• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.141-144
• /
• 2020

한국어에서 문장의 가독성이나 맥락 파악을 위해 띄어쓰기는 매우 중요하다. 또한 자연 언어 처리를 할 때 띄어쓰기 오류가 있는 문장을 사용하면 문장의 구조가 달라지기 때문에 성능에 영향을 미칠 수 있다. 기존 연구에서는 N-gram 기반 통계적인 방법과 형태소 분석기를 이용하여 띄어쓰기 교정을 해왔다. 최근 들어 심층 신경망을 활용하는 많은 띄어쓰기 교정 연구가 진행되고 있다. 기존 심층 신경망을 이용한 연구에서는 문장을 음절 단위 또는 형태소 단위로 처리하여 교정 모델을 만들었다. 본 연구에서는 음절과 형태소 단위 모두 모델의 입력으로 사용하여 두 정보를 결합하여 띄어쓰기 교정 문제를 해결하고자 한다. 모델은 문장의 음절과 형태소 시퀀스에서 지역적 정보를 학습할 수 있는 Convolutional Neural Network와 순서정보를 정방향, 후방향으로 학습할 수 있는 Bidirectional Long Short-Term Memory 구조를 사용한다. 모델의 성능은 음절의 정확도와 어절의 정밀도, 어절의 재현율, 어절의 F1 score를 사용해 평가하였다. 제안한 모델의 성능 평가 결과 어절의 F1 score가 96.06%로 우수한 성능을 냈다.

• Journal of Intelligence and Information Systems
• /
• v.25 no.1
• /
• pp.163-177
• /
• 2019

As smartphones are getting widely used, human activity recognition (HAR) tasks for recognizing personal activities of smartphone users with multimodal data have been actively studied recently. The research area is expanding from the recognition of the simple body movement of an individual user to the recognition of low-level behavior and high-level behavior. However, HAR tasks for recognizing interaction behavior with other people, such as whether the user is accompanying or communicating with someone else, have gotten less attention so far. And previous research for recognizing interaction behavior has usually depended on audio, Bluetooth, and Wi-Fi sensors, which are vulnerable to privacy issues and require much time to collect enough data. Whereas physical sensors including accelerometer, magnetic field and gyroscope sensors are less vulnerable to privacy issues and can collect a large amount of data within a short time. In this paper, a method for detecting accompanying status based on deep learning model by only using multimodal physical sensor data, such as an accelerometer, magnetic field and gyroscope, was proposed. The accompanying status was defined as a redefinition of a part of the user interaction behavior, including whether the user is accompanying with an acquaintance at a close distance and the user is actively communicating with the acquaintance. A framework based on convolutional neural networks (CNN) and long short-term memory (LSTM) recurrent networks for classifying accompanying and conversation was proposed. First, a data preprocessing method which consists of time synchronization of multimodal data from different physical sensors, data normalization and sequence data generation was introduced. We applied the nearest interpolation to synchronize the time of collected data from different sensors. Normalization was performed for each x, y, z axis value of the sensor data, and the sequence data was generated according to the sliding window method. Then, the sequence data became the input for CNN, where feature maps representing local dependencies of the original sequence are extracted. The CNN consisted of 3 convolutional layers and did not have a pooling layer to maintain the temporal information of the sequence data. Next, LSTM recurrent networks received the feature maps, learned long-term dependencies from them and extracted features. The LSTM recurrent networks consisted of two layers, each with 128 cells. Finally, the extracted features were used for classification by softmax classifier. The loss function of the model was cross entropy function and the weights of the model were randomly initialized on a normal distribution with an average of 0 and a standard deviation of 0.1. The model was trained using adaptive moment estimation (ADAM) optimization algorithm and the mini batch size was set to 128. We applied dropout to input values of the LSTM recurrent networks to prevent overfitting. The initial learning rate was set to 0.001, and it decreased exponentially by 0.99 at the end of each epoch training. An Android smartphone application was developed and released to collect data. We collected smartphone data for a total of 18 subjects. Using the data, the model classified accompanying and conversation by 98.74% and 98.83% accuracy each. Both the F1 score and accuracy of the model were higher than the F1 score and accuracy of the majority vote classifier, support vector machine, and deep recurrent neural network. In the future research, we will focus on more rigorous multimodal sensor data synchronization methods that minimize the time stamp differences. In addition, we will further study transfer learning method that enables transfer of trained models tailored to the training data to the evaluation data that follows a different distribution. It is expected that a model capable of exhibiting robust recognition performance against changes in data that is not considered in the model learning stage will be obtained.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.13 no.2
• /
• pp.129-137
• /
• 2020

Currently, The autonomous vehicle market is researching and developing four-level autonomous vehicles beyond the commercialization of three-level autonomous vehicles. Because unlike the level 3, the level 4 autonomous vehicle has to deal with an emergency directly, the most important aspect of a four-level autonomous vehicle is its stability. In this paper, we propose an Optimized Vehicle Routing System (OVRS) that determines the route with the lowest probability of an accident at the destination of the vehicle rather than an immediate response in an emergency. The OVRS analyzes road and surrounding vehicle information collected by The RSU communication to predict road hazards, and sets the route for the safer and faster road. The OVRS can improve the stability of the vehicle by executing the route guidance according to the road situation through the RSU on the road like the network routing method. As a result, the RPNN of the ASICM, one of the OVRS modules, was about 17% better than the CNN and 40% better than the LSTM. However, because the study was conducted in a virtual environment using a PC, the possibility of accident of the VPDM was not actually verified. Therefore, in the future, experiments with high accuracy on VPDM due to the collection of accident data and actual roads should be conducted in real vehicles and RSUs.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.07a
• /
• pp.305-307
• /
• 2021

Alzheimer's Disease (AD) is a cognitive disorder characterized by memory impairment that can be assessed at early stages based on administering clinical tests. However, the AD pathophysiological mechanism is still poorly understood due to the difficulty of distinguishing different levels of AD severity, even using a variety of brain modalities. Therefore, in this study, we present a hybrid EEG-fNIRS modalities to compensate for each other's weaknesses with the help of Machine Learning (ML) techniques for classifying four subject groups, including healthy controls (HC) and three distinguishable groups of AD levels. A concurrent EEF-fNIRS setup was used to record the data from 41 subjects during Oddball and 1-back tasks. We employed both a traditional neural network (NN) and a CNN-LSTM hybrid model for fNIRS and EEG, respectively. The final prediction was then obtained by using majority voting of those models. Classification results indicated that the hybrid EEG-fNIRS feature set achieved a higher accuracy (71.4%) by combining their complementary properties, compared to using EEG (67.9%) or fNIRS alone (68.9%). These findings demonstrate the potential of an EEG-fNIRS hybridization technique coupled with ML-based approaches for further AD studies.

• The Journal of the Convergence on Culture Technology
• /
• v.9 no.6
• /
• pp.883-888
• /
• 2023

Since data centers are places that provide IT services 24 hours a day, 365 days a year, data center power consumption is expected to increase to approximately 10% by 2030, and the introduction of high-density IT equipment will gradually increase. In order to ensure the stable operation of IT equipment, various types of research are required to conserve energy in cooling and improve energy management. This study proposes the following process for energy saving in data centers. We conducted CFD modeling of the data center, proposed an artificial intelligence-based thermal environment prediction model, compared actual measured data, the predicted model, and the CFD results, and finally evaluated the data center's thermal management performance. It can be seen that the predicted values of RCI, RTI, and PUE are also similar according to the normalization used in the normalization method. Therefore, it is judged that the algorithm proposed in this study can be applied and provided as a thermal environment prediction model applied to data centers.

• KNOM Review
• /
• v.23 no.2
• /
• pp.18-28
• /
• 2020

With the development of the virtual community, the benefits that IT technology provides to people in fields such as healthcare, industry, communication, and culture are increasing, and the quality of life is also improving. Accordingly, there are various malicious attacks targeting the developed network environment. Firewalls and intrusion detection systems exist to detect these attacks in advance, but there is a limit to detecting malicious attacks that are evolving day by day. In order to solve this problem, intrusion detection research using machine learning is being actively conducted, but false positives and false negatives are occurring due to imbalance of the learning dataset. In this paper, a Random Oversampling method is used to solve the unbalance problem of the UNSW-NB15 dataset used for network intrusion detection. And through experiments, we compared and analyzed the accuracy, precision, recall, F1-score, training and prediction time, and hardware resource consumption of the models. Based on this study using the Random Oversampling method, we develop a more efficient network intrusion detection model study using other methods and high-performance models that can solve the unbalanced data problem.

• Journal of Intelligence and Information Systems
• /
• v.25 no.4
• /
• pp.105-122
• /
• 2019

Dimensionality reduction is one of the methods to handle big data in text mining. For dimensionality reduction, we should consider the density of data, which has a significant influence on the performance of sentence classification. It requires lots of computations for data of higher dimensions. Eventually, it can cause lots of computational cost and overfitting in the model. Thus, the dimension reduction process is necessary to improve the performance of the model. Diverse methods have been proposed from only lessening the noise of data like misspelling or informal text to including semantic and syntactic information. On top of it, the expression and selection of the text features have impacts on the performance of the classifier for sentence classification, which is one of the fields of Natural Language Processing. The common goal of dimension reduction is to find latent space that is representative of raw data from observation space. Existing methods utilize various algorithms for dimensionality reduction, such as feature extraction and feature selection. In addition to these algorithms, word embeddings, learning low-dimensional vector space representations of words, that can capture semantic and syntactic information from data are also utilized. For improving performance, recent studies have suggested methods that the word dictionary is modified according to the positive and negative score of pre-defined words. The basic idea of this study is that similar words have similar vector representations. Once the feature selection algorithm selects the words that are not important, we thought the words that are similar to the selected words also have no impacts on sentence classification. This study proposes two ways to achieve more accurate classification that conduct selective word elimination under specific regulations and construct word embedding based on Word2Vec embedding. To select words having low importance from the text, we use information gain algorithm to measure the importance and cosine similarity to search for similar words. First, we eliminate words that have comparatively low information gain values from the raw text and form word embedding. Second, we select words additionally that are similar to the words that have a low level of information gain values and make word embedding. In the end, these filtered text and word embedding apply to the deep learning models; Convolutional Neural Network and Attention-Based Bidirectional LSTM. This study uses customer reviews on Kindle in Amazon.com, IMDB, and Yelp as datasets, and classify each data using the deep learning models. The reviews got more than five helpful votes, and the ratio of helpful votes was over 70% classified as helpful reviews. Also, Yelp only shows the number of helpful votes. We extracted 100,000 reviews which got more than five helpful votes using a random sampling method among 750,000 reviews. The minimal preprocessing was executed to each dataset, such as removing numbers and special characters from text data. To evaluate the proposed methods, we compared the performances of Word2Vec and GloVe word embeddings, which used all the words. We showed that one of the proposed methods is better than the embeddings with all the words. By removing unimportant words, we can get better performance. However, if we removed too many words, it showed that the performance was lowered. For future research, it is required to consider diverse ways of preprocessing and the in-depth analysis for the co-occurrence of words to measure similarity values among words. Also, we only applied the proposed method with Word2Vec. Other embedding methods such as GloVe, fastText, ELMo can be applied with the proposed methods, and it is possible to identify the possible combinations between word embedding methods and elimination methods.