• Journal of Multimedia Information System
• /
• v.9 no.2
• /
• pp.93-102
• /
• 2022

Dependency parsing is a decision problem of the syntactic relation between words in a sentence. Recently, deep learning models are used for dependency parsing based on the word representations in a continuous vector space. However, it causes a mislabeled tagging problem for the proper nouns that rarely appear in the training corpus because it is difficult to express out-of-vocabulary (OOV) words in a continuous vector space. To solve the OOV problem in dependency parsing, we explored the proper noun embedding method according to the embedding unit. Before representing words in a continuous vector space, we replace the proper nouns with a special token and train them for the contextual features by using the multi-layer bidirectional LSTM. Two models of the syllable-based and morpheme-based unit are proposed for proper noun embedding and the performance of the dependency parsing is more improved in the ensemble model than each syllable and morpheme embedding model. The experimental results showed that our ensemble model improved 1.69%p in UAS and 2.17%p in LAS than the same arc-eager approach-based Malt parser.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.11a
• /
• pp.1014-1017
• /
• 2020

Many natural language processing (NLP) models utilize pre-trained word embeddings to leverage latent information. One of the most successful word embedding model is the Skip-gram (SG). In this paper, we propose a Skipgram drop (SG-Drop) model, which is a variation of the SG model. The SG-Drop model is designed to reduce training time efficiently. Furthermore, the SG-Drop allows controlling training time with its hyperparameter. It could train word embedding faster than reducing training epochs while better preserving the quality.

• Journal of the Korea Convergence Society
• /
• v.11 no.11
• /
• pp.33-39
• /
• 2020

In the field of Natural Language Processing, Various research such as Translation, POS Tagging, Q&A, and Sentiment Analysis are globally being carried out. Sentiment Analysis shows high classification performance for English single-domain datasets by pretrained sentence embedding models. In this thesis, the classification performance is compared by Korean E-commerce online dataset with various domain attributes and 6 Neural-Net models are built as BOW (Bag Of Word), LSTM[1], Attention, CNN[2], ELMo[3], and BERT(KoBERT)[4]. It has been confirmed that the performance of pretrained sentence embedding models are higher than word embedding models. In addition, practical Neural-Net model composition is proposed after comparing classification performance on dataset with 17 categories. Furthermore, the way of compressing sentence embedding model is mentioned as future work, considering inference time against model capacity on real-time service.

• International Journal of Internet, Broadcasting and Communication
• /
• v.13 no.2
• /
• pp.173-178
• /
• 2021

In this paper, we find a suitable methodology for Korean Sentiment Analysis through a comparative experiment in which methods of embedding and natural network models are learned at the highest accuracy and fastest speed. The embedding method compares word embeddeding and Word2Vec. The model compares and experiments representative neural network models CNN, RNN, LSTM, GRU, Bi-LSTM and Bi-GRU with IKEA review data. Experiments show that Word2Vec and BiGRU had the highest accuracy and second fastest speed with 94.23% accuracy and 42.30 seconds speed. Word2Vec and GRU were found to have the third highest accuracy and fastest speed with 92.53% accuracy and 26.75 seconds speed.

• Journal of Information Processing Systems
• /
• v.18 no.3
• /
• pp.344-358
• /
• 2022

Text summarization is the task of producing a shorter version of a long document while accurately preserving the main contents of the original text. Abstractive summarization generates novel words and phrases using a language generation method through text transformation and prior-embedded word information. However, newly coined words or out-of-vocabulary words decrease the performance of automatic summarization because they are not pre-trained in the machine learning process. In this study, we demonstrated an improvement in summarization quality through the contextualized embedding of BERT with out-of-vocabulary masking. In addition, explicitly providing precise pointing and an optional copy instruction along with BERT embedding, we achieved an increased accuracy than the baseline model. The recall-based word-generation metric ROUGE-1 score was 55.11 and the word-order-based ROUGE-L score was 39.65.

• Journal of Convergence for Information Technology
• /
• v.12 no.2
• /
• pp.47-53
• /
• 2022

Recently, with the change of the intelligent security paradigm, study to apply various information generated from various information security systems to AI-based anomaly detection is increasing. Therefore, in this study, in order to convert log-like time series data into a vector, which is a numerical feature, the CBOW and Skip-gram inference methods of deep learning-based Word2Vec model and statistical method based on the coincidence frequency were used to transform the published ADFA system call data. In relation to this, an experiment was carried out through conversion into various embedding vectors considering the dimension of vector, the length of sequence, and the window size. In addition, the performance of the embedding methods used as well as the detection performance were compared and evaluated through GRU-based anomaly detection model using vectors generated by the embedding model as an input. Compared to the statistical model, it was confirmed that the Skip-gram maintains more stable performance without biasing a specific window size or sequence length, and is more effective in making each event of sequence data into an embedding vector.

• Journal of Digital Contents Society
• /
• v.18 no.7
• /
• pp.1419-1425
• /
• 2017

The ultimate goal of legal knowledge search is to obtain optimal legal information based on laws and precedent. Text mining research is actively being undertaken to meet the needs of efficient retrieval from large scale data. A typical method is to use a word embedding algorithm based on Neural Net. This paper demonstrates how to search relevant information, applying Korean law information to word embedding. First, we extracts reference laws from precedents in order and takes reference laws as input of Law2Vec. The model learns a law by predicting its surrounding context law. The algorithm then moves over each law in the corpus and repeats the training step. After the training finished, we could infer the relationship between the laws via the embedding method. The search performance was evaluated based on precision and the recall rate which are computed from how closely the results are associated to the search terms. The test result proved that what this paper proposes is much more useful compared to existing systems utilizing only keyword search when it comes to extracting related laws.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2020.07a
• /
• pp.199-202
• /
• 2020

본 논문에서는 가짜뉴스 탐지 모델에 워드 임베딩 기법을 접목하여 성능을 향상시키는 방법을 제안한다. 기존의 한국어 가짜뉴스 탐지 연구는 희소 표현인 빈도-역문서 빈도(TF-IDF)를 활용한 탐지 모델들이 주를 이루었다. 하지만 이는 가짜뉴스 탐지의 관점에서 뉴스의 언어적 특성을 파악하는 데 한계가 존재하는데, 특히 문맥에서 드러나는 언어적 특성을 구조적으로 반영하지 못한다. 이에 밀집 표현 기반의 워드 임베딩 기법인 Word2vec을 활용한 텍스트 전처리를 통해 문맥 정보까지 반영한 가짜뉴스 탐지 모델을 본 연구의 제안 모델로 생성한 후 TF-IDF 기반의 가짜뉴스 탐지 모델을 비교 모델로 생성하여 두 모델 간의 비교를 통한 성능 검증을 수행하였다. 그 결과 Word2vec 기반의 제안모형이 더욱 우수하였음을 확인하였다.

• Journal of Information Processing Systems
• /
• v.19 no.6
• /
• pp.830-841
• /
• 2023

Tokenization is the process of segmenting the input text into smaller units of text, and it is a preprocessing task that is mainly performed to improve the efficiency of the machine learning process. Various tokenization methods have been proposed for application in the field of natural language processing, but studies have primarily focused on efficiently segmenting text. Few studies have been conducted on the Korean language to explore what tokenization methods are suitable for document classification task. In this paper, an exploratory study was performed to find the most suitable tokenization method to improve the performance of a representative spatio-temporal document classifier in Korean. For the experiment, a convolutional neural network model was used, and for the final performance comparison, tasks were selected for document classification where performance largely depends on the tokenization method. As a tokenization method for comparative experiments, commonly used Jamo, Character, and Word units were adopted. As a result of the experiment, it was confirmed that the tokenization of word units showed excellent performance in the case of representative spatio-temporal document classification task where the semantic embedding ability of the token itself is important.

• ETRI Journal
• /
• v.44 no.3
• /
• pp.413-425
• /
• 2022

The study of literary texts is one of the earliest disciplines practiced around the globe. Poetry is artistic writing in which words are carefully chosen and arranged for their meaning, sound, and rhythm. Poetry usually has a broad and profound sense that makes it difficult to be interpreted even by humans. The essence of poetry is Rasa, which signifies mood or emotion. In this paper, we propose a poetry classification-based approach to automatically extract similar poems from a repository. Specifically, we perform a novel Rasa-based classification of Hindi poetry. For the task, we primarily used lexical features in a bag-of-words model trained using the support vector machine classifier. In the model, we employed Hindi WordNet, Latent Semantic Indexing, and Word2Vec-based neural word embedding. To extract the rich feature vectors, we prepared a repository containing 37 717 poems collected from various sources. We evaluated the performance of the system on a manually constructed dataset containing 945 Hindi poems. Experimental results demonstrated that the proposed model attained satisfactory performance.