Pub Date : 2023-01-30DOI: 10.5121/ijaia.2023.14101
Simisani Ndaba, E. Thuma, G. Mosweunyane
In this paper, an investigation was done to identify writing style features that can be used for cross-topic and cross-genre documents in the Authorship Identification task from 2003 to 2015. Different writing style features were empirically evaluated that were previously used in single topic and single genre documents for Authorship Identification to determine whether they can be used effectively for cross-topic and crossgenre Authorship Identification using an ablation process. The dataset used was taken from the 2015 PAN CLEF Forum English collection consisting of 100 sets. Furthermore, it was investigated whether combining some of these feature sets can help improve the authorship identification task. Three different classifiers were used: Naïve Bayes, Support Vector Machine, and Random Forest. The results suggest that a combination of a lexical, syntactical, structural, and content feature set can be used effectively for cross topic and cross genre authorship identification, as it achieved an AUC result of 0.837.
{"title":"An Authorship Identification Empirical Evaluation of Writing Style Features in Cross-Topic and Cross-genre Documents","authors":"Simisani Ndaba, E. Thuma, G. Mosweunyane","doi":"10.5121/ijaia.2023.14101","DOIUrl":"https://doi.org/10.5121/ijaia.2023.14101","url":null,"abstract":"In this paper, an investigation was done to identify writing style features that can be used for cross-topic and cross-genre documents in the Authorship Identification task from 2003 to 2015. Different writing style features were empirically evaluated that were previously used in single topic and single genre documents for Authorship Identification to determine whether they can be used effectively for cross-topic and crossgenre Authorship Identification using an ablation process. The dataset used was taken from the 2015 PAN CLEF Forum English collection consisting of 100 sets. Furthermore, it was investigated whether combining some of these feature sets can help improve the authorship identification task. Three different classifiers were used: Naïve Bayes, Support Vector Machine, and Random Forest. The results suggest that a combination of a lexical, syntactical, structural, and content feature set can be used effectively for cross topic and cross genre authorship identification, as it achieved an AUC result of 0.837.","PeriodicalId":391502,"journal":{"name":"International Journal of Artificial Intelligence & Applications","volume":"104 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122588434","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 : 2022-11-30DOI: 10.5121/ijaia.2022.13602
Wei Yao, Jingyi Gu, Wenlu Du, Fadi P. Deek, Guiling Wang
The increasing popularity of crypto assets has resulted in greater cryptocurrency investor interest and more exposure in both industry and academia. Despite the substantial socioeconomic benefits, the anonymous character of cryptocurrency trading makes it prone to abuse and a magnet for illicit purposes, which cause monetary losses for individual traders and erosion in the standing of the tokenomics industry. To regulate the illicit behavior and secure users' privacy for cryptocurrency trading, we present an Anomaly Detection and Privacy-Preserving (ADPP) Framework integrating blockchain and deep learning technologies. Specifically, ADPP leverages blockchain technologies to build a user management platform that ensures anonymity and enhances the privacy-preservation of user information. Atop the user management system, an Anomaly Detection System adapts neural networks and imbalanced learning on topological cryptocurrency flow among users to identify anomalous addresses and maintain a sanction list repository. The experiments on the real-world dataset demonstrate the effectiveness and superior performance of ADPP. The flexible framework can be easily generalized to the crypto assets with public real-time transaction (e.g., Non-fungible Token), which takes up a significant proportion of market capitalization in the domain of tokenomics.
{"title":"ADPP: A Novel Anomaly Detection and Privacy-Preserving Framework using Blockchain and Neural Networks in Tokenomics","authors":"Wei Yao, Jingyi Gu, Wenlu Du, Fadi P. Deek, Guiling Wang","doi":"10.5121/ijaia.2022.13602","DOIUrl":"https://doi.org/10.5121/ijaia.2022.13602","url":null,"abstract":"The increasing popularity of crypto assets has resulted in greater cryptocurrency investor interest and more exposure in both industry and academia. Despite the substantial socioeconomic benefits, the anonymous character of cryptocurrency trading makes it prone to abuse and a magnet for illicit purposes, which cause monetary losses for individual traders and erosion in the standing of the tokenomics industry. To regulate the illicit behavior and secure users' privacy for cryptocurrency trading, we present an Anomaly Detection and Privacy-Preserving (ADPP) Framework integrating blockchain and deep learning technologies. Specifically, ADPP leverages blockchain technologies to build a user management platform that ensures anonymity and enhances the privacy-preservation of user information. Atop the user management system, an Anomaly Detection System adapts neural networks and imbalanced learning on topological cryptocurrency flow among users to identify anomalous addresses and maintain a sanction list repository. The experiments on the real-world dataset demonstrate the effectiveness and superior performance of ADPP. The flexible framework can be easily generalized to the crypto assets with public real-time transaction (e.g., Non-fungible Token), which takes up a significant proportion of market capitalization in the domain of tokenomics.","PeriodicalId":391502,"journal":{"name":"International Journal of Artificial Intelligence & Applications","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126742420","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 : 2022-11-30DOI: 10.5121/ijaia.2022.13604
M. Pazzani, Severine Soltani, Sateesh Kumar, Kamran Alipour, Aadil Ahamed
In explainable AI (XAI) for deep learning, saliency maps, heatmaps, or attention maps are commonly used to identify important regions for the classification of images of explanations. We address two important limitations of heatmaps. First, they do not correspond to type of explanations typically produced by human experts. Second, recent research has shown that many common XAI methods do not accurately identify the regions that human experts consider important. We propose using multitask learning to identify diagnostic features in images and averaging explanations from ensembles of learners to increase the accuracy of explanations. Our technique is general and can be used with multiple deep learning architectures and multiple XAI algorithms. We show that this method decreases the difference between regions of interest of XAI algorithms and those identified by human experts and the multitask learning supports the type of explanations produced by human experts. Furthermore, we show that human experts prefer the explanations produced by ensembles to those of individual networks.
{"title":"Improving Explanations of Image Classifiers: Ensembles and Multitask Learning","authors":"M. Pazzani, Severine Soltani, Sateesh Kumar, Kamran Alipour, Aadil Ahamed","doi":"10.5121/ijaia.2022.13604","DOIUrl":"https://doi.org/10.5121/ijaia.2022.13604","url":null,"abstract":"In explainable AI (XAI) for deep learning, saliency maps, heatmaps, or attention maps are commonly used to identify important regions for the classification of images of explanations. We address two important limitations of heatmaps. First, they do not correspond to type of explanations typically produced by human experts. Second, recent research has shown that many common XAI methods do not accurately identify the regions that human experts consider important. We propose using multitask learning to identify diagnostic features in images and averaging explanations from ensembles of learners to increase the accuracy of explanations. Our technique is general and can be used with multiple deep learning architectures and multiple XAI algorithms. We show that this method decreases the difference between regions of interest of XAI algorithms and those identified by human experts and the multitask learning supports the type of explanations produced by human experts. Furthermore, we show that human experts prefer the explanations produced by ensembles to those of individual networks.","PeriodicalId":391502,"journal":{"name":"International Journal of Artificial Intelligence & Applications","volume":"15 3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127974208","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 : 2022-11-30DOI: 10.5121/ijaia.2022.13601
Wenxin Tian
Cartoons are an important art style, which not only has a unique drawing effect but also reflects the character itself, which is gradually loved by people. With the development of image processing technology, people's research on image research is no longer limited to image recognition, target detection, and tracking, but also images In this paper, we use deep learning based image processing to generate cartoon caricatures of human faces. Therefore, this paper investigates the use of deep learning-based methods to learn face features and convert image styles while preserving the original content features, to automatically generate natural cartoon avatars. In this paper, we study a face cartoon generation method based on content invariance. In the task of image style conversion, the content is fused with different style features based on the invariance of content information, to achieve the style conversion.
{"title":"Converting Real Human Avatar to Cartoon Avatar Utilizing CycleGAN","authors":"Wenxin Tian","doi":"10.5121/ijaia.2022.13601","DOIUrl":"https://doi.org/10.5121/ijaia.2022.13601","url":null,"abstract":"Cartoons are an important art style, which not only has a unique drawing effect but also reflects the character itself, which is gradually loved by people. With the development of image processing technology, people's research on image research is no longer limited to image recognition, target detection, and tracking, but also images In this paper, we use deep learning based image processing to generate cartoon caricatures of human faces. Therefore, this paper investigates the use of deep learning-based methods to learn face features and convert image styles while preserving the original content features, to automatically generate natural cartoon avatars. In this paper, we study a face cartoon generation method based on content invariance. In the task of image style conversion, the content is fused with different style features based on the invariance of content information, to achieve the style conversion.","PeriodicalId":391502,"journal":{"name":"International Journal of Artificial Intelligence & Applications","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123394283","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 : 2022-11-30DOI: 10.5121/ijaia.2022.13603
Madhura Prakash M, Deepthi K Prasad, Meghna S Kulkarni, Spoorthi K, V. S.
Deep learning models are applied seamlessly across various computer vision tasks like object detection, object tracking, scene understanding and further. The application of cutting-edge deep learning (DL) models like U-Net in the classification and segmentation of medical images on different modalities has established significant results in the past few years. Ocular diseases like Diabetic Retinopathy (DR), Glaucoma, Age-Related Macular Degeneration (AMD / ARMD), Hypertensive Retina (HR), Cataract, and dry eyes can be detected at the early stages of disease onset by capturing the fundus image or the anterior image of the subject’s eye. Early detection is key to seeking early treatment and thereby preventing the disease progression, which in some cases may lead to blindness. There is a plethora of deep learning models available which have established significant results in medical image processing and specifically in ocular disease detection. A given task can be solved by using a variety of models and or a combination of them. Deep learning models can be computationally expensive and deploying them on an edge device may be a challenge. This paper provides a comprehensive report and critical evaluation of the various deep learning architectures that can be used to segment and classify ocular diseases namely Glaucoma and Hypertensive Retina on the posterior images of the eye. This review also compares the models based on complexity and edge deployability.
{"title":"A Systematic Study of Deep Learning Architectures for Analysis of Glaucoma and Hypertensive Retinopathy","authors":"Madhura Prakash M, Deepthi K Prasad, Meghna S Kulkarni, Spoorthi K, V. S.","doi":"10.5121/ijaia.2022.13603","DOIUrl":"https://doi.org/10.5121/ijaia.2022.13603","url":null,"abstract":"Deep learning models are applied seamlessly across various computer vision tasks like object detection, object tracking, scene understanding and further. The application of cutting-edge deep learning (DL) models like U-Net in the classification and segmentation of medical images on different modalities has established significant results in the past few years. Ocular diseases like Diabetic Retinopathy (DR), Glaucoma, Age-Related Macular Degeneration (AMD / ARMD), Hypertensive Retina (HR), Cataract, and dry eyes can be detected at the early stages of disease onset by capturing the fundus image or the anterior image of the subject’s eye. Early detection is key to seeking early treatment and thereby preventing the disease progression, which in some cases may lead to blindness. There is a plethora of deep learning models available which have established significant results in medical image processing and specifically in ocular disease detection. A given task can be solved by using a variety of models and or a combination of them. Deep learning models can be computationally expensive and deploying them on an edge device may be a challenge. This paper provides a comprehensive report and critical evaluation of the various deep learning architectures that can be used to segment and classify ocular diseases namely Glaucoma and Hypertensive Retina on the posterior images of the eye. This review also compares the models based on complexity and edge deployability.","PeriodicalId":391502,"journal":{"name":"International Journal of Artificial Intelligence & Applications","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117045365","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 : 2022-09-30DOI: 10.5121/ijaia.2022.13501
Kennedy E. Ehimwenma, Safiya Al Sharji, M. Raheem
The probability of an event is in the range of [0, 1]. In a sample space S, the value of probability determines whether an outcome is true or false. The probability of an event Pr(A) that will never occur = 0. The probability of the event Pr(B) that will certainly occur = 1. This makes both events A and B thus a certainty. Furthermore, the sum of probabilities Pr(E1) + Pr(E2) + … + Pr(En) of a finite set of events in a given sample space S = 1. Conversely, the difference of the sum of two probabilities that will certainly occur is 0. Firstly, this paper discusses Bayes’ theorem, then complement of probability and the difference of probability for occurrences of learning-events, before applying these in the prediction of learning objects in student learning. Given the sum total of 1; to make recommendation for student learning, this paper submits that the difference of argMaxPr(S) and probability of student-performance quantifies the weight of learning objects for students. Using a dataset of skill-set, the computational procedure demonstrates: i) the probability of skill-set events that has occurred that would lead to higher level learning; ii) the probability of the events that has not occurred that requires subject-matter relearning; iii) accuracy of decision tree in the prediction of student performance into class labels; and iv) information entropy about skill-set data and its implication on student cognitive performance and recommendation of learning [1].
{"title":"Difference of Probability and Information Entropy for Skills Classification and Prediction in Student Learning","authors":"Kennedy E. Ehimwenma, Safiya Al Sharji, M. Raheem","doi":"10.5121/ijaia.2022.13501","DOIUrl":"https://doi.org/10.5121/ijaia.2022.13501","url":null,"abstract":"The probability of an event is in the range of [0, 1]. In a sample space S, the value of probability determines whether an outcome is true or false. The probability of an event Pr(A) that will never occur = 0. The probability of the event Pr(B) that will certainly occur = 1. This makes both events A and B thus a certainty. Furthermore, the sum of probabilities Pr(E1) + Pr(E2) + … + Pr(En) of a finite set of events in a given sample space S = 1. Conversely, the difference of the sum of two probabilities that will certainly occur is 0. Firstly, this paper discusses Bayes’ theorem, then complement of probability and the difference of probability for occurrences of learning-events, before applying these in the prediction of learning objects in student learning. Given the sum total of 1; to make recommendation for student learning, this paper submits that the difference of argMaxPr(S) and probability of student-performance quantifies the weight of learning objects for students. Using a dataset of skill-set, the computational procedure demonstrates: i) the probability of skill-set events that has occurred that would lead to higher level learning; ii) the probability of the events that has not occurred that requires subject-matter relearning; iii) accuracy of decision tree in the prediction of student performance into class labels; and iv) information entropy about skill-set data and its implication on student cognitive performance and recommendation of learning [1].","PeriodicalId":391502,"journal":{"name":"International Journal of Artificial Intelligence & Applications","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126956287","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 : 2022-07-31DOI: 10.5121/ijaia.2022.13401
P. Acharya, Tsuchin P. Chu, Khaled R. Ahmed, S. Kharel
Additive manufacturing is an emerging and crucial technology that can overcome the limitations of traditional manufacturing techniques to accurately manufacture highly complex parts. X-ray Computed Tomography (XCT) is a widely used method for non-destructive testing of AM parts. However, detection and segmentation of defects in XCT images of AM have many challenges due to contrast, size, and appearance of defects. This study developed deep learning techniques for detecting and segmenting defects in XCT images of AM. Due to a large number of required defect annotations, this paper applied image processing techniques to automate the defect labeling process. A single-stage object detection algorithm (YOLOv5) was applied to the problem of defect detection in image data. Three different variants of YOLOv5 were implemented and their performances were compared. U-Net was applied for defect segmentation in XCT slices. Finally, this research demonstrates that deep learning techniques can improve the automatic defect detection and segmentation in XCT data of AM.
{"title":"A Deep Learning Approach for Defect Detection and Segmentation in X-Ray Computed Tomography Slices of Additively Manufactured Components","authors":"P. Acharya, Tsuchin P. Chu, Khaled R. Ahmed, S. Kharel","doi":"10.5121/ijaia.2022.13401","DOIUrl":"https://doi.org/10.5121/ijaia.2022.13401","url":null,"abstract":"Additive manufacturing is an emerging and crucial technology that can overcome the limitations of traditional manufacturing techniques to accurately manufacture highly complex parts. X-ray Computed Tomography (XCT) is a widely used method for non-destructive testing of AM parts. However, detection and segmentation of defects in XCT images of AM have many challenges due to contrast, size, and appearance of defects. This study developed deep learning techniques for detecting and segmenting defects in XCT images of AM. Due to a large number of required defect annotations, this paper applied image processing techniques to automate the defect labeling process. A single-stage object detection algorithm (YOLOv5) was applied to the problem of defect detection in image data. Three different variants of YOLOv5 were implemented and their performances were compared. U-Net was applied for defect segmentation in XCT slices. Finally, this research demonstrates that deep learning techniques can improve the automatic defect detection and segmentation in XCT data of AM.","PeriodicalId":391502,"journal":{"name":"International Journal of Artificial Intelligence & Applications","volume":"294 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124221540","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 : 2022-07-31DOI: 10.5121/ijaia.2022.13406
Soroosh Parsai, M. Ahmadi
This paper represents a feature extraction method constructed on the local binary pattern (LBP) structure. The proposed method introduces a new adaptive thresholding function to the LBP method replacing the fixed thresholding at zero. The introduced function is a Gaussian Distribution Function (GDF) variation. The proposed technique uses the global and local information of the image and image blocks to perform the adaptation. The adaptive function adds to the on-hand im-age’s features by preserving the information of the amplitude of the pixel difference rather than just considering the sign of the pixel difference in the process of LBP coding. This feature improves the accuracy of the face recognition system by providing additional information. The proposed method demonstrates a higher recognition rate than other presented techniques (%97.75). The proposed method was also tested with different types of noise to demonstrate its effectiveness in the presence of various levels of noise. The Extended Yale B dataset was used for the testing along with Support Vector Machine (SVM) as classifier.
{"title":"New Local Binary Pattern Feature Extractor with Adaptive Threshold for Face Recognition Applications","authors":"Soroosh Parsai, M. Ahmadi","doi":"10.5121/ijaia.2022.13406","DOIUrl":"https://doi.org/10.5121/ijaia.2022.13406","url":null,"abstract":"This paper represents a feature extraction method constructed on the local binary pattern (LBP) structure. The proposed method introduces a new adaptive thresholding function to the LBP method replacing the fixed thresholding at zero. The introduced function is a Gaussian Distribution Function (GDF) variation. The proposed technique uses the global and local information of the image and image blocks to perform the adaptation. The adaptive function adds to the on-hand im-age’s features by preserving the information of the amplitude of the pixel difference rather than just considering the sign of the pixel difference in the process of LBP coding. This feature improves the accuracy of the face recognition system by providing additional information. The proposed method demonstrates a higher recognition rate than other presented techniques (%97.75). The proposed method was also tested with different types of noise to demonstrate its effectiveness in the presence of various levels of noise. The Extended Yale B dataset was used for the testing along with Support Vector Machine (SVM) as classifier.","PeriodicalId":391502,"journal":{"name":"International Journal of Artificial Intelligence & Applications","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130160357","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 : 2022-07-31DOI: 10.5121/ijaia.2022.13407
Zilvinas Svigaris
Quality risk management in industrial plants involves big calculations, the scale of which is often not only incomprehensible but also difficult to manage due to many parameters that affect the quality of production. Unsurprisingly, artificial intelligence-based quality management models are being introduced in manufacturing, only in niche, narrow areas, mostly for tracking product defects or identifying local quality defects. However, detecting the defect stage already is a late stage of the problem, which is almost always associated with a loss. Here comes the importance of prediction of problems or identifying of problematic patterns at an early stage before having production losses. Such attempts are rare and require a special approach. This type of module is needed for wide range problem forecasting in manufacturing. It should be configurable and clear not only by narrow area professionals, but also by medium-sized factory technologists who can configure such a system themselves to control their production quality risks. So here we are developing an approach whose strengths would be its simplicity, comprehensibility, fastness, and accessibility in its training, allowing us to understand why in one case or another the system predicts one decision or another.
{"title":"Management of Unplanned Changes in Production Processes: AI Control Systems","authors":"Zilvinas Svigaris","doi":"10.5121/ijaia.2022.13407","DOIUrl":"https://doi.org/10.5121/ijaia.2022.13407","url":null,"abstract":"Quality risk management in industrial plants involves big calculations, the scale of which is often not only incomprehensible but also difficult to manage due to many parameters that affect the quality of production. Unsurprisingly, artificial intelligence-based quality management models are being introduced in manufacturing, only in niche, narrow areas, mostly for tracking product defects or identifying local quality defects. However, detecting the defect stage already is a late stage of the problem, which is almost always associated with a loss. Here comes the importance of prediction of problems or identifying of problematic patterns at an early stage before having production losses. Such attempts are rare and require a special approach. This type of module is needed for wide range problem forecasting in manufacturing. It should be configurable and clear not only by narrow area professionals, but also by medium-sized factory technologists who can configure such a system themselves to control their production quality risks. So here we are developing an approach whose strengths would be its simplicity, comprehensibility, fastness, and accessibility in its training, allowing us to understand why in one case or another the system predicts one decision or another.","PeriodicalId":391502,"journal":{"name":"International Journal of Artificial Intelligence & Applications","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129790492","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 : 2022-07-31DOI: 10.5121/ijaia.2022.13403
Glenda Tan Hui En, K. Erhn, Shen Bingquan
More infectious virus variants can arise from rapid mutations in their proteins, creating new infection waves. These variants can evade one’s immune system and infect vaccinated individuals, lowering vaccine efficacy. Hence, to improve vaccine design, this project proposes Optimus PPIme – a deep learning approach to predict future, more infectious variants from an existing virus (exemplified by SARS-CoV-2). The approach comprises an algorithm which acts as a “virus” attacking a host cell. To increase infectivity, the “virus” mutates to bind better to the host’s receptor. 2 algorithms were attempted – greedy search and beam search. The strength of this variant-host binding was then assessed by a transformer network we developed, with a high accuracy of 90%. With both components, beam search eventually proposed more infectious variants. Therefore, this approach can potentially enable researchers to develop vaccines that provide protection against future infectious variants before they emerge, pre-empting outbreaks and saving lives.
{"title":"Predicting more Infectious Virus Variants for Pandemic Prevention through Deep Learning","authors":"Glenda Tan Hui En, K. Erhn, Shen Bingquan","doi":"10.5121/ijaia.2022.13403","DOIUrl":"https://doi.org/10.5121/ijaia.2022.13403","url":null,"abstract":"More infectious virus variants can arise from rapid mutations in their proteins, creating new infection waves. These variants can evade one’s immune system and infect vaccinated individuals, lowering vaccine efficacy. Hence, to improve vaccine design, this project proposes Optimus PPIme – a deep learning approach to predict future, more infectious variants from an existing virus (exemplified by SARS-CoV-2). The approach comprises an algorithm which acts as a “virus” attacking a host cell. To increase infectivity, the “virus” mutates to bind better to the host’s receptor. 2 algorithms were attempted – greedy search and beam search. The strength of this variant-host binding was then assessed by a transformer network we developed, with a high accuracy of 90%. With both components, beam search eventually proposed more infectious variants. Therefore, this approach can potentially enable researchers to develop vaccines that provide protection against future infectious variants before they emerge, pre-empting outbreaks and saving lives.","PeriodicalId":391502,"journal":{"name":"International Journal of Artificial Intelligence & Applications","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127775079","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：