Pub Date : 2021-09-24DOI: 10.1109/ICASI52993.2021.9568450
J. Hung, Jing-Rong Lin, Ling-Yu Zhuang
Neural machine translation (NMT) employs the prevailing deep learning techniques to build a single deep neural network (DNN) that directly maps the input speech utterances of one language to the corresponding texts of the other language. Compared with the conventional statistical machine translation, which separately optimizes each component model (such as acoustic models and language models) in series, NMT can learn the used DNN to directly maximize the overall translation performance. In particular, a novel encoder-decoder DNN structure termed transformer, which Google develops, has been applied in NMT and revealed outstanding translation performance. In this study, we investigate and evaluate the Transformer-based speech translation algorithm by varying the model settings in the training process of the used Transformer. The experiments follow a tutorial script provided in the Tensorflow forum, which are conducted on the TED talk dataset to translate Portuguese to English, which consists of 50,000 utterances for training, 1,100 utterances for validation, and 2,000 utterances for testing. The baseline system, which sets the encoding dimension as 128, the number of encoder/decoder layers as 4, the dropout rate as 0.1 and the negative exponent as −1.5, gives rise to 68.01% in translation accuracy. While the encoding dimension is increased to be 512, the translation accuracy can be promoted to be 76.02%. Also, changing the number of layers to be 2, the dropout rate to be 0.01 and the negative exponent to be 1 can achieve 70.98%, 80.97% and 75.40% in translation accuracy, respectively. The experimental results indicate that we can further improve the translation performance of the transformer by properly arranging the underlying hyper-parameters.
{"title":"The Evaluation Study of the Deep Learning Model Transformer in Speech Translation","authors":"J. Hung, Jing-Rong Lin, Ling-Yu Zhuang","doi":"10.1109/ICASI52993.2021.9568450","DOIUrl":"https://doi.org/10.1109/ICASI52993.2021.9568450","url":null,"abstract":"Neural machine translation (NMT) employs the prevailing deep learning techniques to build a single deep neural network (DNN) that directly maps the input speech utterances of one language to the corresponding texts of the other language. Compared with the conventional statistical machine translation, which separately optimizes each component model (such as acoustic models and language models) in series, NMT can learn the used DNN to directly maximize the overall translation performance. In particular, a novel encoder-decoder DNN structure termed transformer, which Google develops, has been applied in NMT and revealed outstanding translation performance. In this study, we investigate and evaluate the Transformer-based speech translation algorithm by varying the model settings in the training process of the used Transformer. The experiments follow a tutorial script provided in the Tensorflow forum, which are conducted on the TED talk dataset to translate Portuguese to English, which consists of 50,000 utterances for training, 1,100 utterances for validation, and 2,000 utterances for testing. The baseline system, which sets the encoding dimension as 128, the number of encoder/decoder layers as 4, the dropout rate as 0.1 and the negative exponent as −1.5, gives rise to 68.01% in translation accuracy. While the encoding dimension is increased to be 512, the translation accuracy can be promoted to be 76.02%. Also, changing the number of layers to be 2, the dropout rate to be 0.01 and the negative exponent to be 1 can achieve 70.98%, 80.97% and 75.40% in translation accuracy, respectively. The experimental results indicate that we can further improve the translation performance of the transformer by properly arranging the underlying hyper-parameters.","PeriodicalId":103254,"journal":{"name":"2021 7th International Conference on Applied System Innovation (ICASI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130034491","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 : 2021-09-24DOI: 10.1109/ICASI52993.2021.9568451
Hendry, D. Manongga, Yessica Nataliani, Theopilus Hermanus Wellem
In recent years, the growth of machine learning makes the computer can learn many things by using artificial intelligence. One method that is feared nowadays is the computer's capability to imitate something. This capability is called deep-fake. Deep-fake is the capability of the computer to imitate human characteristics such as voice, images, and video through artificial intelligence. Deep-fake is used to combine put the consisted image and video to another source of images and video using machine learning which is known as a generative adversarial network. With these capabilities, deep-fake is already used to make a counterfeit video, signature, voice signature, and much fake news. This paper is about to combine the capabilities of deep learning and the Generative Adversarial Network (GAN) to deal with detecting the fraud in the handwritten signature. We will focus on several types of ways to sign with the characters. The system will recommend if the hand signature of the user is fake or genuine. This is under the capabilities of GAN to synthesize the signature, it can make the computer automatically generate hand signature by using a machine. Many researchers called this capability is deep-fake. This research aims to learn the hand signature to do fraud detection. We propose an architecture to build the anti-counterfeiting hand signature which is utilized deep learning with a self-growing probabilistic method.
{"title":"Anti-Counterfeit Handwritten Signature via DCGAN with SGPD Network","authors":"Hendry, D. Manongga, Yessica Nataliani, Theopilus Hermanus Wellem","doi":"10.1109/ICASI52993.2021.9568451","DOIUrl":"https://doi.org/10.1109/ICASI52993.2021.9568451","url":null,"abstract":"In recent years, the growth of machine learning makes the computer can learn many things by using artificial intelligence. One method that is feared nowadays is the computer's capability to imitate something. This capability is called deep-fake. Deep-fake is the capability of the computer to imitate human characteristics such as voice, images, and video through artificial intelligence. Deep-fake is used to combine put the consisted image and video to another source of images and video using machine learning which is known as a generative adversarial network. With these capabilities, deep-fake is already used to make a counterfeit video, signature, voice signature, and much fake news. This paper is about to combine the capabilities of deep learning and the Generative Adversarial Network (GAN) to deal with detecting the fraud in the handwritten signature. We will focus on several types of ways to sign with the characters. The system will recommend if the hand signature of the user is fake or genuine. This is under the capabilities of GAN to synthesize the signature, it can make the computer automatically generate hand signature by using a machine. Many researchers called this capability is deep-fake. This research aims to learn the hand signature to do fraud detection. We propose an architecture to build the anti-counterfeiting hand signature which is utilized deep learning with a self-growing probabilistic method.","PeriodicalId":103254,"journal":{"name":"2021 7th International Conference on Applied System Innovation (ICASI)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122463986","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 : 2021-09-24DOI: 10.1109/icasi52993.2021.9568434
{"title":"[Title page]","authors":"","doi":"10.1109/icasi52993.2021.9568434","DOIUrl":"https://doi.org/10.1109/icasi52993.2021.9568434","url":null,"abstract":"","PeriodicalId":103254,"journal":{"name":"2021 7th International Conference on Applied System Innovation (ICASI)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123445952","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 : 2021-09-24DOI: 10.1109/ICASI52993.2021.9568448
Angelito A. Silverio, Wen-Yaw Chung, Leandro Silvério
Fall detection technologies kept on advancing to provide better prediction ability, portability and wearability. Existing fallrelated studies include vision-based systems and wearable devices. Though there is already abundant work in the wearable technology space, there is still innovation in making such devices autonomous or standalone. Furthermore, such device could have more functionality embedded to it in one silicon. This served to be the motivation of this work as it attempts to develop a mixed signal system that is fairly stand-alone in terms of fall detection while providing a front-end for heart rate, mean arterial pressure and temperature at post fall condition with inactivity log. For fall detection, a combination of magnitude and spike windowing and thresholding are utilized for a more robust detection. This circuit has been designed using TSMC 0.18um technology obtained from MOSIS wafer test runs and was verified using SPICE.
{"title":"Design of a Mixed Signal Fall Detector System based on a Single 3-D Accelerometer with Dual Mode Physiological Signal Readouts for Stand-Alone Wearable Applications","authors":"Angelito A. Silverio, Wen-Yaw Chung, Leandro Silvério","doi":"10.1109/ICASI52993.2021.9568448","DOIUrl":"https://doi.org/10.1109/ICASI52993.2021.9568448","url":null,"abstract":"Fall detection technologies kept on advancing to provide better prediction ability, portability and wearability. Existing fallrelated studies include vision-based systems and wearable devices. Though there is already abundant work in the wearable technology space, there is still innovation in making such devices autonomous or standalone. Furthermore, such device could have more functionality embedded to it in one silicon. This served to be the motivation of this work as it attempts to develop a mixed signal system that is fairly stand-alone in terms of fall detection while providing a front-end for heart rate, mean arterial pressure and temperature at post fall condition with inactivity log. For fall detection, a combination of magnitude and spike windowing and thresholding are utilized for a more robust detection. This circuit has been designed using TSMC 0.18um technology obtained from MOSIS wafer test runs and was verified using SPICE.","PeriodicalId":103254,"journal":{"name":"2021 7th International Conference on Applied System Innovation (ICASI)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129306282","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 : 2021-09-24DOI: 10.1109/icasi52993.2021.9567901
{"title":"[Title page]","authors":"","doi":"10.1109/icasi52993.2021.9567901","DOIUrl":"https://doi.org/10.1109/icasi52993.2021.9567901","url":null,"abstract":"","PeriodicalId":103254,"journal":{"name":"2021 7th International Conference on Applied System Innovation (ICASI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132064664","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 : 2021-09-24DOI: 10.1109/ICASI52993.2021.9568491
Su-Hua Yang, Jiahua Yang, Zhiyong Chen, C. Ho
An energy efficient dual-WO3 electrode was fabricated for electrochromic window application. A vapor-coated WO3 was prepared using the sputtering deposition method, and a wet-coated WO3 film was prepared with the wet chemical reaction process. The WO3 was amorphous crystalline. The intercalated charge of the electrode with a dual-WO3 layer was 0.0265 C/cm2. High contrast ratio of 55.8% and fast responses of 6 s for charge intercalation and 1.8 s for deintercalation were achieved.
{"title":"Energy Efficient Electrochromic Windows Prepared with Dual WO3-layered Electrodes","authors":"Su-Hua Yang, Jiahua Yang, Zhiyong Chen, C. Ho","doi":"10.1109/ICASI52993.2021.9568491","DOIUrl":"https://doi.org/10.1109/ICASI52993.2021.9568491","url":null,"abstract":"An energy efficient dual-WO<inf>3</inf> electrode was fabricated for electrochromic window application. A vapor-coated WO<inf>3</inf> was prepared using the sputtering deposition method, and a wet-coated WO<inf>3</inf> film was prepared with the wet chemical reaction process. The WO<inf>3</inf> was amorphous crystalline. The intercalated charge of the electrode with a dual-WO<inf>3</inf> layer was 0.0265 C/cm<sup>2</sup>. High contrast ratio of 55.8% and fast responses of 6 s for charge intercalation and 1.8 s for deintercalation were achieved.","PeriodicalId":103254,"journal":{"name":"2021 7th International Conference on Applied System Innovation (ICASI)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127634533","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 : 2021-09-24DOI: 10.1109/ICASI52993.2021.9568429
D. Lin, Han-Lin Chiu, H. Chou
This paper presents a novel design of dipole antenna with H-slots to discuss the radiation characteristics of the antenna. This antenna is designed for unmanned aerial vehicle communication. It has the characteristics of side radiation, and 1.1~1.5GHz is selected as the working frequency band, which is helpful for the unmanned aerial vehicle to carry out long-distance communication. This study reduces the size of the antenna, and uses the surface characteristics of the UAV to match the characteristic impedance of the antenna to improve the performance of the antenna.
{"title":"A Novel Design of Compact Dipole Antenna with H-slots Considering Reflection from UAV Housing","authors":"D. Lin, Han-Lin Chiu, H. Chou","doi":"10.1109/ICASI52993.2021.9568429","DOIUrl":"https://doi.org/10.1109/ICASI52993.2021.9568429","url":null,"abstract":"This paper presents a novel design of dipole antenna with H-slots to discuss the radiation characteristics of the antenna. This antenna is designed for unmanned aerial vehicle communication. It has the characteristics of side radiation, and 1.1~1.5GHz is selected as the working frequency band, which is helpful for the unmanned aerial vehicle to carry out long-distance communication. This study reduces the size of the antenna, and uses the surface characteristics of the UAV to match the characteristic impedance of the antenna to improve the performance of the antenna.","PeriodicalId":103254,"journal":{"name":"2021 7th International Conference on Applied System Innovation (ICASI)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126781031","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}
Plasma electrolytic polishing (PEP) is an attractive innovation process to replace conventional mechanical polishing or electropolishing (EP), because of conventional processes have encountered labor shortages and environmental challenges. In contrast to the EP process which uses acids aggressive media, whereas in PEP, the low-concentration salt solution instead of acids or toxicants as electrolyte with high DC voltage to generate plasma phenomenon is the main process operating characteristics. For further understanding the process characteristic, this study carries series of experiments to observe the process mechanism and also examine the application possibility for kinds of metal workpiece surface finishing.
{"title":"Plasma Electrolytic Polishing Process Mechanism and Application Possibilities Research for Metal Workpiece Surface Finishing","authors":"Chiu-Feng Lin, Zhi-Wen Fan, H. Chen, Yu-Kai Chen, Meilin Liu, Tzu-Hung Chen, Wenchieh Wu","doi":"10.1109/ICASI52993.2021.9568490","DOIUrl":"https://doi.org/10.1109/ICASI52993.2021.9568490","url":null,"abstract":"Plasma electrolytic polishing (PEP) is an attractive innovation process to replace conventional mechanical polishing or electropolishing (EP), because of conventional processes have encountered labor shortages and environmental challenges. In contrast to the EP process which uses acids aggressive media, whereas in PEP, the low-concentration salt solution instead of acids or toxicants as electrolyte with high DC voltage to generate plasma phenomenon is the main process operating characteristics. For further understanding the process characteristic, this study carries series of experiments to observe the process mechanism and also examine the application possibility for kinds of metal workpiece surface finishing.","PeriodicalId":103254,"journal":{"name":"2021 7th International Conference on Applied System Innovation (ICASI)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134441864","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 : 2021-09-24DOI: 10.1109/ICASI52993.2021.9568407
Ahmad Tamim Esmat
Collecting data from remote hydrological and weather stations in Afghanistan has been a severe challenge to the relevant departments, as the lack of internet infrastructures is evident across the country. The staff members commute to the stations regularly to collect the data manually. This manual practice requires constant human labor and substantial financial resources and is inefficient in accessing timely data and managing the massive amount of collected data afterward. Until recent years, some stations were accessible through a satellite link; however, they are now disconnected due to high cost of the service. On the other hand, telecommunication services in Afghanistan had a rapid growth in the last two decades. People even in rural areas now have access to phone call and SMS; however, internet access has remained limited to the cities and populated areas. Hence, this research proposes the SMS solution to address the challenges of collecting data from remote hydrological and weather stations. A survey was conducted to evaluate the coverage of SMS vs. mobile data on 183 stations. A set of lab experiments were also conducted as a proof of concept. The results show that SMS is an applicable solution, considering current conditions in the region.
{"title":"ICT for remote data collection: A study on hydrological and weather stations in Afghanistan","authors":"Ahmad Tamim Esmat","doi":"10.1109/ICASI52993.2021.9568407","DOIUrl":"https://doi.org/10.1109/ICASI52993.2021.9568407","url":null,"abstract":"Collecting data from remote hydrological and weather stations in Afghanistan has been a severe challenge to the relevant departments, as the lack of internet infrastructures is evident across the country. The staff members commute to the stations regularly to collect the data manually. This manual practice requires constant human labor and substantial financial resources and is inefficient in accessing timely data and managing the massive amount of collected data afterward. Until recent years, some stations were accessible through a satellite link; however, they are now disconnected due to high cost of the service. On the other hand, telecommunication services in Afghanistan had a rapid growth in the last two decades. People even in rural areas now have access to phone call and SMS; however, internet access has remained limited to the cities and populated areas. Hence, this research proposes the SMS solution to address the challenges of collecting data from remote hydrological and weather stations. A survey was conducted to evaluate the coverage of SMS vs. mobile data on 183 stations. A set of lab experiments were also conducted as a proof of concept. The results show that SMS is an applicable solution, considering current conditions in the region.","PeriodicalId":103254,"journal":{"name":"2021 7th International Conference on Applied System Innovation (ICASI)","volume":"134 S232","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132905566","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 : 2021-09-24DOI: 10.1109/ICASI52993.2021.9568460
Tsung-Ching Lin, Chih-Hung Lee, Ming-Kun Hsieh, Cheng-Nan Chiu, D. Lin, Hsin-Piao Lin
This paper presents a new type of double ridged horn antenna design for both radiated immunity and emissions test of broadband applications in EMI measurement. The proposed antenna offers excellent performance over the frequency range of 18 to 50 GHz as required by the EMI regulation compliance tests (i.e. FCC, CE, MIL-STD-461, RTCA DO-160, FDA, SAE, etc.). The optimized antenna aperture dimension of flared section of pyramidal structure and improved the ridge profile are yielded the wide antenna beam-width across the bandwidth, lower VSWR, stable gain and symmetric radiation patterns with low side lobe level. Moreover, a coaxial probe structure in the feed section and fitted with by a 2.4 mm connector, while under the increased bandwidth for good impedance matching. To achive the charactistic of this well designed antenna, whose has been implement, measurement, and to reduce the difficulty and cost of daily use in laboratory environments for a standard semi-anechoic chamber.
{"title":"A Broadband Double Ridged Horn Antenna for Radiated Immunity and Emissions Test from 18 GHz to 50 GHz","authors":"Tsung-Ching Lin, Chih-Hung Lee, Ming-Kun Hsieh, Cheng-Nan Chiu, D. Lin, Hsin-Piao Lin","doi":"10.1109/ICASI52993.2021.9568460","DOIUrl":"https://doi.org/10.1109/ICASI52993.2021.9568460","url":null,"abstract":"This paper presents a new type of double ridged horn antenna design for both radiated immunity and emissions test of broadband applications in EMI measurement. The proposed antenna offers excellent performance over the frequency range of 18 to 50 GHz as required by the EMI regulation compliance tests (i.e. FCC, CE, MIL-STD-461, RTCA DO-160, FDA, SAE, etc.). The optimized antenna aperture dimension of flared section of pyramidal structure and improved the ridge profile are yielded the wide antenna beam-width across the bandwidth, lower VSWR, stable gain and symmetric radiation patterns with low side lobe level. Moreover, a coaxial probe structure in the feed section and fitted with by a 2.4 mm connector, while under the increased bandwidth for good impedance matching. To achive the charactistic of this well designed antenna, whose has been implement, measurement, and to reduce the difficulty and cost of daily use in laboratory environments for a standard semi-anechoic chamber.","PeriodicalId":103254,"journal":{"name":"2021 7th International Conference on Applied System Innovation (ICASI)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121918579","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：