• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.5
• /
• pp.1590-1609
• /
• 2021

Laughter is one of the most important nonverbal sound that human generates. It is a means for expressing his emotions. The acoustic and contextual features of this specific sound are different from those of speech and many difficulties arise during their modeling process. During this work, we propose an audio laughter generation system based on unsupervised generative models: the autoencoder (AE) and its variants. This procedure is the association of three main sub-process, (1) the analysis which consist of extracting the log magnitude spectrogram from the laughter database, (2) the generative models training, (3) the synthesis stage which incorporate the involvement of an intermediate mechanism: the vocoder. To improve the synthesis quality, we suggest two hybrid models (LSTM-VAE, GRU-VAE and CNN-VAE) that combine the representation learning capacity of variational autoencoder (VAE) with the temporal modelling ability of a long short-term memory RNN (LSTM) and the CNN ability to learn invariant features. To figure out the performance of our proposed audio laughter generation process, objective evaluation (RMSE) and a perceptual audio quality test (listening test) were conducted. According to these evaluation metrics, we can show that the GRU-VAE outperforms the other VAE models.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.10a
• /
• pp.78-81
• /
• 2021

Through the smart factory construction project, sensors can be installed in manufacturing production facilities and various process data can be collected in real time. Through this, research on real-time facility anomaly detection is being actively conducted to reduce production interruption due to facility abnormality in the manufacturing process. In this paper, to detect abnormalities in production facilities, the manufacturing data was applied to deep learning models Autoencoder(AE), VAE(Variational Autoencoder), and AAE(Adversarial Autoencoder) to derive the results. Manufacturing data was used as input data through a simple moving average technique and preprocessing process, and performance analysis was conducted according to the window size of the simple movement average technique and the feature vector size of the AE model.

• International Journal of Internet, Broadcasting and Communication
• /
• v.15 no.3
• /
• pp.94-102
• /
• 2023

Anomaly detection in human movements can improve safety in indoor workplaces. In this paper, we design a framework for detecting anomalous trajectories of humans in indoor spaces based on a variational autoencoder (VAE) with Bi-LSTM layers. First, the VAE is trained to capture the latent representation of normal trajectories. Then the abnormality of a new trajectory is checked using the trained VAE. In this step, the anomaly score of the trajectory is determined using the trajectory reconstruction error through the VAE. If the anomaly score exceeds a threshold, the trajectory is detected as an anomaly. To select the anomaly threshold, a new metric called D-score is proposed, which measures the difference between recall and precision. The anomaly threshold is selected according to the minimum value of the D-score on the validation set. The MIT Badge dataset, which is a real trajectory dataset of workers in indoor space, is used to evaluate the proposed framework. The experiment results show that our framework effectively identifies abnormal trajectories with 81.22% in terms of the F1-score.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.40 no.2
• /
• pp.233-238
• /
• 2020

Most deep learning (DL) approaches for bridge damage localization based on a structural health monitoring system commonly use supervised learning-based DL models. The supervised learning-based DL model requires the response data obtained from sensors on the bridge and also the label which indicates the damaged state of the bridge. However, it is impractical to accurately obtain the label data in fields, thus, the supervised learning-based DL model has a limitation in that it is not easily applicable in practice. On the other hand, an unsupervised learning-based DL model has the merit of being able to train without label data. Considering this advantage, this study aims to propose and theoretically validate a damage localization approach for bridges using a variational autoencoder, a representative unsupervised learning-based DL network: as a result, this study indicated the feasibility of VAE for damage localization.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.28 no.6
• /
• pp.1463-1474
• /
• 2018

Fuzzing is one of the software testing techniques that find security flaws by inputting invalid values or arbitrary values into the program and various methods have been suggested to increase the efficiency of such fuzzing. In this paper, focusing on the existence of field with high relevance to coverage and software crash, we propose a new method for intensively fuzzing corresponding field part while performing field based fuzzing. In this case, we use a deep learning model called Variational Autoencoder(VAE) to learn the statistical characteristic of input values measured in high coverage and it showed that the coverage of the regenerated files are uniformly higher than that of simple variation. It also showed that new crash could be found by learning the statistical characteristic of the files in which the crash occurred and applying the dropout during the regeneration. Experimental results showed that the coverage is about 10% higher than the files in the queue of the AFL fuzzing tool and in the Hwpviewer binary, we found two new crashes using two crashes that found at the initial fuzzing phase.

• Journal of Broadcast Engineering
• /
• v.28 no.3
• /
• pp.275-284
• /
• 2023

Gene expression data can be utilized in various studies, including the prediction of disease prognosis. However, there are challenges associated with collecting enough data due to cost constraints. In this paper, we propose a gene expression data generation model based on Conditional Variational Autoencoder. Our results demonstrate that the proposed model generates synthetic data with superior quality compared to two other state-of-the-art models for gene expression data generation, namely the Wasserstein Generative Adversarial Network with Gradient Penalty based model and the structured data generation models CTGAN and TVAE.

• Journal of Korean Society for Quality Management
• /
• v.49 no.4
• /
• pp.581-594
• /
• 2021

Purpose: The purpose of this study is to compare machine learning models for anomaly detection of mechanical facility equipment and suggest an anomaly detection system for mechanical facility equipment in subway stations. It helps to predict failures and plan the maintenance of facility. Ultimately it aims to improve the quality of facility equipment. Methods: The data collected from Daejeon Metropolitan Rapid Transit Corporation was used in this experiment. The experiment was performed using Python, Scikit-learn, tensorflow 2.0 for preprocessing and machine learning. Also it was conducted in two failure states of the equipment. We compared and analyzed five unsupervised machine learning models focused on model Long Short-Term Memory Variational Autoencoder(LSTM-VAE). Results: In both experiments, change in vibration and current data was observed when there is a defect. When the rotating body failure was happened, the magnitude of vibration has increased but current has decreased. In situation of axis alignment failure, both of vibration and current have increased. In addition, model LSTM-VAE showed superior accuracy than the other four base-line models. Conclusion: According to the results, model LSTM-VAE showed outstanding performance with more than 97% of accuracy in the experiments. Thus, the quality of mechanical facility equipment will be improved if the proposed anomaly detection system is established with this model used.

• Journal of Korea Society of Industrial Information Systems
• /
• v.25 no.6
• /
• pp.1-13
• /
• 2020

In this paper, the author proposed a nuclear power plant (NPP) instrumentation signal faults identification algorithm. A variational autoencoder (VAE)-based model is trained by using only normal dataset as same as existing anomaly detection method, and trained model predicts which signal within the entire signal set is anomalous. Classification of anomalous signals is performed based on the reconstruction error for each kind of signal and partial derivatives of reconstruction error with respect to the specific part of an input. Simulation was conducted to acquire the data for the experiments. Through the experiments, it was identified that the proposed signal fault identification method can specify the anomalous signals within acceptable range of error.

• ETRI Journal
• /
• v.45 no.4
• /
• pp.678-689
• /
• 2023

This paper describes a novel end-to-end deep generative model-based speaker recognition system using prosodic features. The usefulness of variational autoencoders (VAE) in learning the speaker-specific prosody representations for the speaker recognition task is examined herein for the first time. The speech signal is first automatically segmented into syllable-like units using vowel onset points (VOP) and energy valleys. Prosodic features, such as the dynamics of duration, energy, and fundamental frequency (F₀), are then extracted at the syllable level and used to train/adapt a speaker-dependent VAE from a universal VAE. The initial comparative studies on VAEs and traditional autoencoders (AE) suggest that the former can efficiently learn speaker representations. Investigations on the impact of gender information in speaker recognition also point out that gender-dependent impostor banks lead to higher accuracies. Finally, the evaluation on the NIST SRE 2010 dataset demonstrates the usefulness of the proposed approach for speaker recognition.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.22 no.5
• /
• pp.75-86
• /
• 2022

This study aims to solve the problem of class imbalance in numerical data by using a deep learning-based Variational AutoEncoder and to improve the performance of the learning model by augmenting the learning data. We propose 'D-VAE' to artificially increase the number of records for a given table data. The main features of the proposed technique go through discretization and feature selection in the preprocessing process to optimize the data. In the discretization process, K-means are applied and grouped, and then converted into one-hot vectors by one-hot encoding technique. Subsequently, for memory efficiency, sample data are generated with Variational AutoEncoder using only features that help predict with RFECV among feature selection techniques. To verify the performance of the proposed model, we demonstrate its validity by conducting experiments by data augmentation ratio.