• 한국어정보학회:학술대회논문집
• /
• 2017.10a
• /
• pp.159-162
• /
• 2017

Long Short-Term Memory Recurrent Neural Network(LSTM RNN)는 순차 데이터 모델링에 적합한 딥러닝 모델이다. Bidirectional LSTM RNN(BiLSTM RNN)은 RNN의 그래디언트 소멸 문제(vanishing gradient problem)를 해결한 LSTM RNN을 입력 데이터의 양 방향에 적용시킨 것으로 입력 열의 모든 정보를 볼 수 있는 장점이 있어 자연어처리를 비롯한 다양한 분야에서 많이 사용되고 있다. Highway Network는 비선형 변환을 거치지 않은 입력 정보를 히든레이어에서 직접 사용할 수 있게 LSTM 유닛에 게이트를 추가한 딥러닝 모델이다. 본 논문에서는 Highway Network를 한국어 의미역 결정에 적용하여 기존 연구 보다 더 높은 성능을 얻을 수 있음을 보인다.

• Annual Conference on Human and Language Technology
• /
• 2016.10a
• /
• pp.99-101
• /
• 2016

문자 기반 LSTM CRF는 개체명 인식에서 높은 인식을 보여주고 있는 LSTM-CRF 방식에서 미등록어 문제를 해결하기 위해 단어 단위의 임베딩 뿐만 아니라 단어를 구성하는 문자로부터 단어 임베딩을 합성해 내는 방식으로 기존의 LSTM CRF에서의 성능 향상을 가져왔다. 한편, 개체명 인식에서 어휘 사전은 성능 향상을 위한 외부 리소스원으로 활용하고 있는데 다양한 사전 매칭 방법이 파생될 수 있음에도 이들 자질들에 대한 비교 연구가 이루어지지 않았다. 본 논문에서는 개체명 인식을 위해 다양한 사전 매칭 자질들을 정의하고 이들을 LSTM-CRF의 입력 자질로 활용했을 때의 성능 비교 결과를 제시한다. 실험 결과 사전 자질이 추가된 LSTM-CRF는 ETRI 개체명 말뭉치의 학습데이터에서 F1 measure 기준 최대 89.34%의 성능까지 달성할 수 있었다.

• Annual Conference on Human and Language Technology
• /
• 2016.10a
• /
• pp.105-110
• /
• 2016

개체명 인식이란 문서 내에서 인명, 기관명, 지명, 시간, 날짜 등 고유한 의미를 가지는 개체명을 추출하여 그 종류를 결정하는 것을 말한다. 최근 개체명 인식 연구에서는 bidirectional LSTM CRFs가 가장 우수한 성능을 보여주고 있다. 하지만 LSTM 기반의 딥 러닝 모델은 입력이 되는 단어 표상에 의존적이기 때문에 입력이 되는 단어 표상을 확장하는 방법에 대한 연구가 많이 진행되어지고 있다. 본 논문에서는 한국어 개체명 인식을 위하여 bidirectional LSTM CRFs모델을 사용하고, 그 입력으로 사용되는 단어 표상을 확장하기 위해 사전 학습된 단어 임베딩 벡터, 품사 임베딩 벡터, 그리고 음절 기반에서 확장된 단어 임베딩 벡터를 사용한다. 음절 기반에서 단어 기반 임베딩 벡터로 확장하기 위하여 bidirectional LSTM을 이용하고, 그 입력으로 학습 데이터에서 추출한 개체명 분포를 이용하였다. 그 결과 사전 학습된 단어 임베딩 벡터만 사용한 것보다 4.93%의 성능 향상을 보였다.

• Journal of information and communication convergence engineering
• /
• v.22 no.2
• /
• pp.139-144
• /
• 2024

In this study, four types of fall detection systems - designed with YOLOPose, principal component analysis (PCA), convolutional neural network (CNN), and long short-term memory (LSTM) architectures - were developed and compared in the detection of everyday falls. The experimental dataset encompassed seven types of activities: walking, lying, jumping, jumping in activities of daily living, falling backward, falling forward, and falling sideways. Keypoints extracted from YOLOPose were entered into the following architectures: RAW-LSTM, PCA-LSTM, RAW-PCA-LSTM, and PCA-CNN-LSTM. For the PCA architectures, the reduced input size stemming from a dimensionality reduction enhanced the operational efficiency in terms of computational time and memory at the cost of decreased accuracy. In contrast, the addition of a CNN resulted in higher complexity and lower accuracy. The RAW-LSTM architecture, which did not include either PCA or CNN, had the least number of parameters, which resulted in the best computational time and memory while also achieving the highest accuracy.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2024.01a
• /
• pp.339-341
• /
• 2024

본 논문에서는 전기 자동차 배터리 팩 설계에서 성능 예측을 위해 전산유체해석 및 Long Short-Term Memory (LSTM)를 활용한다. 두 계산 모두의 예측이 상당한 유사성을 나타내며, 전산유체해석은 시스템 유체 역학을 고려한 상세한 물리 모델을 제공하고, LSTM은 시계열 데이터를 기반으로 한 딥러닝 모델로 효과적으로 패턴을 파악, 향후 온도 상승을 예측한다. 결과는 두 접근 모두가 효과적인 예측을 제공하며 향후 전기 자동차 배터리 팩 설계 및 최적화에서 종합적인 접근의 필요성을 강조한다. 특히, LSTM 기반 예측에 소요되는 시간은 계산 유체 역학의 약 25%로, 약 일주일 정도로 빠르게 확인 가능하다. 이는 현대 산업 환경에서 시간적 효율성이 중요한 측면을 강조하며, 계산 유체 역학의 상세한 물리 모델링과 LSTM의 빠른 예측 속도를 결합한 설계 방법론을 제안한다.

• Korean Journal of Remote Sensing
• /
• v.38 no.3
• /
• pp.299-309
• /
• 2022

Recently, The sea water temperature around Korean Peninsula is steadily increasing. Water temperature changes not only affect the fishing ecosystem, but also are closely related to military operations in the sea. The purpose of this study is to suggest which model is more suitable for the field of water temperature prediction by attempting short-term water temperature prediction through various prediction models based on deep learning technology. The data used for prediction are water temperature data from the East Sea (Goseong, Yangyang, Gangneung, and Yeongdeok) from 2016 to 2020, which were observed through marine observation by the National Fisheries Research Institute. In addition, we use Long Short-Term Memory (LSTM), Bidirectional LSTM, and Gated Recurrent Unit (GRU) techniques that show excellent performance in predicting time series data as models for prediction. While the previous study used only LSTM, in this study, the prediction accuracy of each technique and the performance time were compared by applying various techniques in addition to LSTM. As a result of the study, it was confirmed that Bidirectional LSTM and GRU techniques had the least error between actual and predicted values at all observation points based on 1 hour prediction, and GRU was the fastest in learning time. Through this, it was confirmed that a method using Bidirectional LSTM was required for water temperature prediction to improve accuracy while reducing prediction errors. In areas that require real-time prediction in addition to accuracy, such as anti-submarine operations, it is judged that the method of using the GRU technique will be more appropriate.

• The Journal of Bigdata
• /
• v.5 no.1
• /
• pp.197-208
• /
• 2020

Lexical ambiguity means that a word can be interpreted as two or more meanings, such as homonym and polysemy, and there are many cases of word sense ambiguation in words expressing emotions. In terms of projecting human psychology, these words convey specific and rich contexts, resulting in lexical ambiguity. In this study, we propose an emotional classification model that disambiguate word sense using bidirectional LSTM. It is based on the assumption that if the information of the surrounding context is fully reflected, the problem of lexical ambiguity can be solved and the emotions that the sentence wants to express can be expressed as one. Bidirectional LSTM is an algorithm that is frequently used in the field of natural language processing research requiring contextual information and is also intended to be used in this study to learn context. GloVe embedding is used as the embedding layer of this research model, and the performance of this model was verified compared to the model applied with LSTM and RNN algorithms. Such a framework could contribute to various fields, including marketing, which could connect the emotions of SNS users to their desire for consumption.

• Journal of Digital Convergence
• /
• v.17 no.12
• /
• pp.243-248
• /
• 2019

Multi-Task Learning(MTL) is a training method that trains a single neural network with multiple tasks influences each other. In this paper, we compare performance of MTL Named entity recognition(NER) model trained with Korean traditional culture corpus and other NER model. In training process, each Bi-LSTM layer of Part of speech tagging(POS-tagging) and NER are propagated from a Bi-LSTM layer to obtain the joint loss. As a result, the MTL based Bi-LSTM model shows 1.1%~4.6% performance improvement compared to single Bi-LSTM models.

• The Journal of Society for e-Business Studies
• /
• v.26 no.1
• /
• pp.127-140
• /
• 2021

Recently, sequence data containing time information, such as sensor measurement data and purchase history, has been generated in various applications. So far, many methods for finding sequences that are significantly different from other sequences among given sequences have been proposed. However, most of them have a limitation that they consider only the order of elements in the sequences. Therefore, in this paper, we propose a new anomalous sequence detection method that considers both the order of elements and the time interval between elements. The proposed method uses an extended LSTM autoencoder model, which has an additional layer that converts a sequence into a form that can help effectively learn both the order of elements and the time interval between elements. The proposed method learns the features of the given sequences with the extended LSTM autoencoder model, and then detects sequences that the model does not reconstruct well as anomalous sequences. Using experiments on synthetic data that contains both normal and anomalous sequences, we show that the proposed method achieves an accuracy close to 100% compared to the method that uses only the traditional LSTM autoencoder.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.2
• /
• pp.187-192
• /
• 2022

Even the announcing date for the staring date of "With Corona" has been decided, still many people have not completed vaccination, the most important condition for starting the With Corona, because of concerns for its side effects. In addition, although the economy may can be recovered by the With Corona, but the number of infected people may can be surged. In this paper, in order to awaken the people for the awareness of Corona 19 in advance of the With Corona, the Corona 19 is predicted through a non-linear probability process. Here, among the deep learning RNN, BI-LSTM, which is a bidirectional LSTM, and GRU, gates decreased than LSTM have been used. And this has been compared and analyzed through train set, test set, loss function, residual analysis, normal distribution, and autocorrelation, and compared and predicted for which has a better performance.