• Communications for Statistical Applications and Methods
• /
• v.30 no.6
• /
• pp.589-603
• /
• 2023

Deep generative models target to infer the underlying true data distribution, and it leads to a huge success in generating fake-but-realistic data. Regarding such a perspective, the data attributes can be a crucial factor in the data generation process since non-existent counterfactual samples can be generated by altering certain factors. For example, we can generate new portrait images by flipping the gender attribute or altering the hair color attributes. This paper proposes counterfactual disentangled variational autoencoder generative adversarial networks (CDVAE-GAN), specialized for data attribute level counterfactual data generation. The structure of the proposed CDVAE-GAN consists of variational autoencoders and generative adversarial networks. Specifically, we adopt a Gaussian variational autoencoder to extract low-dimensional disentangled data features and auxiliary Bernoulli latent variables to model the data attributes separately. Also, we utilize a generative adversarial network to generate data with high fidelity. By enjoying the benefits of the variational autoencoder with the additional Bernoulli latent variables and the generative adversarial network, the proposed CDVAE-GAN can control the data attributes, and it enables producing counterfactual data. Our experimental result on the CelebA dataset qualitatively shows that the generated samples from CDVAE-GAN are realistic. Also, the quantitative results support that the proposed model can produce data that can deceive other machine learning classifiers with the altered data attributes.

• Advanced Industrial SCIence
• /
• v.3 no.1
• /
• pp.17-22
• /
• 2024

History, which began with the emergence of mankind, has a means of recording. Today, we can check the past through data. Generated data may only be generated and stored at a certain moment, but it is not only continuously generated over a certain time interval from the past to the present, but also occurs in the future, so making predictions using it is an important task. In order to find out trends in the use of time series data among numerous data, this paper analyzes the concept of time series data, analyzes Recurrent Neural Network and Long-Short Term Memory, which are mainly used for time series data analysis in the machine learning field, and analyzes the use of these models. Through case studies, it was confirmed that it is being used in various fields such as medical diagnosis, stock price analysis, and climate prediction, and is showing high predictive results. Based on this, we will explore ways to utilize it in the future.

• The Journal of Society for e-Business Studies
• /
• v.21 no.1
• /
• pp.119-129
• /
• 2016

With the development of web, amount of information are generated in social web. Then many researchers are focused on the extracting and analyzing social issues from various social data. The proposed approach performed gathering the science data and analyzing with LDA algorithm. It generated the clusters which represent the social topics related to 'health'. As a result, we could deduce the relationship between science data and social issues.

• Proceedings of the KOSOMBE Conference
• /
• v.1996 no.05
• /
• pp.27-30
• /
• 1996

PR interval variability has been proposed as a noninvasive tool for in-vestigating the autonomic nervous system as welt as heart rate variability. The goal of this paper is to determine whether PR interval variability is generated from deterministic nonlinear dynamics. The data used in this study is a 24-hour bolter ECGs of 20 healthy adults. We developed an automatic PR interval measurement algorithm, and tested it using MIT ECG Databases. The general discriminants of nonlinear dynamics, such as, correlation dimension and phase space reconstruction are used. Surrogate data is generated from simpler linear models to have similar statistical characteristics with the original data. Nonlinear discriminants are applied to both data, and compared for any significant results. It was concluded that PR interval variability shows non-linear characteristics.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.6
• /
• pp.785-790
• /
• 2006

In RFID-based SCM, The traceability and product information is the important target data. In this paper, efficient items traceability model and the integrated model of the product between RFID network and GDS(Global Data Synchronization) network are studied. Information consists of the dynamic data generated from RFID network and static data generated from GDS Network. The integrated model will provide the interoperability between 2 kinds of networks.

• Journal of information and communication convergence engineering
• /
• v.19 no.3
• /
• pp.166-174
• /
• 2021

A large volume of patient data is generated from various devices used in healthcare applications. With increase in the volume of data generated in the healthcare industry, more wellness monitoring is required. A cloud-enabled analysis of healthcare data that predicts patient risk factors is required. Machine learning techniques have been developed to address these medical care problems. A novel technique called the radix-trie-based Tanimoto kernel regressive infomax boost classification (RT-TKRIBC) technique is introduced to analyze the heterogeneous health data in the cloud to predict the health risks and send alerts. The infomax boost ensemble technique improves the prediction accuracy by finding the maximum mutual information, thereby minimizing the mean square error. The performance evaluation of the proposed RT-TKRIBC technique is realized through extensive simulations in the cloud environment, which provides better prediction accuracy and less prediction time than those provided by the state-of-the-art methods.

• Proceedings of the KSRS Conference
• /
• 2002.10a
• /
• pp.553-558
• /
• 2002

Physical sensor model in pushbroom satellite images can be made from sensor modeling by rotation parameters and attitude parameters on the satellite track. These parameters are determined by the information obtained from GPS, INS, or star tracker. Provided from satellite image, an auxiliary data error is connected directly with an error of rotation parameters and attitude parameters. This paper analyzed how obtaining satellite images influenced errors of rotation parameters and attitude parameters. furthermore, for detailed analysis, this paper generated simulated satellite image, which was changed variously by rotation parameters and attitude parameters of satellite sensor model. Simulated satellite image is generated by using high-resolution digital aerial image and DEM (Digital Elevation Model) data. Moreover, this paper determined correlation of rotation parameter and attitude parameters through error analysis of simulated satellite image that was generated by various rotation parameters and attitude parameters.

• Journal of Environmental Health Sciences
• /
• v.28 no.2
• /
• pp.1-9
• /
• 2002

This study was performed to investigate vibration generated from machines that were used at factories and construction works. Vibrations were measured at three points in a straight line based on distance from the vibration sources, and analyzed to assess the vibration bevels. The average vibration level of factory machines was 65.4dBV at 2m, and that of construction machines was 74.0dBV at 5m. Vibration attenuations was 4.0~8.2dBV by double distance. All such data were applied to gain coefficients of attenuation equations for predicting vibration level by distance from the vibration sources. Data recorded on tapes were analyzed to understand the characteristics of frequency because these characteristics are important factors to design a Plan for installing the vibration-Proof devices. Finally, considering results from these analysis, assessment, and prediction, the methods for reducing vibration generated from machines were discussed.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.539-539
• /
• 2000

This paper shows how to implement force reflection for a needle insertion problem. The target is a needle spine biopsy simulator for tumor inspection by needle insertion. Simulated force is calculated from the relationship of volume graphic data and the orientation and Position of the needle, and it is generated using PHANTOM$^{TM}$. To generate realistic force reflection, the directional force of the needle has been generated by tissue model. The other rotational force is generated using a pivot to keep the needle in the initial inserted direction after puncturing the skin. Since the used haptic device has limitation for generating high stiffness and large damping, scale downed model and digital filter are used to stabilize the system.m.

• Journal of the Korea Society of Computer and Information
• /
• v.23 no.11
• /
• pp.43-49
• /
• 2018

In this paper, we propose a displacement measurement method based on deep learning using image data obtained from tensile tests of a material specimen. We focus on the fact that the sequential images during the tension are generated and the displacement of the specimen is represented in the image data. So, we designed sample generation model which makes sequential images of specimen. The behavior of generated images are similar to the real specimen images under tensile force. Using generated images, we trained and validated our model. In the deep neural network, sequential images are assigned to a multi-channel input to train the network. The multi-channel images are composed of sequential images obtained along the time domain. As a result, the neural network learns the temporal information as the images express the correlation with each other along the time domain. In order to verify the proposed method, we conducted experiments by comparing the deformation measuring performance of the neural network changing the displacement range of images.