• MALSORI
• /
• v.68
• /
• pp.95-114
• /
• 2008

The recent growing popularity of statistical methods in machine translation requires much more large parallel corpora. A Korean-English parallel corpus, however, is not yet enoughly available, little research on this subject is being conducted. In this paper we present a hybrid method of aligning sentences for Korean-English parallel corpora. We use bilingual news wire web pages, reading comprehension materials for English learners, computer-related technical documents and help files of localized software for building a Korean-English parallel corpus. Our hybrid method combines sentence-length based and word-correspondence based methods. We show the results of experimentation and evaluate them. Alignment results from using a full translation model are very encouraging, especially when we apply alignment results to an SMT system: 0.66% for BLEU score and 9.94% for NIST score improvement compared to the previous method.

• Proceedings of the Korean Society for Language and Information Conference
• /
• 2007.11a
• /
• pp.285-292
• /
• 2007

The parallel corpus is an important resource in the research field of data-driven natural language processing, but there are only a few parallel corpora publicly available nowadays, mostly due to the high labor force needed to construct this kind of resource. A novel strategy is brought out to automatically fetch parallel text from the web in this paper, which may help to solve the problem of the lack of parallel corpora with high quality. The system we develop first downloads the web pages from certain hosts. Then candidate parallel page pairs are prepared from the page set based on the outer features of the web pages. The candidate page pairs are evaluated in the last step in which the sentences in the candidate web page pairs are extracted and aligned first, and then the similarity of the two web pages is evaluate based on the similarities of the aligned sentences. The experiments towards a multilingual web site show the satisfactory performance of the system.

• Proceedings of the Korean Society for Emotion and Sensibility Conference
• /
• 2008.10a
• /
• pp.90-93
• /
• 2008

In natural language, it is common that repetitive constituents in an expression are to be left out and it is necessary to figure out the constituents omitted at analyzing the meaning of the sentence. This paper is on recognition of boundaries of parallel noun phrases by figuring out constituents omitted. Recognition of parallel noun phrases can greatly reduce complexity at the phase of sentence parsing. Moreover, in natural language information retrieval, recognition of noun with modifiers can play an important role in making indexes. We propose an unsupervised probabilistic model that identifies parallel cores as well as boundaries of parallel noun phrases conjoined by a conjunctive particle. It is based on the idea of swapping constituents, utilizing symmetry (two or more identical constituents are repeated) and reversibility (the order of constituents is changeable) in parallel structure. Semantic features of the modifiers around parallel noun phrase, are also used the probabilistic swapping model. The model is language-independent and in this paper presented on parallel noun phrases in Korean language. Experiment shows that our probabilistic model outperforms symmetry-based model and supervised machine learning based approaches.

• Smart Media Journal
• /
• v.10 no.1
• /
• pp.16-24
• /
• 2021

The exiting word embedding methodology such as word2vec represents words, which only occur in the raw training corpus, as a fixed-length vector into a continuous vector space, so when mapping the words incorporated in the raw training corpus into a fixed-length vector in morphologically rich language, out-of-vocabulary (OOV) problem often happens. Even for sentence embedding, when representing the meaning of a sentence as a fixed-length vector by synthesizing word vectors constituting a sentence, OOV words make it challenging to meaningfully represent a sentence into a fixed-length vector. In particular, since the agglutinative language, the Korean has a morphological characteristic to integrate lexical morpheme and grammatical morpheme, handling OOV words is an important factor in improving performance. In this paper, we propose parallel Tri-LSTM sentence embedding that is robust to the OOV problem by extending utilizing the morphological information of words into sentence-level. As a result of the sentiment analysis task with corpus in Korean, we empirically found that the character unit is better than the morpheme unit as an embedding unit for Korean sentence embedding. We achieved 86.17% accuracy on the sentiment analysis task with the parallel bidirectional Tri-LSTM sentence encoder.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.3
• /
• pp.125-132
• /
• 2022

Sentence compression is a natural language processing task that generates concise sentences that preserves the important meaning of the original sentence. For grammatically appropriate sentence compression, early studies utilized human-defined linguistic rules. Furthermore, while the sequence-to-sequence models perform well on various natural language processing tasks, such as machine translation, there have been studies that utilize it for sentence compression. However, for the linguistic rule-based studies, all rules have to be defined by human, and for the sequence-to-sequence model based studies require a large amount of parallel data for model training. In order to address these challenges, Deleter, a sentence compression model that leverages a pre-trained language model BERT, is proposed. Because the Deleter utilizes perplexity based score computed over BERT to compress sentences, any linguistic rules and parallel dataset is not required for sentence compression. However, because Deleter compresses sentences only considering perplexity, it does not compress sentences by reflecting the linguistic information of the words in the sentences. Furthermore, since the dataset used for pre-learning BERT are far from compressed sentences, there is a problem that this can lad to incorrect sentence compression. In order to address these problems, this paper proposes a method to quantify the importance of linguistic information and reflect it in perplexity-based sentence scoring. Furthermore, by fine-tuning BERT with a corpus of news articles that often contain proper nouns and often omit the unnecessary modifiers, we allow BERT to measure the perplexity appropriate for sentence compression. The evaluations on the English and Korean dataset confirm that the sentence compression performance of sentence-scoring based models can be improved by utilizing the proposed method.

• Journal of Information Processing Systems
• /
• v.8 no.2
• /
• pp.301-314
• /
• 2012

In this paper, two new approaches to align English-Arabic sentences in bilingual parallel corpora based on the Multi-Class Support Vector Machine (MSVM) and the Hidden Markov Model (HMM) classifiers are presented. A feature vector is extracted from the text pair that is under consideration. This vector contains text features such as length, punctuation score, and cognate score values. A set of manually prepared training data was assigned to train the Multi-Class Support Vector Machine and Hidden Markov Model. Another set of data was used for testing. The results of the MSVM and HMM outperform the results of the length based approach. Moreover these new approaches are valid for any language pairs and are quite flexible since the feature vector may contain less, more, or different features, such as a lexical matching feature and Hanzi characters in Japanese-Chinese texts, than the ones used in the current research.

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.10a
• /
• pp.787-789
• /
• 2004

People frequently feel the need of a general searching tool that frees from language barrier when they find information through the internet. Therefore, it is necessary to have a multilingual parallel corpus to search with a word that includes a search keyword and has a corresponding word in another language, Multilingual parallel corpus can be built and reused effectively through the several processes which are judgment of the web documents, sentence alignment and word alignment. To build a multilingual parallel corpus, multi-lingual dictionary should be constructed in each language and HTML should be simplified. And by understanding the meaning and the statistics of document structure, judgment on translated web documents will be made and the searched web pages will be aligned in sentence unit.

• Korean Journal of Cognitive Science
• /
• v.7 no.4
• /
• pp.179-202
• /
• 1996

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.6
• /
• pp.433-442
• /
• 2013

A set of bilingual terms is one of the most important factors in building language-related applications such as a machine translation system and a cross-lingual information system. In this paper, we introduce a new approach that automatically extracts candidates of English-Korean bilingual terms by using a bilingual parallel corpus and a basic English-Korean lexicon. This approach can be useful even though the size of the parallel corpus is small. A sentence alignment is achieved first for the document-level parallel corpus. We can align words between a pair of aligned sentences by referencing a basic bilingual lexicon. For unaligned words between a pair of aligned sentences, several assumptions are applied in order to align bilingual term candidates of two languages. A location of a sentence, a relation between words, and linguistic information between two languages are examples of the assumptions. An experimental result shows approximately 71.7% accuracy for the English-Korean bilingual term candidates which are automatically extracted from 1,000 bilingual parallel corpus.

• Journal of KIISE
• /
• v.42 no.7
• /
• pp.901-909
• /
• 2015

In this paper, to build a parallel corpus between Korean and English in Wikipedia. We proposed a method to find similar sentences based on language resources and topic modeling. We first applied language resources(Wiki-dictionary, numbers, and online dictionary in Daum) to match word sequentially. We construct the Wiki-dictionary using titles in Wikipedia. In order to take advantages of the Wikipedia, we used translation probability in the Wiki-dictionary for word matching. In addition, we improved the accuracy of sentence similarity measuring method by using word distribution based on topic modeling. In the experiment, a previous study showed 48.4% of F1-score with only language resources based on linear combination and 51.6% with the topic modeling considering entire word distributions additionally. However, our proposed methods with sequential matching added translation probability to language resources and achieved 9.9% (58.3%) better result than the previous study. When using the proposed sequential matching method of language resources and topic modeling after considering important word distributions, the proposed system achieved 7.5%(59.1%) better than the previous study.