• Annual Conference of KIPS
• /
• 2023.05a
• /
• pp.527-529
• /
• 2023

Generative Adversarial Networks (GANs) models have developed rapidly due to the emergence of various variation models and their wide applications. Despite many recent developments in GANs, mode collapse, and instability are still unresolved issues. To address these problems, we focused on the fact that a single GANs model itself cannot realize local failure during the training phase without external standards. This paper introduces a novel training process involving multiple GANs, inspired by auction mechanisms. During the training, auxiliary performance metrics for each GANs are determined by the others through the process of various auction methods.

• International Journal of Advanced Culture Technology
• /
• v.6 no.4
• /
• pp.275-283
• /
• 2018

Cancer show distinct pattern of gene expression when it is compared to normal. This difference results malignant characteristic of cancer. Many cancer drugs are targeting this difference so that it can selectively kill cancer cells. One of the recent demand for personalized treating cancer is retrieving normal tissue from a patient so that the gene expression difference between cancer and normal be assessed. However, in most clinical situation it is hard to retrieve normal tissue from a patient. This is because biopsy of normal tissues may cause damage to the organ function or a risk of infection or side effect what a patient to take. Thus, there is a challenge to estimate normal cell's gene expression where cancers are originated from without taking additional biopsy. In this paper, we propose in-silico based prediction of normal cell's gene expression from gene expression data of a tumor sample. We call this challenge as reverting the cancer into normal. We divided this challenge into two parts. The first part is making a generator that is able to fool a pretrained discriminator. Pretrained discriminator is from the training of public data (9,601 cancers, 7,240 normals) which shows 0.997 of accuracy to discriminate if a given gene expression pattern is cancer or normal. Deceiving this pretrained discriminator means our method is capable of generating very normal-like gene expression data. The second part of the challenge is to address whether generated normal is similar to true reverse form of the input cancer data. We used, cycle-consistent adversarial networks to approach our challenges, since this network is capable of translating one domain to the other while maintaining original domain's feature and at the same time adding the new domain's feature. We evaluated that, if we put cancer data into a cycle-consistent adversarial network, it could retain most of the information from the input (cancer) and at the same time change the data into normal. We also evaluated if this generated gene expression of normal tissue would be the biological reverse form of the gene expression of cancer used as an input.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.7
• /
• pp.311-316
• /
• 2019

Classification problem is to predict the class to which an input data belongs. One of the most popular methods to do this is training a machine learning algorithm using the given dataset. In this case, the dataset should have a well-balanced class distribution for the best performance. However, when the dataset has an imbalanced class distribution, its classification performance could be very poor. To overcome this problem, we propose an over-sampling scheme that balances the number of data by using Conditional Generative Adversarial Networks (CGAN). CGAN is a generative model developed from Generative Adversarial Networks (GAN), which can learn data characteristics and generate data that is similar to real data. Therefore, CGAN can generate data of a class which has a small number of data so that the problem induced by imbalanced class distribution can be mitigated, and classification performance can be improved. Experiments using actual collected data show that the over-sampling technique using CGAN is effective and that it is superior to existing over-sampling techniques.

• Convergence Security Journal
• /
• v.24 no.2
• /
• pp.201-207
• /
• 2024

Voice is a critical element of human communication, and the development of speech recognition models is one of the significant achievements in artificial intelligence, which has recently been applied in various aspects of human life. The application of speech recognition models in the military field is also inevitable. However, before artificial intelligence models can be applied in the military, it is necessary to research their vulnerabilities. In this study, we evaluates the military applicability of the multilingual speech recognition model "Whisper" by examining its vulnerabilities to battlefield noise, white noise, and adversarial attacks. In experiments involving battlefield noise, Whisper showed significant performance degradation with an average Character Error Rate (CER) of 72.4%, indicating difficulties in military applications. In experiments with white noise, Whisper was robust to low-intensity noise but showed performance degradation under high-intensity noise. Adversarial attack experiments revealed vulnerabilities at specific epsilon values. Therefore, the Whisper model requires improvements through fine-tuning, adversarial training, and other methods.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.22 no.6
• /
• pp.797-805
• /
• 2019

Threats targeting cyberspace are becoming more intelligent and increasing day by day. To cope with such cyber threats, it is essential to improve the coping ability of system security officers. In this paper, we propose a simulated threat generator that automatically generates cyber threats for cyber defense training. The proposed Simulated Threat Generator is designed with MITRE ATT & CK(Adversarial Tactics, Techniques and Common Knowledge) framework to easily add an evolving cyber threat and select the next threat based on the threat execution result.

• International Journal of Computer Science & Network Security
• /
• v.24 no.10
• /
• pp.213-220
• /
• 2024

Retinal vessel analysis plays a vital role in the detection of some diseases. For example, diabetic retinopathy which may lead to blindness is one of the most common diseases that cause retinal blood vessel structure to change. However, doctors usually take a lot of time and money to collect and label training sets. Thus, automated vessel segmentation as the first step toward computer-aided analysis of fundus remains an active research avenue. We propose an automated Retinal vessel segmentation method based on the GAN network. Traditional image segmentation networks are unsupervised, and GAN is a new semi-supervised network due to adding a Discriminator. By training the discriminator network, we can capture the quality of the generator's output and drive it closer to the true image features. In our experiment, we use DRIVE dataset for training and testing. The final segmentation effect is represented by the Dice coefficient. Experimental results show that the GAN network can effectively improve the edge effect of image segmentation. Compared with the traditional U-net network, GAN shows about 1.55% higher segmentation accuracy.

• The Bulletin of The Korean Astronomical Society
• /
• v.45 no.1
• /
• pp.50.1-50.1
• /
• 2020

In this topic, I will introduce a recent work on reconstructing the cosmic density field based on the GAN. I will show the performance of the GAN compared to the traditional Unet architecture. I'd also like to discuss a 3-channels-based 2D datasets for the training to recover the 3D density field. Finally, I will present some performance tests based on the test datasets.

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.5
• /
• pp.447-453
• /
• 2020

In this paper, we propose a system to extract effective speaker representations from a speech signal using a deep learning method. Based on the fact that speech signal contains identity unrelated information such as text content, emotion, background noise, and so on, we perform a training such that the extracted features only represent speaker-related information but do not represent speaker-unrelated information. Specifically, we propose an auto-encoder based disentanglement method that outputs both speaker-related and speaker-unrelated embeddings using effective loss functions. To further improve the reconstruction performance in the decoding process, we also introduce a discriminator popularly used in Generative Adversarial Network (GAN) structure. Since improving the decoding capability is helpful for preserving speaker information and disentanglement, it results in the improvement of speaker verification performance. Experimental results demonstrate the effectiveness of our proposed method by improving Equal Error Rate (EER) on benchmark dataset, Voxceleb1.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.12
• /
• pp.1919-1925
• /
• 2022

In this paper, we propose an image-to-image translation approach based on the conditional generative adversarial network for semantic segmentation. Semantic segmentation is the task of clustering parts of an image together which belong to the same object class. Unlike the traditional pixel-wise classification approach, the proposed method parses an input RGB image to its corresponding semantic segmentation mask using a pixel regression approach. The proposed method is based on the Pix2Pix image synthesis method. We employ residual connections-based convolutional neural network architectures for both the generator and discriminator architectures, as the residual connections speed up the training process and generate more accurate results. The proposed method has been trained and tested on the NYU-depthV2 dataset and could achieve a good mIOU value (49.5%). We also compare the proposed approach to the current methods in semantic segmentation showing that the proposed method outperforms most of those methods.

• Journal of Technologic Dentistry
• /
• v.46 no.2
• /
• pp.36-41
• /
• 2024

Purpose: This study aimed to generate virtual mandibular left first molar teeth using deep convolutional generative adversarial networks (DCGANs) and analyze their matching accuracy with actual tooth morphology to propose a new paradigm for using medical data. Methods: Occlusal surface images of the mandibular left first molar scanned using a dental model scanner were analyzed using DCGANs. Overall, 100 training sets comprising 50 original and 50 background-removed images were created, thus generating 1,000 virtual teeth. These virtual teeth were classified based on the number of cusps and occlusal surface ratio, and subsequently, were analyzed for consistency by expert dental technicians over three rounds of examination. Statistical analysis was conducted using IBM SPSS Statistics ver. 23.0 (IBM), including intraclass correlation coefficient for intrarater reliability, one-way ANOVA, and Tukey's post-hoc analysis. Results: Virtual mandibular left first molars exhibited high consistency in the occlusal surface ratio but varied in other criteria. Moreover, consistency was the highest in the occlusal buccal lingual criteria at 91.9%, whereas discrepancies were observed most in the occusal buccal cusp criteria at 85.5%. Significant differences were observed among all groups (p<0.05). Conclusion: Based on the classification of the virtually generated left mandibular first molar according to several criteria, DCGANs can generate virtual data highly similar to real data. Thus, subsequent research in the dental field, including the development of improved neural network structures, is necessary.