing because of an ambiguity and exceptional expression of the sentence. In order to solve these problems, we propose the sentence reduction method which uses templates and POS(part of speech) information without a parsing process. In our proposed method, we create a new sentence using both Sentence Reduction Templates that decide the reduction sentence form and Grammatical POS-based Reduction Rules that compose the grammatical sentence structure. In addition, We use Viterbi algorithms at HMM(Hidden Markov Models) to avoid the exponential calculation problem which occurs under applying to Sentence Reduction Templates. Finally, our experiments show that the proposed method achieves acceptable results in comparison to the previous sentence reduction methods.

• The Journal of the Korea Contents Association
• /
• v.22 no.10
• /
• pp.426-436
• /
• 2022

This study aimed to explore the role of models and reformulations as feedback to English writing in L2 learners' noticing and their writing. 92 participants were placed into three groups; a models group (MG), a reformulations group (RG), a control group (CG), involved in a three-stage writing task. In stage 1, they were asked to perform a 1st draft of writing, while taking notes on the problems they experienced. In stage 2, the MG was asked to compare their writing with a model text and the RG with a reformulated version of it. They were instructed to write down whatever they noticed in their comparison. The CG was asked to just read their writing. In stage 3, all the participants attempted subsequent revisions. The results indicated that all the participants noticed problematic linguistic features the most in a lexical category, and models and reformulations led to higher rate of noticing the problematic linguistic features reported in stage 1 and contributed to subsequent revisions. It was also revealed that the MG and RG significantly improved with their writings of MG and RG on the post-writing test. The findings imply that models and reformulations result in better performance in L2 writing and should be promoted in an English writing class.