Pub Date : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067022
Sudhanshu Arya, Yeon-ho Chung
Orbital angular momentum (OAM) of light has drawn increasing attention due to its intriguingly rich potential for a variety of communication applications. In this paper, we propose a state-of-the-art OAM classification technique using a convolution neural network (CNN) approach for decoding OAM carrying Laguerre-Gaussian beams. We evaluate how well the transmitted alphabet encoded on LG beams is decoded on a noisy channel. From the simulation results, we demonstrate that the OAM beams with different values of OAM mode indexes can readily be classified (or decoded) using the proposed CNN-based approach with average classification accuracy greater than 95%.
{"title":"An OAM Classification Technique using CNN Approach","authors":"Sudhanshu Arya, Yeon-ho Chung","doi":"10.1109/ICAIIC57133.2023.10067022","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067022","url":null,"abstract":"Orbital angular momentum (OAM) of light has drawn increasing attention due to its intriguingly rich potential for a variety of communication applications. In this paper, we propose a state-of-the-art OAM classification technique using a convolution neural network (CNN) approach for decoding OAM carrying Laguerre-Gaussian beams. We evaluate how well the transmitted alphabet encoded on LG beams is decoded on a noisy channel. From the simulation results, we demonstrate that the OAM beams with different values of OAM mode indexes can readily be classified (or decoded) using the proposed CNN-based approach with average classification accuracy greater than 95%.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122101531","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067024
P. Anggraeni, Hilda Khoirunnisa, M. Rizal, Muhammad Fauzian Alfadhila
Modern manufacturing facilities are increasingly leading to highly decentralized systems with self-organized modules that provide flexibility and increase adaptability achieving better performance and efficiency. Therefore, it will be a development of the quadcopter to increase effectiveness in the range of maneuvers but there are still many problems in the communication system, therefore the ROS system is easy to develop in multi-unmanned vehicles and can be implemented in various types of unmanned vehicles so that a multi-unmanned communication system is developed. Communication with a quadcopter with a different firmware will be developed by applying the multi-master ROS. The communication process used in this study uses a wireless LAN with TCP/ IP for connections between multi-masters on ROS embedded on Raspberry and then forwarded using the MavLink serial for each FCU of each ROS multi-master. In this system, a decentralized distribution of data is implemented, where the control center is on a workstation with the ROS system that will control the quadcopter leader. The two quadcopter followers will be controlled or follow the coordinates that have been determined by leader and will form a formation. In testing the whole system, testing data transmission from the workstation to the quadcopter leader is carried out by sending a mode change command to the FCU and producing an average time delay of 0.2s but from the quadcopter leader to the quadcopter follower there are various time delays from each movement of various axes with average delay time for 0.42s. Furthermore, the integration test of this quadcopter has succeeded in forming a formation with the application of this multi-agent communication system with parameters that can follow the existing trajectory.
{"title":"Implementation of WiFi Communication on Multi UAV for Leader-Follower Trajectory based on ROS","authors":"P. Anggraeni, Hilda Khoirunnisa, M. Rizal, Muhammad Fauzian Alfadhila","doi":"10.1109/ICAIIC57133.2023.10067024","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067024","url":null,"abstract":"Modern manufacturing facilities are increasingly leading to highly decentralized systems with self-organized modules that provide flexibility and increase adaptability achieving better performance and efficiency. Therefore, it will be a development of the quadcopter to increase effectiveness in the range of maneuvers but there are still many problems in the communication system, therefore the ROS system is easy to develop in multi-unmanned vehicles and can be implemented in various types of unmanned vehicles so that a multi-unmanned communication system is developed. Communication with a quadcopter with a different firmware will be developed by applying the multi-master ROS. The communication process used in this study uses a wireless LAN with TCP/ IP for connections between multi-masters on ROS embedded on Raspberry and then forwarded using the MavLink serial for each FCU of each ROS multi-master. In this system, a decentralized distribution of data is implemented, where the control center is on a workstation with the ROS system that will control the quadcopter leader. The two quadcopter followers will be controlled or follow the coordinates that have been determined by leader and will form a formation. In testing the whole system, testing data transmission from the workstation to the quadcopter leader is carried out by sending a mode change command to the FCU and producing an average time delay of 0.2s but from the quadcopter leader to the quadcopter follower there are various time delays from each movement of various axes with average delay time for 0.42s. Furthermore, the integration test of this quadcopter has succeeded in forming a formation with the application of this multi-agent communication system with parameters that can follow the existing trajectory.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128407887","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067057
Ida Bagus Krishna Yoga Utama, Duc Hoang Tran, Radityo Fajar Pamungkas, ByungDeok Chung, Y. Jang
Many researchers already perform PM2.5 forecasting. However, the majority of research focuses on predicting PM2.5 concentrations in outdoor environments. In contrast, PM2.5 indoor prediction is rarely conducted, despite being more difficult. This study proposes an LSTM-BNN indoor PM2.5 concentration prediction model. The LSTM in the LSTM-BNN model extracts nonlinear correlations from multivariate time series input while the BNN predicts the PM2.5 concentration. Using multivariable input data, the proposed model estimates PM2.5 values 1 hour, 2 hours, and 3 hours in advance. In addition, the proposed model is compared to RNN, LSTM, Single Dense, Multi Dense, and ConvLSTM. MSE, RMSE, MAE, MAPE, and R2 are employed to evaluate the LSTM-BNN model objectively. The LSTM-BNN model beats other models with 1-hour, 2-hour, and 3-hour prediction MAPE and R2 values of 0.001 and 0.999, 0.004 and 0.996, and 0.004 and 0.999, respectively.
{"title":"Predicting Indoor PM2.5 Concentration using LSTM-BNN in Edge Device","authors":"Ida Bagus Krishna Yoga Utama, Duc Hoang Tran, Radityo Fajar Pamungkas, ByungDeok Chung, Y. Jang","doi":"10.1109/ICAIIC57133.2023.10067057","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067057","url":null,"abstract":"Many researchers already perform PM2.5 forecasting. However, the majority of research focuses on predicting PM2.5 concentrations in outdoor environments. In contrast, PM2.5 indoor prediction is rarely conducted, despite being more difficult. This study proposes an LSTM-BNN indoor PM2.5 concentration prediction model. The LSTM in the LSTM-BNN model extracts nonlinear correlations from multivariate time series input while the BNN predicts the PM2.5 concentration. Using multivariable input data, the proposed model estimates PM2.5 values 1 hour, 2 hours, and 3 hours in advance. In addition, the proposed model is compared to RNN, LSTM, Single Dense, Multi Dense, and ConvLSTM. MSE, RMSE, MAE, MAPE, and R2 are employed to evaluate the LSTM-BNN model objectively. The LSTM-BNN model beats other models with 1-hour, 2-hour, and 3-hour prediction MAPE and R2 values of 0.001 and 0.999, 0.004 and 0.996, and 0.004 and 0.999, respectively.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128567477","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067002
Jaeseob Han, G. Lee, Hyunseo Park, Jun Kyun Choi
As various Internet of Things technologies emerges, IoT monitoring services are rapidly developed. Most IoT sensors deployed in an IoT monitoring environment should reduce the energy consumption of unnecessary data transmission. In this paper, we propose a data accuracy pattern-based transmission period control algorithm. Restoration accuracy patterns of time series data that are missing due to transmission period control are broadly extracted. These restoration accuracy vectors showing similar patterns are clustered into the same cluster. The clustered patterns are modeled based on a logistic function to form a linear weighted sum-based optimization problem that considers the trade-off relationship between the mathematically modeled energy consumption function and the restoration accuracy function. In order to solve the formulated optimization problem, the particle swarm optimization technique is leveraged. The performance evaluations verify that the proposed model simultaneously achieves the best RMSE performance and the second-best energy consumption performance compared to other transmission period control algorithms.
{"title":"Data Accuracy Pattern-based Transmission Period Control Algorithm for IoT networks","authors":"Jaeseob Han, G. Lee, Hyunseo Park, Jun Kyun Choi","doi":"10.1109/ICAIIC57133.2023.10067002","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067002","url":null,"abstract":"As various Internet of Things technologies emerges, IoT monitoring services are rapidly developed. Most IoT sensors deployed in an IoT monitoring environment should reduce the energy consumption of unnecessary data transmission. In this paper, we propose a data accuracy pattern-based transmission period control algorithm. Restoration accuracy patterns of time series data that are missing due to transmission period control are broadly extracted. These restoration accuracy vectors showing similar patterns are clustered into the same cluster. The clustered patterns are modeled based on a logistic function to form a linear weighted sum-based optimization problem that considers the trade-off relationship between the mathematically modeled energy consumption function and the restoration accuracy function. In order to solve the formulated optimization problem, the particle swarm optimization technique is leveraged. The performance evaluations verify that the proposed model simultaneously achieves the best RMSE performance and the second-best energy consumption performance compared to other transmission period control algorithms.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129322600","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10066974
P. Shah, Deepali Sultane, Pratiman Singh
An interesting modern technology called cognitive radio creates new opportunities for the effective utilization of the spectrum. Deep Learning (DL) techniques rely on experimentally recorded data and, when trained properly with a wide range of data, may effectively recognize the radio settings, adapt to different environments, and constantly provide a great performance. Using a variety of signal processing (SP) features, we compare the performance of various deep neural network (DNN) models for spectrum sensing (SS) in this paper. The features that are taken into consideration are differential entropy, energy, Lp-norm and geometric power. Conventional DNN are trained to perform spectrum sensing (SS) in congnitive radio (CR) with two different models of noise. In one noise model we take experimentally recorded data from an unoccupied frequency modulation broadcast channel and in another noise model we consider generalized Gaussian noise (GGN). Through thorough tests based on real-world collected datasets, we find that ResNet and Multilayer perceptron (MLP) architectures provide the most effective result in perspective of likelihood of detection of primary user, for a specific preset value of false-alarm probability.
{"title":"Spectrum Sensing Mechanism For Congnitive Radio using Deep Learning","authors":"P. Shah, Deepali Sultane, Pratiman Singh","doi":"10.1109/ICAIIC57133.2023.10066974","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10066974","url":null,"abstract":"An interesting modern technology called cognitive radio creates new opportunities for the effective utilization of the spectrum. Deep Learning (DL) techniques rely on experimentally recorded data and, when trained properly with a wide range of data, may effectively recognize the radio settings, adapt to different environments, and constantly provide a great performance. Using a variety of signal processing (SP) features, we compare the performance of various deep neural network (DNN) models for spectrum sensing (SS) in this paper. The features that are taken into consideration are differential entropy, energy, Lp-norm and geometric power. Conventional DNN are trained to perform spectrum sensing (SS) in congnitive radio (CR) with two different models of noise. In one noise model we take experimentally recorded data from an unoccupied frequency modulation broadcast channel and in another noise model we consider generalized Gaussian noise (GGN). Through thorough tests based on real-world collected datasets, we find that ResNet and Multilayer perceptron (MLP) architectures provide the most effective result in perspective of likelihood of detection of primary user, for a specific preset value of false-alarm probability.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129911566","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10066978
I. M. Hakim, H. Zakaria, K. Muslim, S. I. Ihsani
Human pose estimation is a field of computer vision science that studies the determination of joint points in the human body based on images or videos. One of the applications of human pose estimation is to evaluate human movement and performance. In this study, 3D markerless human pose estimation was carried out using the direct linear transform and deep learning Blazepose methods. System testing was carried out on the push-up movement by comparing the data results of the markerless system with the marker-based motion capture system. Push-ups are excellent exercises for developing upper body strength or endurance, such as the arms and shoulders. Push-ups are widely used in rehabilitation or recovery after surgical procedures. The American College of Sports Medicine (ACSM) has established a standard for assessing a person's physical endurance based on the number of successful push-ups. Quantitatively, of all the mean absolute errors calculated, 70.9% were below 30 mm, and for measuring joint angles (elbows, hips, and knees) 43% were below 5 degrees. An error value below 30 mm indicates that the system can be used for human movement analysis.
{"title":"3D Human Pose Estimation Using Blazepose and Direct Linear Transform (DLT) for Joint Angle Measurement","authors":"I. M. Hakim, H. Zakaria, K. Muslim, S. I. Ihsani","doi":"10.1109/ICAIIC57133.2023.10066978","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10066978","url":null,"abstract":"Human pose estimation is a field of computer vision science that studies the determination of joint points in the human body based on images or videos. One of the applications of human pose estimation is to evaluate human movement and performance. In this study, 3D markerless human pose estimation was carried out using the direct linear transform and deep learning Blazepose methods. System testing was carried out on the push-up movement by comparing the data results of the markerless system with the marker-based motion capture system. Push-ups are excellent exercises for developing upper body strength or endurance, such as the arms and shoulders. Push-ups are widely used in rehabilitation or recovery after surgical procedures. The American College of Sports Medicine (ACSM) has established a standard for assessing a person's physical endurance based on the number of successful push-ups. Quantitatively, of all the mean absolute errors calculated, 70.9% were below 30 mm, and for measuring joint angles (elbows, hips, and knees) 43% were below 5 degrees. An error value below 30 mm indicates that the system can be used for human movement analysis.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123535087","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067061
Wei Guo, Bailu Jin, S. Sun, Yue Wu, Weijie Qi, J. Zhang
In wireless networks, consumer experience is important for both short monitoring of the Quality of Experience (QoE) as well as long term customer retainment. Current 4G and 5G networks are not equipped to measure QoE in an automated way, and experience is still reported through traditional customer care and drive-testing. In recent years, large-scale social media analytics has enabled researchers to gather statistically significant data on consumer experience and correlate them to major events such as social celebrations or significant network outages. However, the translational pathway from languages to topic-specific emotions (e.g., sentiment) to detecting anomalies in QoE is challenging. This challenge lies in two issues: (1) the social experience data remains sparsely distributed across space, and (2) anomalies in experience jump across sub-topic spaces (e.g., from data rate to signal strength). Here, we solved these two challenges by examining the spectral space of experience across topics using federated learning (FL) to identify anomalies. This can inform telecom operators to pay attention to potential network demand or supply issues in real time using relatively sparse and distributed data. We use real social media data curated for our telecommunication projects across London and the United Kingdom to demonstrate our results. FL was able to achieve 74–92% QoE anomaly detection accuracy, with the benefit of 30–45% reduce data transfer and preserving privacy better than raw data transfer.
{"title":"Federated Learning of Wireless Network Experience Anomalies Using Consumer Sentiment","authors":"Wei Guo, Bailu Jin, S. Sun, Yue Wu, Weijie Qi, J. Zhang","doi":"10.1109/ICAIIC57133.2023.10067061","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067061","url":null,"abstract":"In wireless networks, consumer experience is important for both short monitoring of the Quality of Experience (QoE) as well as long term customer retainment. Current 4G and 5G networks are not equipped to measure QoE in an automated way, and experience is still reported through traditional customer care and drive-testing. In recent years, large-scale social media analytics has enabled researchers to gather statistically significant data on consumer experience and correlate them to major events such as social celebrations or significant network outages. However, the translational pathway from languages to topic-specific emotions (e.g., sentiment) to detecting anomalies in QoE is challenging. This challenge lies in two issues: (1) the social experience data remains sparsely distributed across space, and (2) anomalies in experience jump across sub-topic spaces (e.g., from data rate to signal strength). Here, we solved these two challenges by examining the spectral space of experience across topics using federated learning (FL) to identify anomalies. This can inform telecom operators to pay attention to potential network demand or supply issues in real time using relatively sparse and distributed data. We use real social media data curated for our telecommunication projects across London and the United Kingdom to demonstrate our results. FL was able to achieve 74–92% QoE anomaly detection accuracy, with the benefit of 30–45% reduce data transfer and preserving privacy better than raw data transfer.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120948945","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067079
H. Yamamoto, Masato Iwashita, N. Kondo, Leon Wong, Yusaku Kaneta, Kenta Urano, Takuro Yonezawa, Nobuo Kawaguchi
This article describes the design and prelude implementation of coordinated autonomous networks for remote synchronized video services with the autonomous mobility robots. The domain of this article is the world that human uses autonomous mobility robots (AMR) as a kind of multimedia terminals with 360 degree display and other devices. The video service with AMR and other terminals (e.g. VR goggles) are managed by Bi-directional CDN. And the high quality network infrastructure for the above services are provided by autonomous networks.
{"title":"Coordinated autonomous networks for remote synchronized video services with the autonomous mobility robots - prelude implementation","authors":"H. Yamamoto, Masato Iwashita, N. Kondo, Leon Wong, Yusaku Kaneta, Kenta Urano, Takuro Yonezawa, Nobuo Kawaguchi","doi":"10.1109/ICAIIC57133.2023.10067079","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067079","url":null,"abstract":"This article describes the design and prelude implementation of coordinated autonomous networks for remote synchronized video services with the autonomous mobility robots. The domain of this article is the world that human uses autonomous mobility robots (AMR) as a kind of multimedia terminals with 360 degree display and other devices. The video service with AMR and other terminals (e.g. VR goggles) are managed by Bi-directional CDN. And the high quality network infrastructure for the above services are provided by autonomous networks.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114710405","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067118
Nhu Hai Phung, Chi-Thanh Nguyen, T. Tran, Thi Thu Hang Truong, D. Tran, Thi Trang Nguyen, Duc H. Do
One of the most common congenital anomalies in fetuses is known to be Down syndrome (DS). DS causes various adverse effects on the quality and length of life of children having DS and their families. Therefore, prenatal screening and diagnosis for DS are essential and valuable in antenatal care. Recently, machine learning methods for DS detection have become widespread. However, the existing methods, which use the traditional machine learning models, usually have several limitations while facing imbalanced data and missing data. This paper proposes a multi-branch CNN model combined with a feature rearrangement approach to improve the quality of DS prediction from prenatal screening data. The proposed feature rearrangement approach utilizes Pearson correlation testing and feature grouping to create a proper arrangement for the CNN model. Despite the imbalanced and highly missing data, the experiments show promising results with a Recall of 0.9023, F1-score of 0.8969, and balanced accuracy of 0.9314. These achievements outperform several traditional machine learning and attention-based deep learning models.
{"title":"A Combination of Multi-Branch CNN and Feature Rearrangement for Down Syndrome Prediction","authors":"Nhu Hai Phung, Chi-Thanh Nguyen, T. Tran, Thi Thu Hang Truong, D. Tran, Thi Trang Nguyen, Duc H. Do","doi":"10.1109/ICAIIC57133.2023.10067118","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067118","url":null,"abstract":"One of the most common congenital anomalies in fetuses is known to be Down syndrome (DS). DS causes various adverse effects on the quality and length of life of children having DS and their families. Therefore, prenatal screening and diagnosis for DS are essential and valuable in antenatal care. Recently, machine learning methods for DS detection have become widespread. However, the existing methods, which use the traditional machine learning models, usually have several limitations while facing imbalanced data and missing data. This paper proposes a multi-branch CNN model combined with a feature rearrangement approach to improve the quality of DS prediction from prenatal screening data. The proposed feature rearrangement approach utilizes Pearson correlation testing and feature grouping to create a proper arrangement for the CNN model. Despite the imbalanced and highly missing data, the experiments show promising results with a Recall of 0.9023, F1-score of 0.8969, and balanced accuracy of 0.9314. These achievements outperform several traditional machine learning and attention-based deep learning models.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126229556","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 : 2023-02-20DOI: 10.1109/ICAIIC57133.2023.10067113
Jong-jin Jung, Kyung-Won Kim
This paper introduces a study to verify whether the expanded data through various data augmentation methods are valid in terms of accuracy and bias. Data augmentation is a method of processing and generating other types of data with similar characteristics based on the characteristics of the obtained data, rather than directly collecting data when there is not enough data to increase analysis accuracy. However, unverified and augmented data may actually degrade the results of the analysis. Before using the amplified data for analysis, it is a very important verification factor whether it is accurately propagated in terms of similarity to the source data, and whether bias occurs because only a specific part is concentrated and propagated as a result of the propagation. Therefore, in this paper, a verification method is presented from these two perspectives.
{"title":"Study of validation methods for augmented data","authors":"Jong-jin Jung, Kyung-Won Kim","doi":"10.1109/ICAIIC57133.2023.10067113","DOIUrl":"https://doi.org/10.1109/ICAIIC57133.2023.10067113","url":null,"abstract":"This paper introduces a study to verify whether the expanded data through various data augmentation methods are valid in terms of accuracy and bias. Data augmentation is a method of processing and generating other types of data with similar characteristics based on the characteristics of the obtained data, rather than directly collecting data when there is not enough data to increase analysis accuracy. However, unverified and augmented data may actually degrade the results of the analysis. Before using the amplified data for analysis, it is a very important verification factor whether it is accurately propagated in terms of similarity to the source data, and whether bias occurs because only a specific part is concentrated and propagated as a result of the propagation. Therefore, in this paper, a verification method is presented from these two perspectives.","PeriodicalId":105769,"journal":{"name":"2023 International Conference on Artificial Intelligence in Information and Communication (ICAIIC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126522165","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}