Pub Date : 2022-05-16DOI: 10.1109/CITDS54976.2022.9914232
István Csoba, Roland Kunkli
The point-spread function (PSF) is the diffraction pattern of an infinitesimal light source and plays an important role in the study and simulation of human vision. It forms the backbone of a multitude of vision-rendering algorithms, as it can be used to obtain the necessary kernels for convolution. Its computation is often performed via ray-tracing or the fast Fourier transform (FFT), but recently we also demonstrated that the Extended Nijboer-Zernike (ENZ) approach can be a more efficient alternative, which reduces the computation time of large PSF sets to just a few minutes. In this paper, we present a significantly faster, GPU-based computation scheme of the ENZ approach to further improve the computation process for such large PSF sets. Our algorithm works by reformulating the core $V_{n}^{m}$ function to reusable subterms that are efficient to accumulate in parallel. We demonstrate that our proposed method leads to substantial performance improvements and facilitates the interactive exploration of visual aberrations when paired with our existing vision simulation algorithm.
{"title":"Fast, GPU-based Computation of Large Point-Spread Function Sets for the Human Eye using the Extended Nijboer-Zernike Approach","authors":"István Csoba, Roland Kunkli","doi":"10.1109/CITDS54976.2022.9914232","DOIUrl":"https://doi.org/10.1109/CITDS54976.2022.9914232","url":null,"abstract":"The point-spread function (PSF) is the diffraction pattern of an infinitesimal light source and plays an important role in the study and simulation of human vision. It forms the backbone of a multitude of vision-rendering algorithms, as it can be used to obtain the necessary kernels for convolution. Its computation is often performed via ray-tracing or the fast Fourier transform (FFT), but recently we also demonstrated that the Extended Nijboer-Zernike (ENZ) approach can be a more efficient alternative, which reduces the computation time of large PSF sets to just a few minutes. In this paper, we present a significantly faster, GPU-based computation scheme of the ENZ approach to further improve the computation process for such large PSF sets. Our algorithm works by reformulating the core $V_{n}^{m}$ function to reusable subterms that are efficient to accumulate in parallel. We demonstrate that our proposed method leads to substantial performance improvements and facilitates the interactive exploration of visual aberrations when paired with our existing vision simulation algorithm.","PeriodicalId":271992,"journal":{"name":"2022 IEEE 2nd Conference on Information Technology and Data Science (CITDS)","volume":"274 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115667127","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 : 2022-05-16DOI: 10.1109/CITDS54976.2022.9914316
Ermiyas Birihanu, Áron Barcsa-Szabó, I. Lendák
Industrial Control Systems (ICSs) utilize different sensors and various embedded systems to operate. Devices often communicate using protocols like Siemens Step 7 and Modbus, which were designed for use in closed networks many years ago and are vulnerable to attacks. The goal of this study is to detect anomalies in industrial control systems using a proximity-based approach on the Securing Water Treatment (SWaT) dataset. We encoded categorical data using one hot encoding and normalized numerical data using min max scaling. The experiment shown that by adopting a proximity-based approach, we can obtain state-of-the-art 99% precision and 98% recall and able to identify 35 out of 37 attack points, indicating that the suggested methodology is suitable for usage in industrial control system scenarios.
{"title":"Proximity-based anomaly detection in Securing Water Treatment","authors":"Ermiyas Birihanu, Áron Barcsa-Szabó, I. Lendák","doi":"10.1109/CITDS54976.2022.9914316","DOIUrl":"https://doi.org/10.1109/CITDS54976.2022.9914316","url":null,"abstract":"Industrial Control Systems (ICSs) utilize different sensors and various embedded systems to operate. Devices often communicate using protocols like Siemens Step 7 and Modbus, which were designed for use in closed networks many years ago and are vulnerable to attacks. The goal of this study is to detect anomalies in industrial control systems using a proximity-based approach on the Securing Water Treatment (SWaT) dataset. We encoded categorical data using one hot encoding and normalized numerical data using min max scaling. The experiment shown that by adopting a proximity-based approach, we can obtain state-of-the-art 99% precision and 98% recall and able to identify 35 out of 37 attack points, indicating that the suggested methodology is suitable for usage in industrial control system scenarios.","PeriodicalId":271992,"journal":{"name":"2022 IEEE 2nd Conference on Information Technology and Data Science (CITDS)","volume":"160 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117104890","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 : 2022-05-16DOI: 10.1109/CITDS54976.2022.9914350
Chuangtao Ma, B. Molnár, Á. Tarcsi, A. Benczúr
Schema matching is a process of creating the correspondences and mappings from the various schemas, which is a critical phase of migrating and integrating heterogeneous databases from multiple sources. However, the semantic heterogeneity in various schemas brings some obstacles while establishing the correspondences between source schema and target schema, hence human interventions and domain knowledge are required to tackle some complex mapping tasks for heterogeneous data integration. To reduce human intervention and improve the ability to handle complex matching tasks, we present a knowledge-enriched schema matching framework. In this framework, the schema matching task is treated as a classification problem, thereby, a schema matching network is designed as a classifier to give the mapping result. In particular, the external knowledge bases are injected into the schema matching network to capture the background knowledge and provide the common knowledge to handle the semantic heterogeneity of complex mapping tasks. Additionally, the main components of the presented framework and their roles are analyzed, and the feasibility of our framework and the future work are highlighted.
{"title":"Knowledge Enriched Schema Matching Framework for Heterogeneous Data Integration","authors":"Chuangtao Ma, B. Molnár, Á. Tarcsi, A. Benczúr","doi":"10.1109/CITDS54976.2022.9914350","DOIUrl":"https://doi.org/10.1109/CITDS54976.2022.9914350","url":null,"abstract":"Schema matching is a process of creating the correspondences and mappings from the various schemas, which is a critical phase of migrating and integrating heterogeneous databases from multiple sources. However, the semantic heterogeneity in various schemas brings some obstacles while establishing the correspondences between source schema and target schema, hence human interventions and domain knowledge are required to tackle some complex mapping tasks for heterogeneous data integration. To reduce human intervention and improve the ability to handle complex matching tasks, we present a knowledge-enriched schema matching framework. In this framework, the schema matching task is treated as a classification problem, thereby, a schema matching network is designed as a classifier to give the mapping result. In particular, the external knowledge bases are injected into the schema matching network to capture the background knowledge and provide the common knowledge to handle the semantic heterogeneity of complex mapping tasks. Additionally, the main components of the presented framework and their roles are analyzed, and the feasibility of our framework and the future work are highlighted.","PeriodicalId":271992,"journal":{"name":"2022 IEEE 2nd Conference on Information Technology and Data Science (CITDS)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124402163","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 : 2022-05-16DOI: 10.1109/CITDS54976.2022.9914386
Máté Vágner, Dénes Palkovics, László Kovács
Localization and position estimation are crucial tasks in autonomous driving. In addition to the importance of positioning, with good quality position tracking, it is possible to implement sophisticated data collection procedures and use advanced machine learning methods such as reinforced learning. Global navigation satellite systems offer very accurate positioning, but their use is cumbersome or impossible under certain laboratory conditions. In our work, we applied an indoor positioning system, that is integrated into our 1:16 self-driving car.
{"title":"3D Localization and Data Quality Estimation with Marvelmind","authors":"Máté Vágner, Dénes Palkovics, László Kovács","doi":"10.1109/CITDS54976.2022.9914386","DOIUrl":"https://doi.org/10.1109/CITDS54976.2022.9914386","url":null,"abstract":"Localization and position estimation are crucial tasks in autonomous driving. In addition to the importance of positioning, with good quality position tracking, it is possible to implement sophisticated data collection procedures and use advanced machine learning methods such as reinforced learning. Global navigation satellite systems offer very accurate positioning, but their use is cumbersome or impossible under certain laboratory conditions. In our work, we applied an indoor positioning system, that is integrated into our 1:16 self-driving car.","PeriodicalId":271992,"journal":{"name":"2022 IEEE 2nd Conference on Information Technology and Data Science (CITDS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129555333","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 : 2019-09-25DOI: 10.1109/CITDS54976.2022.9914244
Adrián Csiszárik, Beatrix Benko, D. Varga
Modern deep artificial neural networks have achieved great success in the domain of computer vision and beyond. However, their application to many real-world tasks is undermined by certain limitations, such as overconfident uncertainty estimates on out-of-distribution data or performance deterioration under data distribution shifts. Several types of deep learning models used for density estimation through probabilistic generative modeling have been shown to fail to detect out-of-distribution samples by assigning higher likelihoods to anomalous data. We investigate this failure mode in Variational Autoencoder models, which are also prone to this, and improve upon the out-of-distribution generalization performance of the model by employing an alternative training scheme utilizing negative samples. We present a fully unsupervised version: when the model is trained in an adversarial manner, the generator’s own outputs can be used as negative samples. We demonstrate empirically the effectiveness of the approach in reducing the overconfident likelihood estimates of out-of-distribution inputs on image data.
{"title":"Negative Sampling in Variational Autoencoders","authors":"Adrián Csiszárik, Beatrix Benko, D. Varga","doi":"10.1109/CITDS54976.2022.9914244","DOIUrl":"https://doi.org/10.1109/CITDS54976.2022.9914244","url":null,"abstract":"Modern deep artificial neural networks have achieved great success in the domain of computer vision and beyond. However, their application to many real-world tasks is undermined by certain limitations, such as overconfident uncertainty estimates on out-of-distribution data or performance deterioration under data distribution shifts. Several types of deep learning models used for density estimation through probabilistic generative modeling have been shown to fail to detect out-of-distribution samples by assigning higher likelihoods to anomalous data. We investigate this failure mode in Variational Autoencoder models, which are also prone to this, and improve upon the out-of-distribution generalization performance of the model by employing an alternative training scheme utilizing negative samples. We present a fully unsupervised version: when the model is trained in an adversarial manner, the generator’s own outputs can be used as negative samples. We demonstrate empirically the effectiveness of the approach in reducing the overconfident likelihood estimates of out-of-distribution inputs on image data.","PeriodicalId":271992,"journal":{"name":"2022 IEEE 2nd Conference on Information Technology and Data Science (CITDS)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131443524","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 : 2016-06-01DOI: 10.1109/CITDS54976.2022.9914043
B. Almási, T. Bérczes, A. Kuki, J. Sztrik, Jinting Wang
Understanding the impatient behaviour of users and customers has a critical importance for every organization to remain at the forefront in today’s competitive business world. Customers’ most prevalent impatient behaviours are balking and reneging. Customers are discouraged about receiving service when they notice large queues ahead (balking); they may even exit the system after joining if their wait time exceeds expectations (reneging). Nevertheless, in the investment-related industry, the opposite of balking is true, the desire to join a business is great if the number of customers is high, as this can be a very attractive factor for new investors. If the number of existing clients is large, the possibility of connecting to such a business is significant. Thus, the more crowded the system, the more joiners and vice versa (reverse balking).In this article, we study the concepts of reneging and reverse balking in the context of a Cognitive Radio Network. The more crowded our network is, the more likely new calls join, and vice versa. These calls, may also get irritated and abandon the whole system as a result of a lengthy delay. The system’s key performance measures are visually illustrated and acquired using simulation.
{"title":"Performance modeling of finite-source cognitive radio networks with reverse balking and reneging using simulation","authors":"B. Almási, T. Bérczes, A. Kuki, J. Sztrik, Jinting Wang","doi":"10.1109/CITDS54976.2022.9914043","DOIUrl":"https://doi.org/10.1109/CITDS54976.2022.9914043","url":null,"abstract":"Understanding the impatient behaviour of users and customers has a critical importance for every organization to remain at the forefront in today’s competitive business world. Customers’ most prevalent impatient behaviours are balking and reneging. Customers are discouraged about receiving service when they notice large queues ahead (balking); they may even exit the system after joining if their wait time exceeds expectations (reneging). Nevertheless, in the investment-related industry, the opposite of balking is true, the desire to join a business is great if the number of customers is high, as this can be a very attractive factor for new investors. If the number of existing clients is large, the possibility of connecting to such a business is significant. Thus, the more crowded the system, the more joiners and vice versa (reverse balking).In this article, we study the concepts of reneging and reverse balking in the context of a Cognitive Radio Network. The more crowded our network is, the more likely new calls join, and vice versa. These calls, may also get irritated and abandon the whole system as a result of a lengthy delay. The system’s key performance measures are visually illustrated and acquired using simulation.","PeriodicalId":271992,"journal":{"name":"2022 IEEE 2nd Conference on Information Technology and Data Science (CITDS)","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114989506","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：