Manufacturing companies widely use interactive configurators to specify product variants. These variants are created from different combinations of individual components. Constraints are usually imposed as relations between components to rule out invalid product configurations. The configurator and its underlying algorithm ensure the product specified by the user satisfies all constraints. We provide a method that models the online configurator as a series of highly parallelizable boolean satisfiability problems (SAT). Our methodology is facilitated by state-of-the-art tools such as the Microsoft Z3 theorem prover. Additionally, we confirm that the configurator problem is NP-complete. Hence, using SAT solving is not only natural but canonical.
{"title":"Interactive Online Configurator via Boolean Satisfiability Modeling","authors":"Tao Tao, D. Plaisted","doi":"10.1145/3468784.3468785","DOIUrl":"https://doi.org/10.1145/3468784.3468785","url":null,"abstract":"Manufacturing companies widely use interactive configurators to specify product variants. These variants are created from different combinations of individual components. Constraints are usually imposed as relations between components to rule out invalid product configurations. The configurator and its underlying algorithm ensure the product specified by the user satisfies all constraints. We provide a method that models the online configurator as a series of highly parallelizable boolean satisfiability problems (SAT). Our methodology is facilitated by state-of-the-art tools such as the Microsoft Z3 theorem prover. Additionally, we confirm that the configurator problem is NP-complete. Hence, using SAT solving is not only natural but canonical.","PeriodicalId":341589,"journal":{"name":"The 12th International Conference on Advances in Information Technology","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131270058","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}
Tawatchai Perdpunya, Siranee Nuchitprasitchai, P. Boonrawd
A machine is an essential factor for industrial production. Industry 4.0 is the revolution that causes improvement of machines to have higher efficiency. Accordingly, inspection and maintenance are becoming more important. However, most of factories are not changed the operating process, there is no data logging for evaluation and analysis for preventive maintenance. This research aims to develop a model for machine inspection using augmented reality with object detection and marker techniques on real world machines and mask R-CNN algorithm allowing inspector to perform inspections. This study, we demonstrate the process of development of the proposed model by showing steps of data acquisition from a machine in a factory. The dataset is images of machines in different perspectives, and they were used for training and testing the model. The testing is done on a mobile device of an inspector. With computer vision technique and the proposed model, the instant precision tracking and detection are provided. Then the trained model is transferred to the mobile devices for testing without any modification by an expert. Some images of machines are randomly selected to verify the accuracy of the model. The result shows that the efficiency of the model is acceptable in real usage.
{"title":"Augmented Reality with Mask R-CNN (ARR-CNN) inspection for Intelligent Manufacturing","authors":"Tawatchai Perdpunya, Siranee Nuchitprasitchai, P. Boonrawd","doi":"10.1145/3468784.3468788","DOIUrl":"https://doi.org/10.1145/3468784.3468788","url":null,"abstract":"A machine is an essential factor for industrial production. Industry 4.0 is the revolution that causes improvement of machines to have higher efficiency. Accordingly, inspection and maintenance are becoming more important. However, most of factories are not changed the operating process, there is no data logging for evaluation and analysis for preventive maintenance. This research aims to develop a model for machine inspection using augmented reality with object detection and marker techniques on real world machines and mask R-CNN algorithm allowing inspector to perform inspections. This study, we demonstrate the process of development of the proposed model by showing steps of data acquisition from a machine in a factory. The dataset is images of machines in different perspectives, and they were used for training and testing the model. The testing is done on a mobile device of an inspector. With computer vision technique and the proposed model, the instant precision tracking and detection are provided. Then the trained model is transferred to the mobile devices for testing without any modification by an expert. Some images of machines are randomly selected to verify the accuracy of the model. The result shows that the efficiency of the model is acceptable in real usage.","PeriodicalId":341589,"journal":{"name":"The 12th International Conference on Advances in Information Technology","volume":"102 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132741333","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}
A call center is an important communication channel between a business and its customers. The call center staffs are responsible for resolving customer problems and fulfilling their needs. It is undeniable that customer satisfaction can be increased if personalized services relating to their characteristics such as personality are provided. Researches are suggesting that a person's personality can be recognized from his/her conversational voice. Thus, a machine learning model that recognizes each customer's personality from one's conversational voice in a call center would enable the cell center to give that one appropriate response. This study focuses on developing a personality recognition model to predict each MPI (Maudsley Personality Inventory) personality dimension from each conversational voice. The MPI personality dimension includes E-scale (representing extraversion and introversion) and N-scale (representing neuroticism and stability). Audio files of conversational voice were collected from 92 volunteers instructed to make conversation in the simulated call center context. Logistic regression, LinearSVC, Random forest, and Artificial neural networks were used in the modeling process. The result shows that the model generated by using Artificial neural networks has the best performance on predicting the E-scale. The model has the positive predictive value (Introversion prediction) and the negative predictive value (Extraversion prediction) equal to 0.71 and 0.75, respectively. No model shows satisfying performance on neuroticism and stability prediction. This study shows a piece of evidence that extraversion and introversion in MPI, which have implications in businesses, can be effectively recognized from each person's conversational voice made through call centers. The model can be beneficial in many business applications such as call center management, personalized product offering, and personalized advertisement.
{"title":"A Development of Personality Recognition Model from Conversation Voice in Call Center Context","authors":"Nakorn Srinarong, J. Mongkolnavin","doi":"10.1145/3468784.3469180","DOIUrl":"https://doi.org/10.1145/3468784.3469180","url":null,"abstract":"A call center is an important communication channel between a business and its customers. The call center staffs are responsible for resolving customer problems and fulfilling their needs. It is undeniable that customer satisfaction can be increased if personalized services relating to their characteristics such as personality are provided. Researches are suggesting that a person's personality can be recognized from his/her conversational voice. Thus, a machine learning model that recognizes each customer's personality from one's conversational voice in a call center would enable the cell center to give that one appropriate response. This study focuses on developing a personality recognition model to predict each MPI (Maudsley Personality Inventory) personality dimension from each conversational voice. The MPI personality dimension includes E-scale (representing extraversion and introversion) and N-scale (representing neuroticism and stability). Audio files of conversational voice were collected from 92 volunteers instructed to make conversation in the simulated call center context. Logistic regression, LinearSVC, Random forest, and Artificial neural networks were used in the modeling process. The result shows that the model generated by using Artificial neural networks has the best performance on predicting the E-scale. The model has the positive predictive value (Introversion prediction) and the negative predictive value (Extraversion prediction) equal to 0.71 and 0.75, respectively. No model shows satisfying performance on neuroticism and stability prediction. This study shows a piece of evidence that extraversion and introversion in MPI, which have implications in businesses, can be effectively recognized from each person's conversational voice made through call centers. The model can be beneficial in many business applications such as call center management, personalized product offering, and personalized advertisement.","PeriodicalId":341589,"journal":{"name":"The 12th International Conference on Advances in Information Technology","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131314736","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}
Privacy violation is a serious issue that must be considered when datasets are released for public use. To address this issue, a well-known privacy preservation model, l-Diversity, is proposed. Unfortunately, l-Diversity is generally proposed to address privacy violation issues in datasets that are focused on performing one-time data releasing. For this reason, l-Diversity could be inadequate to preserve the privacy data if datasets are dynamic and released at all times. To rid this vulnerability of l-Diversity, a new privacy preservation model for sequential data releasing to be proposed in this work, so called as ε-Error and l-Diversity. Aside from privacy preservation constraints, the complexity and the data utility are also maintained in the privacy preservation constraint of the proposed model.
隐私侵犯是一个严重的问题，当数据集发布给公众使用时必须考虑。为了解决这个问题，我们提出了一个著名的隐私保护模型——l-Diversity。不幸的是，l-Diversity通常被提议用于解决数据集中的隐私侵犯问题，这些数据集中于执行一次性数据发布。由于这个原因，如果数据集是动态的，并且在任何时候都是发布的，那么l-Diversity可能不足以保护隐私数据。为了消除l-Diversity的这一漏洞，本文提出了一种新的序列数据发布隐私保护模型ε-Error and l-Diversity。除了隐私保护约束外，该模型还保持了隐私保护约束的复杂性和数据实用性。
{"title":"Privacy Preservation Techniques for Sequential Data Releasing","authors":"Surapon Riyana, Noppamas Riyana, Srikul Nanthachumphu","doi":"10.1145/3468784.3470468","DOIUrl":"https://doi.org/10.1145/3468784.3470468","url":null,"abstract":"Privacy violation is a serious issue that must be considered when datasets are released for public use. To address this issue, a well-known privacy preservation model, l-Diversity, is proposed. Unfortunately, l-Diversity is generally proposed to address privacy violation issues in datasets that are focused on performing one-time data releasing. For this reason, l-Diversity could be inadequate to preserve the privacy data if datasets are dynamic and released at all times. To rid this vulnerability of l-Diversity, a new privacy preservation model for sequential data releasing to be proposed in this work, so called as ε-Error and l-Diversity. Aside from privacy preservation constraints, the complexity and the data utility are also maintained in the privacy preservation constraint of the proposed model.","PeriodicalId":341589,"journal":{"name":"The 12th International Conference on Advances in Information Technology","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114537086","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}
Manifold learning has been an important research area as from literature it is evident that patterns in most real-life data sets can be embedded in low-dimensional space while maintaining the original structure of high-dimensional space. This work concentrates on one of the major research areas of manifold learning, which is the segregation of manifolds where more than one non-intersecting manifolds are present. The proposed method presents a solution to the problem by detecting the number of manifolds in a dataset using the Laplacian graph matrix and segregate the manifolds using agglomerative clustering. Eventually, locally linear embedding has been used for dimensionality reduction of every individual manifold in such a way that manifolds remain segregated and also holds the original global structure. The proposed method achieves finer results when applied on benchmark synthetic data sets SCurve, SwissRoll, Helix and real-life datasets COIL-20, optical digit recognition, att_faces, extended Yale Face Database B. While the state of the art methods fails to detect the number of manifolds in a dataset, the proposed method not only eclipses the performance of them but also carry the separable structure in the lower dimensional space.
{"title":"Unsupervised Segmentation of Non-Intersecting Manifolds","authors":"Subhadip Boral, Sumedha Dhar, Ashish Ghosh","doi":"10.1145/3468784.3470467","DOIUrl":"https://doi.org/10.1145/3468784.3470467","url":null,"abstract":"Manifold learning has been an important research area as from literature it is evident that patterns in most real-life data sets can be embedded in low-dimensional space while maintaining the original structure of high-dimensional space. This work concentrates on one of the major research areas of manifold learning, which is the segregation of manifolds where more than one non-intersecting manifolds are present. The proposed method presents a solution to the problem by detecting the number of manifolds in a dataset using the Laplacian graph matrix and segregate the manifolds using agglomerative clustering. Eventually, locally linear embedding has been used for dimensionality reduction of every individual manifold in such a way that manifolds remain segregated and also holds the original global structure. The proposed method achieves finer results when applied on benchmark synthetic data sets SCurve, SwissRoll, Helix and real-life datasets COIL-20, optical digit recognition, att_faces, extended Yale Face Database B. While the state of the art methods fails to detect the number of manifolds in a dataset, the proposed method not only eclipses the performance of them but also carry the separable structure in the lower dimensional space.","PeriodicalId":341589,"journal":{"name":"The 12th International Conference on Advances in Information Technology","volume":"79 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124840835","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}
RFID is a smart label technology that is used in several real-life applications such as inventory management, asset tracking, personnel tracking, controlling access to restricted areas, ID badging, supply chain management, counterfeit prevention (e.g., in the pharmaceutical industry), and smart farming. Generally, the data collection of RFIDs consists of the users’ visited locations and their visiting time, so called as trajectory datasets. Aside from applications, trajectory datasets can also be released for public use. For this reason, they could lead to being privacy violation issues. To address these issues in trajectory datasets, LKC-Privacy is proposed. Unfortunately, in this work, we demonstrate that LKC-Privacy still has a serious vulnerability that must be improved. To rid the demonstrated vulnerability of LKC-Privacy, a privacy preservation model is proposed in this work. Furthermore, the proposed mode is evaluated by extensive experiments. From the experimental results, they indicate that the proposed model is highly secure and more efficient than LKC-Privacy.
{"title":"A Privacy Preservation Model for RFID Data-Collections is Highly Secure and More Efficient than LKC-Privacy","authors":"Surapon Riyana, Noppamas Riyana","doi":"10.1145/3468784.3469853","DOIUrl":"https://doi.org/10.1145/3468784.3469853","url":null,"abstract":"RFID is a smart label technology that is used in several real-life applications such as inventory management, asset tracking, personnel tracking, controlling access to restricted areas, ID badging, supply chain management, counterfeit prevention (e.g., in the pharmaceutical industry), and smart farming. Generally, the data collection of RFIDs consists of the users’ visited locations and their visiting time, so called as trajectory datasets. Aside from applications, trajectory datasets can also be released for public use. For this reason, they could lead to being privacy violation issues. To address these issues in trajectory datasets, LKC-Privacy is proposed. Unfortunately, in this work, we demonstrate that LKC-Privacy still has a serious vulnerability that must be improved. To rid the demonstrated vulnerability of LKC-Privacy, a privacy preservation model is proposed in this work. Furthermore, the proposed mode is evaluated by extensive experiments. From the experimental results, they indicate that the proposed model is highly secure and more efficient than LKC-Privacy.","PeriodicalId":341589,"journal":{"name":"The 12th International Conference on Advances in Information Technology","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127357556","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}
Precipitation nowcasting with ground-based weather radars and satellite-based precipitation data based on deep learning method will open a new avenue of weather prediction. However, it is limited to regions where ground-based weather radars can operate for nowcasting. We propose an improved deep learning precipitation prediction by integrating the precipitation data from Japan Aerospace Exploration Agency (JAXA)’s Global Rainfall Watch (GSMAP) with the precipitation data from WEATHERNEWS Co., Ltd., which provides precipitation data with Type C Doppler radars that detect precipitation in the atmosphere. It has been demonstrated that our proposed method can improve precipitation data coverage areas and the efficiency of precipitation nowcasting by the proposed deep learning technique in many extreme weather cases, i.e., typhoons.
{"title":"Precipitation Nowcasting Using Deep Learning on Radar Data Augmented with Satellite Data","authors":"Wikom Tosiri, Nutnaree Kleawsirikul, Patamawadee Leepaisomboon, Natnapat Gaviphatt, Hidetomo Sakaino, P. Vateekul","doi":"10.1145/3468784.3470469","DOIUrl":"https://doi.org/10.1145/3468784.3470469","url":null,"abstract":"Precipitation nowcasting with ground-based weather radars and satellite-based precipitation data based on deep learning method will open a new avenue of weather prediction. However, it is limited to regions where ground-based weather radars can operate for nowcasting. We propose an improved deep learning precipitation prediction by integrating the precipitation data from Japan Aerospace Exploration Agency (JAXA)’s Global Rainfall Watch (GSMAP) with the precipitation data from WEATHERNEWS Co., Ltd., which provides precipitation data with Type C Doppler radars that detect precipitation in the atmosphere. It has been demonstrated that our proposed method can improve precipitation data coverage areas and the efficiency of precipitation nowcasting by the proposed deep learning technique in many extreme weather cases, i.e., typhoons.","PeriodicalId":341589,"journal":{"name":"The 12th International Conference on Advances in Information Technology","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126649438","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}
Phongsathorn Kittiworapanya, Kitsuchart Pasupa, P. Auer
We assessed several state-of-the-art deep learning algorithms and computer vision techniques for estimating the particle size of mixed commercial waste from images. In waste management, the first step is often coarse shredding, using the particle size to set up the shredder machine. The difficulty is separating the waste particles in an image, which can not be performed well. This work focused on estimating size by using the texture from the input image, captured at a fixed height from the camera lens to the ground. We found that EfficientNet achieved the best performance of 0.72 on F1-Score and 75.89% on accuracy.
{"title":"Particle Size Estimation in Mixed Commercial Waste Images Using Deep Learning","authors":"Phongsathorn Kittiworapanya, Kitsuchart Pasupa, P. Auer","doi":"10.1145/3468784.3471273","DOIUrl":"https://doi.org/10.1145/3468784.3471273","url":null,"abstract":"We assessed several state-of-the-art deep learning algorithms and computer vision techniques for estimating the particle size of mixed commercial waste from images. In waste management, the first step is often coarse shredding, using the particle size to set up the shredder machine. The difficulty is separating the waste particles in an image, which can not be performed well. This work focused on estimating size by using the texture from the input image, captured at a fixed height from the camera lens to the ground. We found that EfficientNet achieved the best performance of 0.72 on F1-Score and 75.89% on accuracy.","PeriodicalId":341589,"journal":{"name":"The 12th International Conference on Advances in Information Technology","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133067738","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}
Kanlayanee Kaweesinsakul, Siranee Nuchitprasitchai, Joshua M. Pearce
There is a growing interest in cactus cultivation because of numerous cacti uses from houseplants to food and medicinal applications. Various diseases impact the growth of cacti. To develop an automated model for the analysis of cactus disease and to be able to quickly treat and prevent damage to the cactus. The Faster R-CNN and YOLO algorithm technique were used to analyze cactus diseases automatically distributed into six groups: 1) anthracnose, 2) canker, 3) lack of care, 4) aphid, 5) rusts and 6) normal group. Based on the experimental results the YOLOv5 algorithm was found to be more effective at detecting and identifying cactus disease than the Faster R-CNN algorithm. Data training and testing with YOLOv5S model resulted in a precision of 89.7% and an accuracy (recall) of 98.5%, which is effective enough for further use in a number of applications in cactus cultivation. Overall the YOLOv5 algorithm had a test time per image of only 26 milliseconds. Therefore, the YOLOv5 algorithm was found to suitable for mobile applications and this model could be further developed into a program for analyzing cactus disease.
{"title":"Open source disease analysis system of cactus by artificial intelligence and image processing","authors":"Kanlayanee Kaweesinsakul, Siranee Nuchitprasitchai, Joshua M. Pearce","doi":"10.1145/3468784.3469075","DOIUrl":"https://doi.org/10.1145/3468784.3469075","url":null,"abstract":"There is a growing interest in cactus cultivation because of numerous cacti uses from houseplants to food and medicinal applications. Various diseases impact the growth of cacti. To develop an automated model for the analysis of cactus disease and to be able to quickly treat and prevent damage to the cactus. The Faster R-CNN and YOLO algorithm technique were used to analyze cactus diseases automatically distributed into six groups: 1) anthracnose, 2) canker, 3) lack of care, 4) aphid, 5) rusts and 6) normal group. Based on the experimental results the YOLOv5 algorithm was found to be more effective at detecting and identifying cactus disease than the Faster R-CNN algorithm. Data training and testing with YOLOv5S model resulted in a precision of 89.7% and an accuracy (recall) of 98.5%, which is effective enough for further use in a number of applications in cactus cultivation. Overall the YOLOv5 algorithm had a test time per image of only 26 milliseconds. Therefore, the YOLOv5 algorithm was found to suitable for mobile applications and this model could be further developed into a program for analyzing cactus disease.","PeriodicalId":341589,"journal":{"name":"The 12th International Conference on Advances in Information Technology","volume":"86 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132504881","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：