Pub Date : 2023-07-27DOI: 10.1109/ICSSE58758.2023.10227146
Le Hoang Lam, Nguyen Hoang Truong, N. Huy, T. Q. Thanh, D. Tran
This study investigates a force control method for free abrasive polishing to achieve a shiny metal surface finish. Despite the potential of this technique, this method is not widely used in metal surface polishing due to challenges in controlling the force during movement. To address the force control issue, we proposed an approach that uses the feedback force value indirectly calculated through the current value of the AC driver and applies it in a closed-loop control system. The proposed method is evaluated experimentally by polishing different metal pieces with varying particle sizes and movement speeds. The experimental results demonstrate that the proposed force control method effectively maintains a constant force during the polishing process, which leads to an improved surface finish. They also show that decreasing particle size and movement speed can improve the surface finish, while the contact force has a limited effect. These findings contribute to a better understanding of the relationship between force, movement speed, and surface finish quality in metal polishing.
{"title":"The Metal Polishing System for Finishing Shiny Metal Surfaces by Free Abrasive Polishing","authors":"Le Hoang Lam, Nguyen Hoang Truong, N. Huy, T. Q. Thanh, D. Tran","doi":"10.1109/ICSSE58758.2023.10227146","DOIUrl":"https://doi.org/10.1109/ICSSE58758.2023.10227146","url":null,"abstract":"This study investigates a force control method for free abrasive polishing to achieve a shiny metal surface finish. Despite the potential of this technique, this method is not widely used in metal surface polishing due to challenges in controlling the force during movement. To address the force control issue, we proposed an approach that uses the feedback force value indirectly calculated through the current value of the AC driver and applies it in a closed-loop control system. The proposed method is evaluated experimentally by polishing different metal pieces with varying particle sizes and movement speeds. The experimental results demonstrate that the proposed force control method effectively maintains a constant force during the polishing process, which leads to an improved surface finish. They also show that decreasing particle size and movement speed can improve the surface finish, while the contact force has a limited effect. These findings contribute to a better understanding of the relationship between force, movement speed, and surface finish quality in metal polishing.","PeriodicalId":280745,"journal":{"name":"2023 International Conference on System Science and Engineering (ICSSE)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128708968","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-27DOI: 10.1109/ICSSE58758.2023.10227248
Tran Kim Toai, V. Hanh, Vo Minh Huan
This paper proposes the hybrid random forest and long short-term memory (LSTM) to mitigate overfitting issue in time series data in stock market. There are many techniques that reduce the overfitting such as data augmentation, regularization, feature selection, dimension reduction, and so on. We propose the model based on feature selection to reduce the model complexity. First, the model selects the stock data features by random forest model. As the result, the selected features are inputted to the LSTM to predict the stock price. By doing so, the proposed model can improve model accuracy in both training and test dataset and generalize well unseen data to mitigate overfitting. The hybrid random forest and LSTM is compared with hybrid ridge and LSTM, and single LSTM model in ability to mitigate overfitting. The MAE, RSME and R2 are used as performance evaluation metrics. We also conduct the study on various stock datasets to evaluate the performance of overcoming the overfitting problems.
{"title":"Hybrid Random Forest and Long Short-Term Memory to Mitigate Overfitting Issue in Time Series Stock Data","authors":"Tran Kim Toai, V. Hanh, Vo Minh Huan","doi":"10.1109/ICSSE58758.2023.10227248","DOIUrl":"https://doi.org/10.1109/ICSSE58758.2023.10227248","url":null,"abstract":"This paper proposes the hybrid random forest and long short-term memory (LSTM) to mitigate overfitting issue in time series data in stock market. There are many techniques that reduce the overfitting such as data augmentation, regularization, feature selection, dimension reduction, and so on. We propose the model based on feature selection to reduce the model complexity. First, the model selects the stock data features by random forest model. As the result, the selected features are inputted to the LSTM to predict the stock price. By doing so, the proposed model can improve model accuracy in both training and test dataset and generalize well unseen data to mitigate overfitting. The hybrid random forest and LSTM is compared with hybrid ridge and LSTM, and single LSTM model in ability to mitigate overfitting. The MAE, RSME and R2 are used as performance evaluation metrics. We also conduct the study on various stock datasets to evaluate the performance of overcoming the overfitting problems.","PeriodicalId":280745,"journal":{"name":"2023 International Conference on System Science and Engineering (ICSSE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128957053","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-27DOI: 10.1109/ICSSE58758.2023.10227199
Q. Ho, P. S. Minh, T. Do
This study proposes a method for detecting chatter during machining by classifying the recorded sound produced during machining. The sound is recorded during machining with different cutting speeds, and abnormal vibration is determined based on the quality of the machined surface (stable or vibrating). After machining, the sound is processed and classified using the deep learning model VGG16. The sound data is represented as a spectrogram image, which is used to train the image classification model. The results showed that the model achieved an accuracy of 94%. Furthermore, the results demonstrate that sound data is sufficient to identify chatter during machining, and sound classification can be used to develop remote monitoring tools to improve productivity and quality of mechanical machining products.
{"title":"A Study on Machine Learning Application by Convolutional Neural Network Model Classifying Audio to Identify Vibration Phenomenon in the Turning Process","authors":"Q. Ho, P. S. Minh, T. Do","doi":"10.1109/ICSSE58758.2023.10227199","DOIUrl":"https://doi.org/10.1109/ICSSE58758.2023.10227199","url":null,"abstract":"This study proposes a method for detecting chatter during machining by classifying the recorded sound produced during machining. The sound is recorded during machining with different cutting speeds, and abnormal vibration is determined based on the quality of the machined surface (stable or vibrating). After machining, the sound is processed and classified using the deep learning model VGG16. The sound data is represented as a spectrogram image, which is used to train the image classification model. The results showed that the model achieved an accuracy of 94%. Furthermore, the results demonstrate that sound data is sufficient to identify chatter during machining, and sound classification can be used to develop remote monitoring tools to improve productivity and quality of mechanical machining products.","PeriodicalId":280745,"journal":{"name":"2023 International Conference on System Science and Engineering (ICSSE)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127461345","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-27DOI: 10.1109/ICSSE58758.2023.10227208
Truong-Dong Do, Nguyen Xuan Mung, H. Jeong, Yong-Seok Lee, Chang-Woo Sung, S. Hong
Over the past decades, quadcopters have been investigated, due to their mobility and flexibility to operate in a wide range of environments. They have been used in various areas, including surveillance and monitoring. During a mission, drones do not have to remain active once they have reached a target location. To conserve energy and maintain a static position, it is possible to perch and stop the motors in such situations. The problem of achieving a reliable and highly accurate perching method remains a challenge and promising. In this paper, a vision-based autonomous perching approach for nano quadcopters onto a predefined perching target on horizontal surfaces is proposed. First, a perching target with a small marker inside a larger one is designed to improve detection capability at a variety of ranges. Second, a monocular camera is used to calculate the relative poses of the flying vehicle from the markers detected. Then, a Kalman filter is applied to determine the pose more reliably, especially when measurement data is missing. Next, we introduce an algorithm for merging the pose data from multiple markers. Finally, the poses are sent to the perching planner to conduct the real flight test to align the drone with the target’s center and steer it there. Based on the experimental results, the approach proved to be effective and feasible. The drone can successfully perch on the center of markers within two centimeters of precision.
{"title":"Vision-based Autonomous Perching of Quadrotors on Horizontal Surfaces","authors":"Truong-Dong Do, Nguyen Xuan Mung, H. Jeong, Yong-Seok Lee, Chang-Woo Sung, S. Hong","doi":"10.1109/ICSSE58758.2023.10227208","DOIUrl":"https://doi.org/10.1109/ICSSE58758.2023.10227208","url":null,"abstract":"Over the past decades, quadcopters have been investigated, due to their mobility and flexibility to operate in a wide range of environments. They have been used in various areas, including surveillance and monitoring. During a mission, drones do not have to remain active once they have reached a target location. To conserve energy and maintain a static position, it is possible to perch and stop the motors in such situations. The problem of achieving a reliable and highly accurate perching method remains a challenge and promising. In this paper, a vision-based autonomous perching approach for nano quadcopters onto a predefined perching target on horizontal surfaces is proposed. First, a perching target with a small marker inside a larger one is designed to improve detection capability at a variety of ranges. Second, a monocular camera is used to calculate the relative poses of the flying vehicle from the markers detected. Then, a Kalman filter is applied to determine the pose more reliably, especially when measurement data is missing. Next, we introduce an algorithm for merging the pose data from multiple markers. Finally, the poses are sent to the perching planner to conduct the real flight test to align the drone with the target’s center and steer it there. Based on the experimental results, the approach proved to be effective and feasible. The drone can successfully perch on the center of markers within two centimeters of precision.","PeriodicalId":280745,"journal":{"name":"2023 International Conference on System Science and Engineering (ICSSE)","volume":"176 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125811348","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-27DOI: 10.1109/ICSSE58758.2023.10227227
Van Thinh Pham, Hoang Long Nguyen, Hai-Chau Le, M. Nguyen
The Internet of Things is one of the cutting-edge technologies applied in many fields in real life. However, drawbacks that have to resist exist, one of the most dangerous of them being cyber security issues. Distributed Denial of Service (DDoS) is an extreme cyber attack that has long existed and brought many negative effects to IoT networks. Besides, other kinds of intrusions also cause a lot of damage and losses. Therefore, an ML-based IDS is proposed and validated by the IoT-23 dataset. Two independent scenarios are generated, the first scheme for DDoS detection and the second scheme for other attack identification. Random Forest feature importance, sequential forward procedure, and 5-fold cross-validation are carried out in both schemes to find two optimal feature sets that can optimize the classification performance and reduce data dimensionality. As a result, the DDoS detection rate of the top five most important features is extremely impressive with 99.89% accuracy and 99.94% Fl-Score. Besides, other intrusions’ classification result is also outstanding, specifically, 98.89% accuracy and 98.83% Fl-Score with only the six highest-ranking features. The results indicate that DDoS attacks and other irregular activities can be identified efficiently with the proposed approach, which will bring more practical value for solving security problems in IoT environments.
{"title":"Machine Learning-based Intrusion Detection System for DDoS Attack in the Internet of Things","authors":"Van Thinh Pham, Hoang Long Nguyen, Hai-Chau Le, M. Nguyen","doi":"10.1109/ICSSE58758.2023.10227227","DOIUrl":"https://doi.org/10.1109/ICSSE58758.2023.10227227","url":null,"abstract":"The Internet of Things is one of the cutting-edge technologies applied in many fields in real life. However, drawbacks that have to resist exist, one of the most dangerous of them being cyber security issues. Distributed Denial of Service (DDoS) is an extreme cyber attack that has long existed and brought many negative effects to IoT networks. Besides, other kinds of intrusions also cause a lot of damage and losses. Therefore, an ML-based IDS is proposed and validated by the IoT-23 dataset. Two independent scenarios are generated, the first scheme for DDoS detection and the second scheme for other attack identification. Random Forest feature importance, sequential forward procedure, and 5-fold cross-validation are carried out in both schemes to find two optimal feature sets that can optimize the classification performance and reduce data dimensionality. As a result, the DDoS detection rate of the top five most important features is extremely impressive with 99.89% accuracy and 99.94% Fl-Score. Besides, other intrusions’ classification result is also outstanding, specifically, 98.89% accuracy and 98.83% Fl-Score with only the six highest-ranking features. The results indicate that DDoS attacks and other irregular activities can be identified efficiently with the proposed approach, which will bring more practical value for solving security problems in IoT environments.","PeriodicalId":280745,"journal":{"name":"2023 International Conference on System Science and Engineering (ICSSE)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125910872","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-27DOI: 10.1109/ICSSE58758.2023.10227203
Truong Hoang Bao Huy, D. Vo, H. Nguyen, Hoa Phuoc Truong, K. Dang, K. H. Truong
As energy demand increases rapidly, short-term load forecasting is becoming progressively vital in power system dispatch and demand response. This study proposes a short-term load forecasting approach for the power system in Vietnam. In this regard, a gated recurrent unit-based deep learning model is applied to use the historical load sequences to forecast the single-step and multi-step ahead values of the load consumption. The hourly load consumption dataset is provided by Ho Chi Minh City Power Corporation (EVNHCMC). Simulation results prove the effectiveness of the developed prediction algorithm for short-term load forecasting, especially for multi-step forecasting.
{"title":"Short-Term Load Forecasting in Power System Using Recurrent Neural Network","authors":"Truong Hoang Bao Huy, D. Vo, H. Nguyen, Hoa Phuoc Truong, K. Dang, K. H. Truong","doi":"10.1109/ICSSE58758.2023.10227203","DOIUrl":"https://doi.org/10.1109/ICSSE58758.2023.10227203","url":null,"abstract":"As energy demand increases rapidly, short-term load forecasting is becoming progressively vital in power system dispatch and demand response. This study proposes a short-term load forecasting approach for the power system in Vietnam. In this regard, a gated recurrent unit-based deep learning model is applied to use the historical load sequences to forecast the single-step and multi-step ahead values of the load consumption. The hourly load consumption dataset is provided by Ho Chi Minh City Power Corporation (EVNHCMC). Simulation results prove the effectiveness of the developed prediction algorithm for short-term load forecasting, especially for multi-step forecasting.","PeriodicalId":280745,"journal":{"name":"2023 International Conference on System Science and Engineering (ICSSE)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126052941","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-27DOI: 10.1109/ICSSE58758.2023.10227142
Phuc Hong Lam, Khoa Nguyen Dang Tran, T. Nguyen, Hoa Truong Phuoc, H. Nguyen, D. M. Pham
DC microgrids offer a potential solution for efficient and reliable energy sharing within local communities. Nonetheless, conventional control techniques employed in these systems are hindered by their inability to ensure optimal energy distribution and load balancing, particularly under conditions of fluctuating loads and renewable power sources. To overcome these problems, this paper presents an adaptive droop controller that adjusts the droop parameters in real time using the primary current sharing loops to reduce the load current deviation and shifts the droop lines through the second loop to eliminate bus voltage deviation in DC microgrids. The proposed system adjusts its droop coefficients dynamically to regulate the load power. The performance of the proposed system is evaluated through PLECS software simulation. Simulation results show that the proposed system achieves better load balancing and stability compared to traditional control methods. Also, this study provides valuable insights into the development and implementation of adaptive droop control systems for automatic-energy sharing in DC microgrids, which can contribute to the development of sustainable and reliable energy systems in local communities.
{"title":"Adaptive Droop Control System for Automatic Voltage Restoration in DC Microgrids","authors":"Phuc Hong Lam, Khoa Nguyen Dang Tran, T. Nguyen, Hoa Truong Phuoc, H. Nguyen, D. M. Pham","doi":"10.1109/ICSSE58758.2023.10227142","DOIUrl":"https://doi.org/10.1109/ICSSE58758.2023.10227142","url":null,"abstract":"DC microgrids offer a potential solution for efficient and reliable energy sharing within local communities. Nonetheless, conventional control techniques employed in these systems are hindered by their inability to ensure optimal energy distribution and load balancing, particularly under conditions of fluctuating loads and renewable power sources. To overcome these problems, this paper presents an adaptive droop controller that adjusts the droop parameters in real time using the primary current sharing loops to reduce the load current deviation and shifts the droop lines through the second loop to eliminate bus voltage deviation in DC microgrids. The proposed system adjusts its droop coefficients dynamically to regulate the load power. The performance of the proposed system is evaluated through PLECS software simulation. Simulation results show that the proposed system achieves better load balancing and stability compared to traditional control methods. Also, this study provides valuable insights into the development and implementation of adaptive droop control systems for automatic-energy sharing in DC microgrids, which can contribute to the development of sustainable and reliable energy systems in local communities.","PeriodicalId":280745,"journal":{"name":"2023 International Conference on System Science and Engineering (ICSSE)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131214448","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-27DOI: 10.1109/ICSSE58758.2023.10227247
Loc Nguyen-Van-Thanh, Tan Do-Duy
In this paper, we propose an improved Low-Density Parity-Check (LDPC) code design scheme based on the existing genetic optimization-based LDPC code design scheme proposed in [1]. In particular, we perform the removal of the girth-4 property of the parity check matrix (H-matrix) and utilize the min-sum decoding algorithm instead of the Belief Propagation (BP) algorithm in order to enhance the performance of the LDPC code. Furthermore, an (32,64) LDPC code is considered in this paper. Finally, we evaluate the block error rate (BLER) of the LDPC code over white Gaussian noise channels. By means of evaluation results using Matlab, we indicate that our proposed approach can achieve a gain of more than 11% in terms of BLER compared to the existing schemes without significantly increasing the complexity of the decoding scheme.
{"title":"Efficient Genetic Algorithm-based LDPC Code Design for IoT Applications","authors":"Loc Nguyen-Van-Thanh, Tan Do-Duy","doi":"10.1109/ICSSE58758.2023.10227247","DOIUrl":"https://doi.org/10.1109/ICSSE58758.2023.10227247","url":null,"abstract":"In this paper, we propose an improved Low-Density Parity-Check (LDPC) code design scheme based on the existing genetic optimization-based LDPC code design scheme proposed in [1]. In particular, we perform the removal of the girth-4 property of the parity check matrix (H-matrix) and utilize the min-sum decoding algorithm instead of the Belief Propagation (BP) algorithm in order to enhance the performance of the LDPC code. Furthermore, an (32,64) LDPC code is considered in this paper. Finally, we evaluate the block error rate (BLER) of the LDPC code over white Gaussian noise channels. By means of evaluation results using Matlab, we indicate that our proposed approach can achieve a gain of more than 11% in terms of BLER compared to the existing schemes without significantly increasing the complexity of the decoding scheme.","PeriodicalId":280745,"journal":{"name":"2023 International Conference on System Science and Engineering (ICSSE)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133923846","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-27DOI: 10.1109/ICSSE58758.2023.10227245
T. Le, M. Hsieh, Phan-Thanh Nguyen, M. Nguyen
This paper proposes a novel approach to improve the traditional speed controller of the field-oriented control (FOC) strategy for permanent magnet synchronous motor (PMSM) drives. The performance and robustness of the speed controller for PMSM drives are limited when using the traditional proportional-integral (PI) method. The proposed approach is the terminal sliding mode high-order control (TSMHC), designed to ensure fast and accurate tracking for PMSM drives. The TSMHC approach integrates the advantages of terminal sliding mode (TSM) and high-order control law. TSM brings faster tracking with smaller steady-state errors, while high-order control law can reduce the reaching time between the initial system state and the sliding-mode surface with slight chattering. The stability of the TSMHC is evaluated using the Lyapunov stability theory. The simulation results validate the efficiency and superiority of the proposed TSMHC approach.
{"title":"Speed Control for Permanent Magnet Synchronous Motor Based on Terminal Sliding Mode High-order Control","authors":"T. Le, M. Hsieh, Phan-Thanh Nguyen, M. Nguyen","doi":"10.1109/ICSSE58758.2023.10227245","DOIUrl":"https://doi.org/10.1109/ICSSE58758.2023.10227245","url":null,"abstract":"This paper proposes a novel approach to improve the traditional speed controller of the field-oriented control (FOC) strategy for permanent magnet synchronous motor (PMSM) drives. The performance and robustness of the speed controller for PMSM drives are limited when using the traditional proportional-integral (PI) method. The proposed approach is the terminal sliding mode high-order control (TSMHC), designed to ensure fast and accurate tracking for PMSM drives. The TSMHC approach integrates the advantages of terminal sliding mode (TSM) and high-order control law. TSM brings faster tracking with smaller steady-state errors, while high-order control law can reduce the reaching time between the initial system state and the sliding-mode surface with slight chattering. The stability of the TSMHC is evaluated using the Lyapunov stability theory. The simulation results validate the efficiency and superiority of the proposed TSMHC approach.","PeriodicalId":280745,"journal":{"name":"2023 International Conference on System Science and Engineering (ICSSE)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131236435","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-07-27DOI: 10.1109/ICSSE58758.2023.10227159
Hoa Le Viet, Toai Tran Hoang Cong, Tuan Trinh Nguyen Bao, Duy Tran Ngoc Bao, N. H. Tuong
Despite the growing interest in learning Vietnamese, pronunciation remains a significant challenge for many language learners. This study explores the use of deep learning techniques to automatically detect incorrect pronunciation in Vietnamese. Our approach utilizes a multi-task setup that incorporates an Audio Encoder and a Phoneme Recognizer, enabling the model to learn the alignment between phonemes and acoustic features. This alignment information is then employed by the Incorrect Pronunciation Detector to identify words with incorrect pronunciation. Notably, we propose a novel strategy for generating pronunciation features, which involves “manually” grouping phonemes of the same word, thereby facilitating the model’s learning process. To evaluate the effectiveness of the proposed method, we build a small non-native (L2) Vietnamese speech dataset for training and testing. Compared to the baseline model, our final result improves the accuracy by 5.2% and $F_{1}$ score by 21.14%.
{"title":"A Deep Learning-Based Strategy for Vietnamese Incorrect Pronunciation Detection","authors":"Hoa Le Viet, Toai Tran Hoang Cong, Tuan Trinh Nguyen Bao, Duy Tran Ngoc Bao, N. H. Tuong","doi":"10.1109/ICSSE58758.2023.10227159","DOIUrl":"https://doi.org/10.1109/ICSSE58758.2023.10227159","url":null,"abstract":"Despite the growing interest in learning Vietnamese, pronunciation remains a significant challenge for many language learners. This study explores the use of deep learning techniques to automatically detect incorrect pronunciation in Vietnamese. Our approach utilizes a multi-task setup that incorporates an Audio Encoder and a Phoneme Recognizer, enabling the model to learn the alignment between phonemes and acoustic features. This alignment information is then employed by the Incorrect Pronunciation Detector to identify words with incorrect pronunciation. Notably, we propose a novel strategy for generating pronunciation features, which involves “manually” grouping phonemes of the same word, thereby facilitating the model’s learning process. To evaluate the effectiveness of the proposed method, we build a small non-native (L2) Vietnamese speech dataset for training and testing. Compared to the baseline model, our final result improves the accuracy by 5.2% and $F_{1}$ score by 21.14%.","PeriodicalId":280745,"journal":{"name":"2023 International Conference on System Science and Engineering (ICSSE)","volume":"66 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134565535","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：