• Journal of KIISE
• /
• v.44 no.12
• /
• pp.1290-1295
• /
• 2017

Document classification, a task of classifying the category of each document based on text, is one of the fundamental areas for natural language processing. Document classification may be used in various fields such as topic classification and sentiment classification. Neural network models for document classification can be divided into two categories: word-level models and character-level models that treat words and characters as basic units respectively. In this study, we propose a neural network model that combines character-level and word-level models to improve performance of document classification. The proposed model extracts the feature vector of each word by combining information obtained from a word embedding matrix and information encoded by a character-level neural network. Based on feature vectors of words, the model classifies documents with a hierarchical structure wherein recurrent neural networks with attention mechanisms are used for both the word and the sentence levels. Experiments on real life datasets demonstrate effectiveness of our proposed model.

• Phonetics and Speech Sciences
• /
• v.11 no.2
• /
• pp.57-64
• /
• 2019

In this paper, we construct a Korean text-to-speech system using the Merlin toolkit which is an open source system for speech synthesis. In the text-to-speech system, the HMM-based statistical parametric speech synthesis method is widely used, but it is known that the quality of synthesized speech is degraded due to limitations of the acoustic modeling scheme that includes context factors. In this paper, we propose an acoustic modeling architecture that uses deep neural network technique, which shows excellent performance in various fields. Fully connected deep feedforward neural network (DNN), recurrent neural network (RNN), gated recurrent unit (GRU), long short-term memory (LSTM), bidirectional LSTM (BLSTM) are included in the architecture. Experimental results have shown that the performance is improved by including sequence modeling in the architecture, and the architecture with LSTM or BLSTM shows the best performance. It has been also found that inclusion of delta and delta-delta components in the acoustic feature parameters is advantageous for performance improvement.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2017.11a
• /
• pp.50-53
• /
• 2017

Not only growth of importance to understanding economic trends, but also the prediction to overcome the uncertainty is coming up for long-term maritime recession. This paper discussed about the prediction of BDI with artificial neural networks (ANN). ANN is one of emerging applications that can be the finest solution to the knotty problems that may not easy to achieve by humankind. Proposed a prediction by implementing neural networks that have recurrent architecture which are a Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM). And for the reason of comparison, trained Multi Layer Perceptron (MLP) from 2009.04.01 to 2017.07.31. Also made a comparison with conventional statistics, prediction tools; ARIMA. As a result, recurrent net, especially RNN outperformed and also could discover the applicability of LSTM to specific time-series (BDI).

• Information Systems Review
• /
• v.20 no.4
• /
• pp.25-41
• /
• 2018

Applying predictive analytics to enterprise processes is an effective way to reduce operation costs and enhance productivity. Accordingly, the ability to predict business processes and performance indicators are regarded as a core capability. Recently, several works have predicted processes using deep learning in the form of recurrent neural networks (RNN). In particular, the approach of predicting the next step of activity using static or dynamic RNN has excellent results. However, few studies have given attention to applying deep learning in the form of dynamic RNN to predictions of process performance indicators. To fill this knowledge gap, the study developed an approach to using process mining and dynamic RNN. By utilizing actual data from a large domestic company, it has applied the suggested approach in estimating timely stocking in purchasing process, which is an important indicator of the process. The analytic methods and results of this study were presented and some implications and limitations are also discussed.

• Journal of Internet of Things and Convergence
• /
• v.9 no.6
• /
• pp.111-117
• /
• 2023

There is a lot of research on using a combination of graph neural networks and recurrent neural networks as a way to account for both temporal and spatial dependencies. In particular, graph neural networks are an emerging area of research. Seoul's bicycle rental service (aka Daereungi) has rental stations all over the city of Seoul, and the rental information at each station is a time series that is faithfully recorded. The rental information of each rental station has temporal characteristics that show periodicity over time, and regional characteristics are also thought to have important effects on the rental status. Regional correlations can be well understood using graph neural networks. In this study, we reconstructed the time series data of Seoul's bicycle rental service into a graph and developed a rental prediction model that combines a graph neural network and a recurrent neural network. We considered temporal characteristics such as periodicity over time, regional characteristics, and the degree importance of each rental station.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.9
• /
• pp.1305-1311
• /
• 2022

With the recent rapid development of mobile terminals and personal computers and the advent of neural network technology, real-time face swapping using images has become possible. In particular, the cycle generative adversarial network made it possible to replace faces using uncorrelated image data. In this paper, we propose an input data processing scheme that can improve the quality of face swapping with less training data and time. The proposed scheme can improve the image quality while preserving facial structure and expression information by combining facial landmarks extracted through a pre-trained neural network with major information that affects the structure and expression of the face. Using the blind/referenceless image spatial quality evaluator (BRISQUE) score, which is one of the AI-based non-reference quality metrics, we quantitatively analyze the performance of the proposed scheme and compare it to the conventional schemes. According to the numerical results, the proposed scheme obtained BRISQUE scores improved by about 4.6% to 14.6%, compared to the conventional schemes.

• Journal of The Korean Association of Information Education
• /
• v.21 no.2
• /
• pp.161-169
• /
• 2017

We often use spellcheckers in order to correct the syntactic errors in our documents. However, these computer programs are not enough for elementary school students, because their sentences are not smooth even after correcting the syntactic errors in many cases. In this paper, we introduce an automated method for evaluating the smoothness of two synonymous sentences. This method uses a recurrent neural network to solve the problem of long-term dependencies and exploits subwords to cope with the rare word problem. We trained the recurrent neural network language model based on a monolingual corpus of about two million English sentences. In our experiments, the trained model successfully selected the more smooth sentences for all of nine types of test set. We expect that our approach will help in elementary school writing after being implemented as an application for smart devices.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.592-595
• /
• 2021

양자 알고리즘인 그루버나 쇼어 알고리즘에 의해 현존하는 암호 체계들이 무너질 수 있으며, 블록체인 네트워크를 기반으로 타원곡선 암호 및 타원곡선 전자서명을 사용하는 암호화폐의 안전성 또한 위협받고 있다. 따라서 암호화폐에도 양자 컴퓨터에 대한 대응책이 필요하다. 본 논문에서는 시계열 예측에 적합한 순환형 신경망을 활용하여 양자 저항성을 가지는 암호화폐들의 가격을 예측하고 분석한다. 데이터가 부족하였으나 학습 결과 0.005 이하의 손실을 달성하였으며, 최근 15일의 데이터를 통해 예측한 결과, 모두 소폭 상승할 것으로 나타났다. 향후에는 더 많은 데이터를 통해 더 정확한 예측이 가능한 신경망을 설계하고 다양한 양자 관련 이슈들을 참고하여 분석을 수행하고자 한다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.473-474
• /
• 2023

본 논문에서는 딥러닝 기법 중 하나인 순환신경망(RNN)을 활용하여 자연어를 처리할 수 있는 모델 개발에 대하여 연구를 진행하였다. 다양한 주제에 대한 사용자들의 의견을 확보할 수 있는 유튜브 플랫픔을 활용하여 데이터를 확보하였으며, 감성 분류를 진행하는 만큼 학습 데이터셋으로는 네이버 영화 리뷰 데이터셋을 활용하였다. 사용자는 직접 데이터 파일을 삽입하거나 혹은 유튜브 댓글과 같이 데이터를 외부에서 확보하여 감성을 분석할 수 있으며, 자연어 속 등장하는 단어의 빈도수를 종합하여 해당 데이터들 속 키워드는 무엇인지를 분석할 수 있도록 하였다. 나아가 종합 데이터 분석 관리 플랫폼을 제작하기 위하여 해당 데이터를 데이터베이스에 저장하고GUI 프로그램을 통하여 접근 및 관리가 가능하도록 하였다.

• Annual Conference on Human and Language Technology
• /
• 2016.10a
• /
• pp.321-323
• /
• 2016

최근 연구에서 기계학습 중 지도학습 방법으로 개체명 인식을 하고 있다. 그러나 지도 학습 방법은 데이터를 만드는 비용과 시간이 많이 필요로 한다. 본 연구에서는 주석 된 말뭉치를 사용하여 지도 학습 방법을 사용 한다. 의생명 개체명 인식은 Protein, RNA, DNA, Cell type, Cell line 등을 포함한 텍스트 처리에 중요한 기초 작업입니다. 그리고 의생명 지식 검색에서 가장 기본과 핵심 작업 중 하나이다. 본 연구에서는 순환형 신경망과 워드 임베딩을 자질로 사용한 조건적 임의 필드에 대한 성능을 비교한다. 조건적 임의 필드에 N_Gram만을 자질로 사용한 것을 기준점으로 설정 하였고, 기준점의 결과는 70.09% F1 Score이다. RNN의 jordan type은 60.75% F1 Score, elman type은 58.80% F1 Score의 성능을 보여준다. 조건적 임의 필드에 CCA, GLOVE, WORD2VEC을 사용 한 결과는 각각 72.73% F1 Score, 72.74% F1 Score, 72.82% F1 Score의 성능을 얻을 수 있다.