• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.8
• /
• pp.355-364
• /
• 2023

As advanced cyber threats continue to increase in recent years, it is difficult to detect new types of cyber attacks with existing pattern or signature-based intrusion detection method. Therefore, research on anomaly detection methods using data learning-based artificial intelligence technology is increasing. In addition, supervised learning-based anomaly detection methods are difficult to use in real environments because they require sufficient labeled data for learning. Research on an unsupervised learning-based method that learns from normal data and detects an anomaly by finding a pattern in the data itself has been actively conducted. Therefore, this study aims to extract a latent vector that preserves useful sequence information from sequence log data and develop an anomaly detection learning model using the extracted latent vector. Word2Vec was used to create a dense vector representation corresponding to the characteristics of each sequence, and an unsupervised autoencoder was developed to extract latent vectors from sequence data expressed as dense vectors. The developed autoencoder model is a recurrent neural network GRU (Gated Recurrent Unit) based denoising autoencoder suitable for sequence data, a one-dimensional convolutional neural network-based autoencoder to solve the limited short-term memory problem that GRU can have, and an autoencoder combining GRU and one-dimensional convolution was used. The data used in the experiment is time-series-based NGIDS (Next Generation IDS Dataset) data, and as a result of the experiment, an autoencoder that combines GRU and one-dimensional convolution is better than a model using a GRU-based autoencoder or a one-dimensional convolution-based autoencoder. It was efficient in terms of learning time for extracting useful latent patterns from training data, and showed stable performance with smaller fluctuations in anomaly detection performance.

• The Korean Journal of Applied Statistics
• /
• v.37 no.2
• /
• pp.209-224
• /
• 2024

Speaker verification is becoming popular as a method of non-face-to-face identity authentication. It involves determining whether two voice data belong to the same speaker. In cases where the criminal's voice remains at the crime scene, it is vital to establish a speaker verification system that can accurately compare the two voice evidence. In this study, to achieve this, a new speaker verification system was built using a deep learning model for Korean language. High-dimensional voice data with a high variability like background noise made it necessary to use deep learning-based methods for speaker matching. To construct the matching algorithm, the ECAPA-TDNN model, known as the most famous deep learning system for speaker verification, was selected. A large dataset of the voice data, Voxceleb, collected from people of various nationalities without Korean. To study the appropriate form of datasets necessary for learning the Korean language, experiments were carried out to find out how Korean voice data affects the matching performance. The results showed that when comparing models learned only with Voxceleb and models learned with datasets combining Voxceleb and Korean datasets to maximize language and speaker diversity, the performance of learning data, including Korean, is improved for all test sets.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.6
• /
• pp.736-741
• /
• 2006

The design of a classification system generally involves data acquisition module, learning module and decision module, considering their functions and it is often an important component of intelligent systems. The learning module provides a priori information and it has been playing a key role for the classification. The conventional learning techniques for classification are based on a winner take all fashion which does not reflect the description of real data where boundarues might be fuzzy Moreover they need all data for the learning of its problem domain. Generally, in many practical applications, it is not possible to prepare them at a time. In this paper, we design an adaptive classification model using incremental training fuzzy neural networks, FNN-I. To have a more useful information, it introduces the representation and membership degree by fuzzy theory. And it provides an incremental learning algorithm for continuously gathered data. We present tie experimental results on computer virus data. They show that the proposed system can learn incrementally and classify new viruses effectively.

• KIPS Transactions on Software and Data Engineering
• /
• v.6 no.3
• /
• pp.149-154
• /
• 2017

The MCT-Adaboost algorithm chooses an optimal set of features in each rounds. On each round, it chooses the best feature by calculate minimizing error rate using feature index and MCT kernel distribution. The involved process of weak classification executed by a hard decision. This decision occurs some problems when it chooses ambiguous kernel feature. In this paper, we propose the modified MCT-Adaboost classification using soft decision. The typical MCT-Adaboost assigns a same initial weights to each datum. This is because, they assume that all information of database is blind. We assign different initial weights with our propose new algorithm using some statistical properties of involved features. In experimental results, we confirm that our method shows better performance than the traditional one.

• Information and Communications Magazine
• /
• v.33 no.5
• /
• pp.48-54
• /
• 2016

우리는 물리적인 현실 세계와 디지털의 가상 세계에서 매일 끊임없이 데이터를 양산해내고 있다. 구글, 아마존, MS, IBM 등의 유수 기업들은 이미 데이터를 수집하고 분석하여 특정 사용자나 불특정 다수에게 다양한 서비스를 제공하면서 새로운 형태의 이윤을 창출하고 있다. 가까운 미래에 사물인터넷(Internet of Things)이 본격적으로 활성화된다면 사람뿐만 아니라 모든 사물들이 인터넷을 통해 데이터를 양산하고 서로 교환하는 그야말로 데이터 빅뱅의 시대가 도래할 것으로 예상된다. 이러한 변혁의 시대에 우리는 사물인터넷을 통해 수집되는 수많은 데이터를 어떻게 활용할 것인지에 대해 진지하게 고민하고 연구할 필요가 있다. 본고에서는 사물인터넷을 통해 수집된 데이터를 효과적으로 활용하기 위해 필요한 핵심기술 중 하나인 기계학습(Machine Learning)에 대해 기본 개념, 종류, 평가방법 등을 설명하고 기계학습 알고리즘 중 딥 러닝(Deep Learning)에 대한 기술 동향을 살펴본 후, 사물인터넷에서 기계학습 프레임워크에 대해 간략히 소개한다.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2005.04a
• /
• pp.428-432
• /
• 2005

영상 데이터와 같은 대용량의 데이터를 분류하고자 할 경우, 입력 데이터의 차원을 줄여서 특징 벡터를 뽑아내는 전처리 과정은 필수적이다. 이 경우 특징 벡터가 입력 데이터의 정보를 최대한 포함하도록 하는 것이 중요하다. 특징 벡터를 뽑는 대표적인 방법으로는 PCA, ICA, LDA, MLP와 같은 특징 추출(feature extraction) 방법을 들 수 있다. PCA와 LDA는 무감독 학습 방식이고, LDA, MLP는 감독 학습 방식에 해당한다. 감독학습 방식의 경우 입력 정보와 함께 클래스 정보를 사용하기 때문에 데이터를 분류하기에 더 좋은 특징들을 뽑아낼 수 있는 장점이 있다. 본 논문에서는 무감독 학습 방식인 PCA에 클래스에 대한 정보를 함께 사용하여 특징을 추출함으로써 데이터 분류에 더욱 적합한 특징들을 뽑는 방법을 제안하였다. 그리고, Yale face database를 사용하여 제안한 알고리즘의 성능을 기존의 알고리즘과 비교, 테스트하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.712-715
• /
• 2023

딥러닝 기술의 발전은 고품질의 대규모 데이터에 크게 의존한다. 그러나, 데이터의 품질과 일관성을 유지하는 것은 상당한 비용과 시간이 소요된다. 이러한 문제를 해결하기 위해 최근 연구에서 최소한의 비용으로 최대의 성능을 추구하는 액티브 러닝(active learning) 기법이 주목받고 있는데, 액티브 러닝은 모델 관점에서 불확실성(uncertainty)이 높은 데이터들을 샘플링 하는데 중점을 둔다. 하지만, 레이블 생성에 있어서 여전히 많은 시간적, 자원적 비용이 불가피한 점을 고려할 때 보완이 불가피 하다. 본 논문에서는 의사-라벨링(pseudo labeling)을 활용한 준지도학습(semi-supervised learning) 방식과 학습 손실을 동시에 사용하여 모델의 불확실성(uncertainty)을 측정하는 방법론을 제안한다. 제안 방식은 레이블의 신뢰도(confidence)와 학습 손실의 최적화를 통해 비용 효율적인 데이터 레이블 생성 방식을 제안한다. 특히, 레이블 데이터의 품질(quality) 및 일관성(consistency) 측면에서 딥러닝 모델의 정확도 성능을 높임과 동시에 적은 데이터만으로도 효과적인 학습이 가능할 수 있는 메커니즘을 제안한다.

• Journal of Intelligence and Information Systems
• /
• v.28 no.4
• /
• pp.27-40
• /
• 2022

Medical AI, which has lately made significant advances, is playing a vital role, such as assisting clinicians with diagnosis and decision-making. The field of chest X-rays, in particular, is attracting a lot of attention since it is important for accessibility and identification of chest diseases, as well as the current COVID-19 pandemic. However, despite the vast amount of data, there remains a limit to developing an effective AI model due to a lack of labeled data. A research that used federated learning on chest X-ray data to lessen this difficulty has emerged, although it still has the following limitations. 1) It does not consider the problems that may occur in the Non-IID environment. 2) Even in the federated learning environment, there is still a shortage of labeled data of clients. We propose a method to solve the above problems by using the self-supervised learning model as a global model of federated learning. To that aim, we investigate a self-supervised learning methods suited for federated learning using chest X-ray data and demonstrate the benefits of adopting the self-supervised learning model for federated learning.

• Smart Media Journal
• /
• v.12 no.1
• /
• pp.24-31
• /
• 2023

Meta-learning is metacognition that instantly distinguishes between knowing and unknown. It is a learning method that adapts and solves new problems by self-learning with a small amount of data.A few-shot learning method is a type of meta-learning method that accurately predicts query data even with a very small support set. In this study, we propose a method to solve the limitations of the prototype created with the mean-point vector of each class. For this purpose, we use the few-shot learning method that created the prototype used in the few-shot learning method as the median prototype. For quantitative evaluation, a handwriting recognition dataset and mini-Imagenet dataset were used and compared with the existing method. Through the experimental results, it was confirmed that the performance was improved compared to the existing method.

• Journal of the Korean Society for Library and Information Science
• /
• v.54 no.1
• /
• pp.175-195
• /
• 2020

The purpose of this study is to propose a data curation service to support teaching and learning activities of university libraries. To this end, the concept of data curation has been summarized, and how to understand information services and teaching-learning information data curation of university libraries. As a traditional extension of library information services, the use of university library data curation services has been proposed. As a result of research, data curation suggests that it is appropriate to understand the data by focusing more on the utilization of data. The university library emphasized that the data curation service is needed through the establishment of an institutional repository of teaching-learning activity materials for the visible effects that contribute to the educational service of the parent university. Finally, the research was completed by presenting five levels of teaching and learning information data curation framework.