• KIISE Transactions on Computing Practices
• /
• v.23 no.4
• /
• pp.226-231
• /
• 2017

Word representation, an important area in natural language processing(NLP) used machine learning, is a method that represents a word not by text but by distinguishable symbol. Existing word embedding employed a large number of corpora to ensure that words are positioned nearby within text. However corpus-based word embedding needs several corpora because of the frequency of word occurrence and increased number of words. In this paper word embedding is done using dictionary definitions and semantic relationship information(hypernyms and antonyms). Words are trained using the feature mirror model(FMM), a modified Skip-Gram(Word2Vec). Sense similar words have similar vector. Furthermore, it was possible to distinguish vectors of antonym words.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.4
• /
• pp.1400-1418
• /
• 2020

In the development of commercial promotion, chatbot is known as one of significant skill by application of natural language processing (NLP). Conventional design methods are using bag-of-words model (BOW) alone based on Google database and other online corpus. For one thing, in the bag-of-words model, the vectors are Irrelevant to one another. Even though this method is friendly to discrete features, it is not conducive to the machine to understand continuous statements due to the loss of the connection between words in the encoded word vector. For other thing, existing methods are used to test in state-of-the-art online corpus but it is hard to apply in real applications such as telemarketing data. In this paper, we propose an improved chatbot design way using hybrid bag-of-words model and skip-gram model based on the real telemarketing data. Specifically, we first collect the real data in the telemarketing field and perform data cleaning and data classification on the constructed corpus. Second, the word representation is adopted hybrid bag-of-words model and skip-gram model. The skip-gram model maps synonyms in the vicinity of vector space. The correlation between words is expressed, so the amount of information contained in the word vector is increased, making up for the shortcomings caused by using bag-of-words model alone. Third, we use the term frequency-inverse document frequency (TF-IDF) weighting method to improve the weight of key words, then output the final word expression. At last, the answer is produced using hybrid retrieval model and generate model. The retrieval model can accurately answer questions in the field. The generate model can supplement the question of answering the open domain, in which the answer to the final reply is completed by long-short term memory (LSTM) training and prediction. Experimental results show which the hybrid word vector expression model can improve the accuracy of the response and the whole system can communicate with humans.

• 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.

• Language and Information
• /
• v.14 no.1
• /
• pp.113-144
• /
• 2010

The primary goal of this paper is to find a feasible way to answer the question: Does the similarity in meaning between verbs relate to the similarity in their subcategorization? In order to answer this question in a rather concrete way on the basis of a large set of English verbs, this study made use of various language resources, tools, and statistical methodologies. We first compiled a list of 678 verbs that were selected from the most and second most frequent word lists from the Colins Cobuild English Dictionary, which also appeared in WordNet 3.0. We calculated similarity measures between all the pairs of the words based on the 'jcn' algorithm (Jiang and Conrath, 1997) implemented in the WordNet::Similarity module (Pedersen, Patwardhan, and Michelizzi, 2004). The clustering process followed, first building similarity matrices out of the similarity measure values, next drawing dendrograms on the basis of the matricies, then finally getting 177 meaningful clusters (covering 437 verbs) that passed a certain level set by z-score. The subcategorization frames and their frequency values were taken from the ICE-GB. In order to calculate the Selectional Preference Strength (SPS) of the relationship between a verb and its subcategorizations, we relied on the Kullback-Leibler Divergence model (Resnik, 1996). The SPS values of the verbs in the same cluster were compared with each other, which served to give the statistical values that indicate how much the SPS values overlap between the subcategorization frames of the verbs. Our final analysis shows that the degree of overlap, or the relationship between semantic similarity and the subcategorization frames of the verbs in English, is equally spread out from the 'very strongly related' to the 'very weakly related'. Some semantically similar verbs share a lot in terms of their subcategorization frames, and some others indicate an average degree of strength in the relationship, while the others, though still semantically similar, tend to share little in their subcategorization frames.

• Journal of Information Processing Systems
• /
• v.18 no.4
• /
• pp.535-548
• /
• 2022

Using a rich resource language to classify sentiments in a language with few resources is a popular subject of research in natural language processing. Burmese is a low-resource language. In light of the scarcity of labeled training data for sentiment classification in Burmese, in this study, we propose a method of transfer learning for sentiment analysis of a language that uses the feature transfer technique on sentiments in English. This method generates a cross-language word-embedding representation of Burmese vocabulary to map Burmese text to the semantic space of English text. A model to classify sentiments in English is then pre-trained using a convolutional neural network and an attention mechanism, where the network shares the model for sentiment analysis of English. The parameters of the network layer are used to learn the cross-language features of the sentiments, which are then transferred to the model to classify sentiments in Burmese. Finally, the model was tuned using the labeled Burmese data. The results of the experiments show that the proposed method can significantly improve the classification of sentiments in Burmese compared to a model trained using only a Burmese corpus.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2019.11a
• /
• pp.40-41
• /
• 2019

The purpose of this study is to analyze the vocabulary related to safety accidents based on the reports recorded on the violation of safety rules at the construction sites. We used Word2Vec and Topic Model as natural language processing techniques to analyze the safety accidents presented in the reports of the large enterprise. The words that appeared based on the occupational accident types such as the fall, falling objects, and others were derived and visualized. We derive the frequency and similarity of the words and topics of the accident that occur at the construction site. In future studies, we will be able to proceed with the generation of texts from pictures based on images and this reports.

• Proceedings of the Korean Society for Cognitive Science Conference
• /
• 2006.06a
• /
• pp.67-71
• /
• 2006

현재까지 제안된 자동 띄어쓰기 교정 모델들은 그 중의 대다수가 입력 문장에서 공백을 제거한 후에 교정 작업을 수행한다. 이러한 교정 방식은 입력 문장의 띄어쓰기가 잘 되어 있는 경우에 입력 문장보다 좋지 못한 교정 문장을 생성하는 경우가 있다. 본 논문에서는 이러한 문제점을 해결하기 위하여 입력 문장의 띄어쓰기를 고려한 자동 띄어쓰기 교정모델을 제안한다. 이 모델은 입력 문장의 음절단위 띄어쓰기 오류가 5%일 때 약 8%의 성능 향상을 보였으며, 10%의 오류가 존재할 때 약 5%의 성능 향상을 보였다.

• Proceedings of the KSPS conference
• /
• 2007.05a
• /
• pp.327-329
• /
• 2007

This paper presents a FSN-based LVCSR system and it's application to the speech TV program guide. Unlike the most popular statistical language model-based system, we used FSN grammar based on the graph theory-based FSN optimization algorithm and knowledge-based advanced word boundary modeling. For the memory and latency efficiency, we implemented the dynamic pruning scheduling based on the histogram of active words and their likelihood distribution. We achieved a 10.7% word accuracy improvement with 57.3% speedup.

• Annual Conference on Human and Language Technology
• /
• 2011.10a
• /
• pp.55-60
• /
• 2011

단어 클러스터링 (word clustering) 또는 군집화는 자연어처리에서 데이터 부족 문제로 인하여 단어 간의 의미관계와 관련된 정보를 사용하기 어렵게 만드는 문제에 대처할 수 있는 중요한 기술이다. 단어 클러스터링과 관련하여 알려진 가장 대표적인 기법으로는 클래스-기반 n-gram 언어모델의 개발을 위하여 제안된 Brown 단어 클러스터링 기법이다. 그러나 Brown 클러스터링 기법을 이용하는데 있어서 부딪치는 가장 큰 문제점은 시간과 공간적인 면에서 자원 소요량이 너무 방대하다는 점이다. 본 연구는 이 클러스터링 기법의 효율성을 개선하는 실험을 수행하였다. 실험 결과 가장 단순한(naive) 접근에 비하여 약 7.9배 이상의 속도 향상을 이룰 수 있음을 관찰하였다.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.8
• /
• pp.315-324
• /
• 2022

In this paper, we propose KoEPT, a Transformer-based generative model for automatic math word problems solving. A math word problem written in human language which describes everyday situations in a mathematical form. Math word problem solving requires an artificial intelligence model to understand the implied logic within the problem. Therefore, it is being studied variously across the world to improve the language understanding ability of artificial intelligence. In the case of the Korean language, studies so far have mainly attempted to solve problems by classifying them into templates, but there is a limitation in that these techniques are difficult to apply to datasets with high classification difficulty. To solve this problem, this paper used the KoEPT model which uses 'expression' tokens and pointer networks. To measure the performance of this model, the classification difficulty scores of IL, CC, and ALG514, which are existing Korean mathematical sentence problem datasets, were measured, and then the performance of KoEPT was evaluated using 5-fold cross-validation. For the Korean datasets used for evaluation, KoEPT obtained the state-of-the-art(SOTA) performance with 99.1% in CC, which is comparable to the existing SOTA performance, and 89.3% and 80.5% in IL and ALG514, respectively. In addition, as a result of evaluation, KoEPT showed a relatively improved performance for datasets with high classification difficulty. Through an ablation study, we uncovered that the use of the 'expression' tokens and pointer networks contributed to KoEPT's state of being less affected by classification difficulty while obtaining good performance.