• Journal of Digital Convergence
• /
• v.19 no.10
• /
• pp.253-263
• /
• 2021

Machine learning is the process of constructing a cost function using learning data used for learning and an artificial neural network to predict the data, and finding parameters that minimize the cost function. Parameters are changed by using the gradient-based method of the cost function. The more complex the digital signal and the more complex the problem to be learned, the more complex and deeper the structure of the artificial neural network. Such a complex and deep neural network structure can cause over-fitting problems. In order to avoid over-fitting, a weight decay regularization method of parameters is used. We additionally use the value of the cost function in this method. In this way, the accuracy of machine learning is improved, and the superiority is confirmed through numerical experiments. These results derive accurate values for a wide range of artificial intelligence data through machine learning.

• Korean Chemical Engineering Research
• /
• v.59 no.4
• /
• pp.532-541
• /
• 2021

Recently, research on the application of artificial intelligence in the chemical process has been increasing rapidly. However, overfitting is a significant problem that prevents the model from being generalized well to predict unseen data on test data, as well as observed training data. Cross validation is one of the ways to solve the overfitting problem. In this study, the time-series cross validation method was applied to optimize the number of batch and epoch in the hyperparameters of the prediction model for the 2,3-BDO distillation process, and it compared with K-fold cross validation generally used. As a result, the RMSE of the model with time-series cross validation was lower by 9.06%, and the MAPE was higher by 0.61% than the model with K-fold cross validation. Also, the calculation time was 198.29 sec less than the K-fold cross validation method.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.2
• /
• pp.148-154
• /
• 2021

Text-Independent speaker verification needs to extract text-independent speaker embedding to improve generalization performance. However, deep neural networks that depend on training data have the potential to overfit text information instead of learning the speaker information when repeatedly learning from the identical time series. In this paper, to prevent the overfitting, we propose a segment unit shuffling layer that divides and rearranges the input layer or a hidden layer along the time axis, thus mixes the time series information. Since the segment unit shuffling layer can be applied not only to the input layer but also to the hidden layers, it can be used as generalization technique in the hidden layer, which is known to be effective compared to the generalization technique in the input layer, and can be applied simultaneously with data augmentation. In addition, the degree of distortion can be adjusted by adjusting the unit size of the segment. We observe that the performance of text-independent speaker verification is improved compared to the baseline when the proposed segment unit shuffling layer is applied.

• Journal of Broadcast Engineering
• /
• v.28 no.3
• /
• pp.267-274
• /
• 2023

There is a problem that the size of the dataset is insufficient due to the limitation of the collection of the medical image public data, so there is a possibility that the existing studies are overfitted to the public dataset. In this paper, we compare the performance of eight (Unet, X-Net, HarDNet, SegNet, PSPNet, SwinUnet, 3D-ResU-Net, UNETR) medical image semantic segmentation models to revalidate the superiority of existing models. Anatomical Tracings of Lesions After Stroke (ATLAS) V1.2, a public dataset for stroke diagnosis, is used to compare the performance of the models and the performance of the models in ATLAS V2.0. Experimental results show that most models have similar performance in V1.2 and V2.0, but X-net and 3D-ResU-Net have higher performance in V1.2 datasets. These results can be interpreted that the models may be overfitted to V1.2.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.2
• /
• pp.235-241
• /
• 2022

Several studies using Convolutional Neural Network (CNN) have been conducted to detect and classify the sounds of marine mammals in underwater acoustic data collected through passive acoustic monitoring. In this study, the possibility of automatic classification of bearded seal sounds was confirmed using a CNN model based on the underwater acoustic spectrogram images collected from August 2017 to August 2018 in East Siberian Sea. When only the clear seal sound was used as training dataset, overfitting due to memorization was occurred. By evaluating the entire training data by replacing some training data with data containing noise, it was confirmed that overfitting was prevented as the model was generalized more than before with accuracy (0.9743), precision (0.9783), recall (0.9520). As a result, the performance of the classification model for bearded seals signal has improved when the noise was included in the training data.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.10a
• /
• pp.492-494
• /
• 2021

The decision tree technique is used as a classification technique in machine learning. However, the decision tree has a problem of consuming a lot of speed or resources due to the problem of overfitting. To solve this problem, there are bagging and boosting techniques. Bagging creates multiple samplings and models them using them, and boosting models the sampled data and adjusts weights to reduce overfitting. In addition, recently, techniques Xgboost have been introduced to improve performance. Therefore, in this paper, we collect wifi signal data for indoor positioning, apply it to the existing method and Xgboost, and perform performance evaluation through it.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2017.06a
• /
• pp.145-146
• /
• 2017

본 논문에서는 남성의 하반신을 촬영한 MRI 영상으로부터 전립선을 분할하는 알고리즘을 제안한다. 우선 3 차원 입체 영상을 학습하기 위해 3D 컨볼루션 계층(convolutional layer) 및 3D 풀링 계층(pooling layer)에 기반한 네트워크를 제안한다. 다음으로 네트워크의 최후단에 해당하는 전연결 계층(fully connected layer)의 강인한 학습을 돕는 잡음 계층을 제안한다. 잡음 계층은 네트워크의 학습 파라미터 혹은 출력 영상에 가우시안 잡음를 더함으로써 드롭 아웃과 같이 훈련 영상에 대한 과적합(overfitting)을 막고 테스트 영상에 강인한 네트워크의 학습을 돕는다. 마지막으로 실험을 통해 제안하는 기법이 기존 기법에 비해 우수한 분할 성능을 보임을 확인한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.10b
• /
• pp.307-309
• /
• 2001

분류문제에서 유용한 학습패턴은 클래스들간의 분류경계에 근접한 정상패턴들을 말한다. 본 연구에서는 다양한 구조와 학습 파라미터를 가진 신경망 앙상블을 구성하고 그 출력값의 편기와 분산에 기초한 패턴절수를 정의한다. 전체 학습패턴 중 일정한 임계값 이상의 패턴점수를 가진 패턴들만이 학습패턴으로 선정된다. 제안한 방법은 두 개의 인공문제와 두 개의 실제문제 (UCI Repository)에 적응, 검증되었다. 그 결과 선택된 패턴만으로 학습한 경우, 메모리 공간 절약 및 계산시간 단축의 효과뿐만 아니라 복잡도가 큰 모델이라도 과적합을 하지 않았고 실험적으로 안정된 결과를 산출했으며, 적은 수의 학습패턴만으로도 일반화 성능을 향상시키거나 적어도 저하시키지 않았다는 것을 보였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2017.11a
• /
• pp.139-141
• /
• 2017

화재는 다른 재난보다 확산 속도가 빠르기 때문에 신속하고 정확한 감지와 지속적인 감시가 요구된다. 최근, 신속하고 정확한 화재 감지를 위해, CCTV(Closed-Circuit TeleVision)으로 획득한 이미지를 기계학습(Machine Learning)을 이용해 화재 발생 여부를 감지하는 화재 감지 시스템이 주목받고 있다. 본 논문에서는 기계학습의 기술 중 정확도가 가장 높은 딥러닝(Deep Learning)기반의 CCTV 화재 감지 시스템을 제안한다. 본 논문의 시스템은 딥러닝 기술 적용뿐만이 아니라, CCTV 이미지 전처리 과정을 보완함으로써 딥러닝에서의 미지 데이터(unseen data)의 낮은 분류 정확도 문제인 과적합(overfitting)문제를 해결하였다. 본 논문의 시스템은 약 80,000 개의 CCTV 이미지 데이터를 학습하여, 90% 이상의 화재 이미지 분류 정확도의 성능을 보여주었다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2018.07a
• /
• pp.33-34
• /
• 2018

기계학습이란 인공지능의 한 분야이다. 컴퓨터에 명시적인 프로그램 없이 배울 수 있는 능력을 부여하는 연구 분야이며, 사람이 학습하듯이 컴퓨터에도 데이터들을 줘서 학습하게 함으로써 새로운 지식을 얻어내게 하는 분야이다. 기계학습 종류에는 크게 Supervised Learning, Unsupervised Learning, Reinforcement Learning이 있다. 본 논문에서는 기계학습 종류 및 컴퓨터가 데이터들을 학습하면서 생기는 문제점을 알아보고, 문제점의 종류 및 해결방안을 제시한다.