• 한국인터넷방송통신학회논문지
• /
• 제20권1호
• /
• pp.11-18
• /
• 2020

본 연구에서는 모바일 환경에서의 기존 맥락인증기법의 부족한 성능을 보완하고자 한다. 사용된 데이터는 GPS, CDR(Call Detail Record), App usage이며 GPS의 처리과정에서 인구밀집지역의 타인을 세밀하게 구분하고자 GPS밀도에 따른 지역구분을 시행하였다. 또한 전처리에서 데이터 수집에서 발생할 수 있는 결측치를 처리한다. 인증 모델은 두 개의 LSTM(Long-Short Term Memory)와 그들 결과를 종합하는 하나의 ANN(Artificial Neural Network)로 구성하며 이를 통해 최종적으로 인증 점수를 산출한다. 본 논문에서는 기존 연구와의 정확도를 비교하고 타인을 구별해내는데 필요한 인증 시도 횟수를 비교하여 평균 11.6%의 정확도 향상과 검증 데이터의 약 60%에 대하여 더 적은 시도에 구별해 낼 수 있었다.

• International Journal of Contents
• /
• 제17권4호
• /
• pp.1-15
• /
• 2021

Named entity recognition (NER) is a crucial task for NLP, which aims to extract information from texts. To build NER systems, deep learning (DL) models are learned with dictionary features by mapping each word in the dataset to dictionary features and generating a unique index. However, this technique might generate noisy labels, which pose significant challenges for the NER task. In this paper, we proposed DL-dictionary features, and evaluated them on two datasets, including the OntoNotes 5.0 dataset and our new infectious disease outbreak dataset named GFID. We used (1) a Bidirectional Long Short-Term Memory (BiLSTM) character and (2) pre-trained embedding to concatenate with (3) our proposed features, named the Convolutional Neural Network (CNN), BiLSTM, and self-attention dictionaries, respectively. The combined features (1-3) were fed through BiLSTM - Conditional Random Field (CRF) to predict named entity classes as outputs. We compared these outputs with other predictions of the BiLSTM character, pre-trained embedding, and dictionary features from previous research, which used the exact matching and partial matching dictionary technique. The findings showed that the model employing our dictionary features outperformed other models that used existing dictionary features. We also computed the F1 score with the GFID dataset to apply this technique to extract medical or healthcare information.

• 한국수자원학회논문집
• /
• 제54권spc1호
• /
• pp.1107-1118
• /
• 2021

지구온난화로 인해 발생한 기후변화는 한반도의 홍수, 가뭄 등의 발생빈도를 증가시켰으며, 이로 인해 인적, 물적 피해가 증가한 것으로 나타났다. 수재해 대비 및 대응을 위해서는 국가 차원의 수자원 관리 계획 수립이 필요하며, 유역 단위 수자원 관리를 위해서는 장기간 관측된 유량 자료를 이용하여 도출된 유량지속곡선이 필요하다. 전통적으로 수자원 분야에서 유량지속곡선을 도출하기 위하여 물리적 기반의 강우-유출 모형이 많이 사용되고 있으며, 최근에는 데이터 기반의 딥러닝 기법을 이용한 유출량 예측 기법에 관한 연구가 진행된 바 있다. 물리적 기반의 모형은 수문학적으로 신뢰도 높은 결과를 도출할 수 있으나, 사용자의 높은 이해도가 요구되며, 모형 구동 시간이 오래 걸릴 수 있는 단점이 있다. 데이터 기반의 딥러닝 기법의 경우 입력 자료가 간단하며, 모형 구동 시간이 비교적 짧으나 입력 및 출력자료 간의 관계가 블랙박스로 처리되어 수리·수문학적 특성을 반영할 수 없는 단점이 있다. 본 연구에서는 물리적 기반 모형으로 국내외에서 적용성이 검증된 Soil Water Assessment Tool (SWAT)의 매개변수 보정(Calibration)을 통해 장기간의 결측치 없는 데이터를 산출하고, 이를 데이터 기반 딥러닝 기법인 Long Short-term Memory (LSTM)의 훈련(Training) 데이터로 활용하였다. 시계열 데이터 분석 결과 검·보정 전체 기간('07-'18) 동안 Nash-Sutcliffe Efficiency (NSE)와 적합도 비교를 위한 결정계수는 각각 0.04, 0.03 높게 도출되어 모형에서 도출된 SWAT의 결과가 LSTM보다 전반적으로 우수한 것으로 나타났다. 또한, 모형에서 도출된 연도별 시계열 자료를 내림차순하여 산정된 유량지속곡선과 관측유량 기반의 유량지속곡선과 비교한 결과 NSE는 SWAT과 LSTM 각각 0.95, 0.91로 나타났으며, 결정계수는 0.96, 0.92로 두 모형 모두 우수한 성능을 보였다. LSTM 모형의 경우 저유량 부분 모의의 정확도 개선이 필요하나, 방대한 입력 자료로 인해 모형 구축 및 구동 시간이 오래 걸리는 대유역과 입력 자료가 부족한 미계측 유역의 유량지속곡선 산정 등에 활용성이 높을 것으로 판단된다.

• 한국산업융합학회 논문집
• /
• 제25권2_1호
• /
• pp.169-176
• /
• 2022

Prognostics and health management (PHM) is recently converging throughout the industry, one of the trending issue is to detect abnormal conditions at decanter centrifuge during water treatment facilities. Wastewater treatment operation produces corrosive gas which results failures on attached sensors. This scenario causes frequent sensor replacement and requires highly qualified manager's visual inspection while replacing important parts such as bearings and screws. In this paper, we propose anomaly detection by measuring the vibration of the decanter centrifuge based on the video camera images. Measuring the vibration of the screw decanter by applying the optical flow technique, the amount of movement change of the corresponding pixel is measured and fed into the LST M model. As a result, it is possible to detect the normal/warning/dangerous state based on LSTM classification. In the future work, we aim to gather more abnormal data in order to increase the further accuracy so that it can be utilized in the field of industry.

• Journal of Information Processing Systems
• /
• 제16권6호
• /
• pp.1407-1423
• /
• 2020

A graph is a data structure consisting of nodes and edges between these nodes. Graph embedding is to generate a low dimensional vector for a given graph that best represents the characteristics of the graph. Recently, there have been studies on graph embedding, especially using deep learning techniques. However, until now, most deep learning-based graph embedding techniques have focused on unweighted graphs. Therefore, in this paper, we propose a graph embedding technique for weighted graphs based on long short-term memory (LSTM) autoencoders. Given weighted graphs, we traverse each graph to extract node-weight sequences from the graph. Each node-weight sequence represents a path in the graph consisting of nodes and the weights between these nodes. We then train an LSTM autoencoder on the extracted node-weight sequences and encode each nodeweight sequence into a fixed-length vector using the trained LSTM autoencoder. Finally, for each graph, we collect the encoding vectors obtained from the graph and combine them to generate the final embedding vector for the graph. These embedding vectors can be used to classify weighted graphs or to search for similar weighted graphs. The experiments on synthetic and real datasets show that the proposed method is effective in measuring the similarity between weighted graphs.

• 시스템엔지니어링학술지
• /
• 제18권2호
• /
• pp.75-93
• /
• 2022

Machine learning (ML) data-driven meta-model is proposed as a surrogate model to reduce the excessive computational cost of the physics-based model and facilitate the real-time prediction of a nuclear power plant's transient response. To forecast the transient response three machine learning (ML) meta-models based on recurrent neural networks (RNNs); specifically, Long Short Term Memory (LSTM), Gated Recurrent Unit (GRU), and a sequence combination of Convolutional Neural Network (CNN) and LSTM are developed. The chosen accident scenario is a control element assembly withdrawal at power concurrent with the Loss Of Offsite Power (LOOP). The transient response was obtained using the best estimate thermal hydraulics code, MARS-KS, and cross-validated against the Design and control document (DCD). DAKOTA software is loosely coupled with MARS-KS code via a python interface to perform the Best Estimate Plus Uncertainty Quantification (BEPU) analysis and generate a time series database of the system response to train, test and validate the ML meta-models. Key uncertain parameters identified as required by the CASU methodology were propagated using the non-parametric Monte-Carlo (MC) random propagation and Latin Hypercube Sampling technique until a statistically significant database (181 samples) as required by Wilk's fifth order is achieved with 95% probability and 95% confidence level. The three ML RNN models were built and optimized with the help of the Talos tool and demonstrated excellent performance in forecasting the most probable NPP transient response. This research was guided by the Systems Engineering (SE) approach for the systematic and efficient planning and execution of the research.

• International Journal of Contents
• /
• 제18권2호
• /
• pp.68-80
• /
• 2022

The health of the human heart is commonly measured using ECG (Electrocardiography) signals. To identify any anomaly in the human heart, the time-sequence of ECG signals is examined manually by a cardiologist or cardiac electrophysiologist. Lightweight anomaly detection on ECG signals in an embedded system is expected to be popular in the near future, because of the increasing number of heart disease symptoms. Some previous research uses deep learning networks such as LSTM and BiLSTM to detect anomaly signals without any handcrafted feature. Unfortunately, lightweight LSTMs show low precision and heavy LSTMs require heavy computing powers and volumes of labeled dataset for symptom classification. This paper proposes an ECG anomaly detection system based on two level BiLSTM for acceptable precision with lightweight networks, which is lightweight and usable at home. Also, this paper presents a new threshold technique which considers statistics of the current ECG pattern. This paper's proposed model with BiLSTM detects ECG signal anomaly in 0.467 ~ 1.0 F₁ score, compared to 0.426 ~ 0.978 F₁ score of the similar model with LSTM except one highly noisy dataset.

• 한국컴퓨터정보학회논문지
• /
• 제29권8호
• /
• pp.11-21
• /
• 2024

본 연구는 가상현실 (Virtual Reality, VR) 기술을 활용하여 골프 스윙의 주요 구간을 식별하고 레이블링 (Labeling) 하는 방법을 탐구한다. 기존 가상현실 기기의 제한점을 해결하기 위해 OpenVR SDK (Software Development Kit)와 SteamVR을 활용하여 다양한 VR 기기에서 운동 데이터를 수집할 수 있는 플랫폼을 개발하였다. 또한, 장단기 메모리 (Long Short-Term Memory) 기반의 시계열 데이터 분석을 통해 운동 동작의 시간적 변화를 식별하고 레이블링하는 반자동 레이블링 기술을 개발하였다. 실험은 소년, 청년, 중년, 장년 세대별 각 20명씩 총 80명의 참가자가 각 5회의 스윙 데이터를 수집하여 총 400개의 운동 데이터 세트를 구축하였다. 제안하는 기술은 골프 스윙의 7가지 주요 구간에 대해 모든 연령대에서 일관되게 높은 정확도(0.94 이상)와 F1 점수(0.95 이상)를 달성하였다. 이 기술은 운동 데이터를 세분화하고 구간별로 운동 능력을 정밀하게 평가할 수 있는 기반을 마련하여, 향후 교육 및 훈련 과정에서 개별 사용자에게 맞춤형 피드백을 제공하는 데 목적이 있다.

• International Journal of Advanced Culture Technology
• /
• 제9권4호
• /
• pp.268-273
• /
• 2021

In this paper, we propose a sentiment analysis model that improves performance on small-scale data. A sentiment analysis model for small-scale data is proposed and verified through experiments. To this end, we propose Bagging-Bi-GRU, which combines Bi-GRU, which learns GRU, which is a variant of LSTM (Long Short-Term Memory) with excellent performance on sequential data, in both directions and the bagging technique, which is one of the ensembles learning methods. In order to verify the performance of the proposed model, it is applied to small-scale data and large-scale data. And by comparing and analyzing it with the existing machine learning algorithm, Bi-GRU, it shows that the performance of the proposed model is improved not only for small data but also for large data.

• ETRI Journal
• /
• 제46권2호
• /
• pp.205-217
• /
• 2024

We propose a bandwidth prediction approach based on deep learning. The approach is intended to accurately predict the bandwidth of various types of mobile networks. We first use a machine learning technique, namely, the gradient boosting algorithm, to recognize the connected mobile network. Second, we apply a handover detection algorithm based on network recognition to account for vertical handover that causes the bandwidth variance. Third, as the communication performance offered by 3G, 4G, and 5G networks varies, we suggest a bidirectional long short-term memory model with time distribution for bandwidth prediction per network. To increase the prediction accuracy, pretraining and fine-tuning are applied for each type of network. We use a dataset collected at University College Cork for network recognition, handover detection, and bandwidth prediction. The performance evaluation indicates that the handover detection algorithm achieves 88.5% accuracy, and the bandwidth prediction model achieves a high accuracy, with a root-mean-square error of only 2.12%.