Aye Aye Mar, Kiyoaki Shirai, Natthawut Kertkeidkachorn
Aspect-based sentiment analysis (ABSA) is a process to extract an aspect of a product from a customer review and identify its polarity. Most previous studies of ABSA focused on explicit aspects, but implicit aspects have not yet been the subject of much attention. This paper proposes a novel weakly supervised method for implicit aspect extraction, which is a task to classify a sentence into a pre-defined implicit aspect category. A dataset labeled with implicit aspects is automatically constructed from unlabeled sentences as follows. First, explicit sentences are obtained by extracting explicit aspects from unlabeled sentences, while sentences that do not contain explicit aspects are preserved as candidates of implicit sentences. Second, clustering is performed to merge the explicit and implicit sentences that share the same aspect. Third, the aspect of the explicit sentence is assigned to the implicit sentences in the same cluster as the implicit aspect label. Then, the BERT model is fine-tuned for implicit aspect extraction using the constructed dataset. The results of the experiments show that our method achieves 82% and 84% accuracy for mobile phone and PC reviews, respectively, which are 20 and 21 percentage points higher than the baseline.
{"title":"Weakly Supervised Learning Approach for Implicit Aspect Extraction","authors":"Aye Aye Mar, Kiyoaki Shirai, Natthawut Kertkeidkachorn","doi":"10.3390/info14110612","DOIUrl":"https://doi.org/10.3390/info14110612","url":null,"abstract":"Aspect-based sentiment analysis (ABSA) is a process to extract an aspect of a product from a customer review and identify its polarity. Most previous studies of ABSA focused on explicit aspects, but implicit aspects have not yet been the subject of much attention. This paper proposes a novel weakly supervised method for implicit aspect extraction, which is a task to classify a sentence into a pre-defined implicit aspect category. A dataset labeled with implicit aspects is automatically constructed from unlabeled sentences as follows. First, explicit sentences are obtained by extracting explicit aspects from unlabeled sentences, while sentences that do not contain explicit aspects are preserved as candidates of implicit sentences. Second, clustering is performed to merge the explicit and implicit sentences that share the same aspect. Third, the aspect of the explicit sentence is assigned to the implicit sentences in the same cluster as the implicit aspect label. Then, the BERT model is fine-tuned for implicit aspect extraction using the constructed dataset. The results of the experiments show that our method achieves 82% and 84% accuracy for mobile phone and PC reviews, respectively, which are 20 and 21 percentage points higher than the baseline.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"58 3","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136346763","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}
In the prediction of time series, Empirical Mode Decomposition (EMD) generates subsequences and separates short-term tendencies from long-term ones. However, a single prediction model, including attention mechanism, has varying effects on each subsequence. To accurately capture the regularities of subsequences using an attention mechanism, we propose an integrated model for time series prediction based on signal decomposition and two attention mechanisms. This model combines the results of three networks—LSTM, LSTM-self-attention, and LSTM-temporal attention—all trained using subsequences obtained from EMD. Additionally, since previous research on EMD has been limited to single series analysis, this paper includes multiple series by employing two data pre-processing methods: ‘overall normalization’ and ‘respective normalization’. Experimental results on various datasets demonstrate that compared to models without attention mechanisms, temporal attention improves the prediction accuracy of short- and medium-term decomposed series by 15~28% and 45~72%, respectively; furthermore, it reduces the overall prediction error by 10~17%. The integrated model with temporal attention achieves a reduction in error of approximately 0.3%, primarily when compared to models utilizing only general forms of attention mechanisms. Moreover, after normalizing multiple series separately, the predictive performance is equivalent to that achieved for individual series.
{"title":"An Integrated Time Series Prediction Model Based on Empirical Mode Decomposition and Two Attention Mechanisms","authors":"Xianchang Wang, Siyu Dong, Rui Zhang","doi":"10.3390/info14110610","DOIUrl":"https://doi.org/10.3390/info14110610","url":null,"abstract":"In the prediction of time series, Empirical Mode Decomposition (EMD) generates subsequences and separates short-term tendencies from long-term ones. However, a single prediction model, including attention mechanism, has varying effects on each subsequence. To accurately capture the regularities of subsequences using an attention mechanism, we propose an integrated model for time series prediction based on signal decomposition and two attention mechanisms. This model combines the results of three networks—LSTM, LSTM-self-attention, and LSTM-temporal attention—all trained using subsequences obtained from EMD. Additionally, since previous research on EMD has been limited to single series analysis, this paper includes multiple series by employing two data pre-processing methods: ‘overall normalization’ and ‘respective normalization’. Experimental results on various datasets demonstrate that compared to models without attention mechanisms, temporal attention improves the prediction accuracy of short- and medium-term decomposed series by 15~28% and 45~72%, respectively; furthermore, it reduces the overall prediction error by 10~17%. The integrated model with temporal attention achieves a reduction in error of approximately 0.3%, primarily when compared to models utilizing only general forms of attention mechanisms. Moreover, after normalizing multiple series separately, the predictive performance is equivalent to that achieved for individual series.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"24 18","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135087043","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}
Weiting Ding, Jialu Li, Heyang Ma, Yeru Wu, Hailong He
As a powerful statistical method, meta-analysis has been applied increasingly in agricultural science with remarkable progress. However, meta-analysis research reports in the agricultural discipline still need to be systematically combed. Scientometrics is often used to quantitatively analyze research on certain themes. In this study, the literature from a 30-year period (1992–2021) was retrieved based on the Web of Science database, and a quantitative analysis was performed using the VOSviewer and CiteSpace visual analysis software packages. The objective of this study was to investigate the current application of meta-analysis in agricultural sciences, the latest research hotspots, and trends, and to identify influential authors, research institutions, countries, articles, and journal sources. Over the past 30 years, the volume of the meta-analysis literature in agriculture has increased rapidly. We identified the top three authors (Sauvant D, Kebreab E, and Huhtanen P), the top three contributing organizations (Chinese Academy of Sciences, National Institute for Agricultural Research, and Northwest A&F University), and top three productive countries (the USA, China, and France). Keyword cluster analysis shows that the meta-analysis research in agricultural sciences falls into four categories: climate change, crop yield, soil, and animal husbandry. Jeffrey (2011) is the most influential and cited research paper, with the highest utilization rate for the Journal of Dairy Science. This paper objectively evaluates the development of meta-analysis in the agricultural sciences using bibliometrics analysis, grasps the development frontier of agricultural research, and provides insights into the future of related research in the agricultural sciences.
元分析作为一种强大的统计方法，在农业科学中的应用日益广泛，取得了显著进展。然而，农业学科的meta分析研究报告仍需系统梳理。科学计量学通常用于定量分析某些主题的研究。本研究基于Web of Science数据库检索近30年(1992-2021)的文献，利用VOSviewer和CiteSpace可视化分析软件包进行定量分析。本研究的目的是调查meta分析在农业科学中的应用现状、最新研究热点和趋势，并确定有影响力的作者、研究机构、国家、文章和期刊来源。在过去的30年里，农业荟萃分析文献的数量迅速增加。我们确定了前三位作者(Sauvant D, Kebreab E和Huhtanen P)，前三位贡献机构(中国科学院，国家农业研究所和西北农林科技大学)和前三位生产国家(美国，中国和法国)。关键词聚类分析表明，农业科学的元分析研究可分为气候变化、作物产量、土壤和畜牧业四类。Jeffrey(2011)是《Journal of Dairy Science》最具影响力和被引率最高的研究论文。本文运用文献计量学分析客观评价农业科学元分析的发展，把握农业研究的发展前沿，展望农业科学相关研究的未来。
{"title":"Science Mapping of Meta-Analysis in Agricultural Science","authors":"Weiting Ding, Jialu Li, Heyang Ma, Yeru Wu, Hailong He","doi":"10.3390/info14110611","DOIUrl":"https://doi.org/10.3390/info14110611","url":null,"abstract":"As a powerful statistical method, meta-analysis has been applied increasingly in agricultural science with remarkable progress. However, meta-analysis research reports in the agricultural discipline still need to be systematically combed. Scientometrics is often used to quantitatively analyze research on certain themes. In this study, the literature from a 30-year period (1992–2021) was retrieved based on the Web of Science database, and a quantitative analysis was performed using the VOSviewer and CiteSpace visual analysis software packages. The objective of this study was to investigate the current application of meta-analysis in agricultural sciences, the latest research hotspots, and trends, and to identify influential authors, research institutions, countries, articles, and journal sources. Over the past 30 years, the volume of the meta-analysis literature in agriculture has increased rapidly. We identified the top three authors (Sauvant D, Kebreab E, and Huhtanen P), the top three contributing organizations (Chinese Academy of Sciences, National Institute for Agricultural Research, and Northwest A&F University), and top three productive countries (the USA, China, and France). Keyword cluster analysis shows that the meta-analysis research in agricultural sciences falls into four categories: climate change, crop yield, soil, and animal husbandry. Jeffrey (2011) is the most influential and cited research paper, with the highest utilization rate for the Journal of Dairy Science. This paper objectively evaluates the development of meta-analysis in the agricultural sciences using bibliometrics analysis, grasps the development frontier of agricultural research, and provides insights into the future of related research in the agricultural sciences.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"39 11","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135086865","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}
Politically polarizing issues are a growing concern around the world, creating divisions along ideological lines, which was also confirmed during the 2022 United States midterm elections. The purpose of this study was to explore the relationship between the results of the 2022 U.S. midterm elections and the topics that were covered during the campaign. A dataset consisting of 52,688 tweets in total was created by collecting tweets of senators, representatives and governors who participated in the elections one month before the start of the elections. Using unsupervised machine learning, topic modeling is built on the collected data and visualized to represent topics. Furthermore, supervised machine learning is used to classify tweets to the corresponding political party, whereas sentiment analysis is carried out in order to detect polarity and subjectivity. Tweets from participating politicians, U.S. states and involved parties were found to correlate with polarizing topics. This study hereby explored the relationship between the topics that were creating a divide between Democrats and Republicans during their campaign and the 2022 U.S. midterm election outcomes. This research found that polarizing topics permeated the Twitter (today known as X) campaign, and that all elections were classified as highly subjective. In the Senate and House elections, this classification analysis showed significant misclassification rates of 21.37% and 24.15%, respectively, indicating that Republican tweets often aligned with traditional Democratic narratives.
{"title":"Polarizing Topics on Twitter in the 2022 United States Elections","authors":"Josip Katalinić, Ivan Dunđer, Sanja Seljan","doi":"10.3390/info14110609","DOIUrl":"https://doi.org/10.3390/info14110609","url":null,"abstract":"Politically polarizing issues are a growing concern around the world, creating divisions along ideological lines, which was also confirmed during the 2022 United States midterm elections. The purpose of this study was to explore the relationship between the results of the 2022 U.S. midterm elections and the topics that were covered during the campaign. A dataset consisting of 52,688 tweets in total was created by collecting tweets of senators, representatives and governors who participated in the elections one month before the start of the elections. Using unsupervised machine learning, topic modeling is built on the collected data and visualized to represent topics. Furthermore, supervised machine learning is used to classify tweets to the corresponding political party, whereas sentiment analysis is carried out in order to detect polarity and subjectivity. Tweets from participating politicians, U.S. states and involved parties were found to correlate with polarizing topics. This study hereby explored the relationship between the topics that were creating a divide between Democrats and Republicans during their campaign and the 2022 U.S. midterm election outcomes. This research found that polarizing topics permeated the Twitter (today known as X) campaign, and that all elections were classified as highly subjective. In the Senate and House elections, this classification analysis showed significant misclassification rates of 21.37% and 24.15%, respectively, indicating that Republican tweets often aligned with traditional Democratic narratives.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":" 3","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135191850","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}
This study proposes a framework for a systems engineering-based approach to context-aware personalization, which is applied to e-commerce through the understanding and modeling of user behavior from their interactions with sales channels and media. The framework is practical and built on systems engineering principles. It combines three conceptual components to produce signals that provide content relevant to the users based on their behavior, thus enhancing their experience. These components are the ‘recognition and knowledge’ of the users and their behavior (persona); the awareness of users’ current contexts; and the comprehension of their situation and projection of their future status (intent prediction). The persona generator is implemented by leveraging an unsupervised machine learning algorithm to assign users into cohorts and learn cohort behavior while preserving their privacy in an ethical framework. The component of the users’ current context is fulfilled as a microservice that adopts novel e-commerce data interpretations. The best result of 97.3% accuracy for the intent prediction component was obtained by tokenizing categorical features with a pre-trained BERT (bidirectional encoder representations from transformers) model and passing these, as the contextual embedding input, to an LSTM (long short-term memory) neural network. Paired cohort-directed prescriptive action is generated from learned behavior as a recommended alternative to users’ shopping steps. The practical implementation of this e-commerce personalization framework is demonstrated in this study through the empirical evaluation of experimental results.
{"title":"Context-Aware Personalization: A Systems Engineering Framework","authors":"Olurotimi Oguntola, Steven Simske","doi":"10.3390/info14110608","DOIUrl":"https://doi.org/10.3390/info14110608","url":null,"abstract":"This study proposes a framework for a systems engineering-based approach to context-aware personalization, which is applied to e-commerce through the understanding and modeling of user behavior from their interactions with sales channels and media. The framework is practical and built on systems engineering principles. It combines three conceptual components to produce signals that provide content relevant to the users based on their behavior, thus enhancing their experience. These components are the ‘recognition and knowledge’ of the users and their behavior (persona); the awareness of users’ current contexts; and the comprehension of their situation and projection of their future status (intent prediction). The persona generator is implemented by leveraging an unsupervised machine learning algorithm to assign users into cohorts and learn cohort behavior while preserving their privacy in an ethical framework. The component of the users’ current context is fulfilled as a microservice that adopts novel e-commerce data interpretations. The best result of 97.3% accuracy for the intent prediction component was obtained by tokenizing categorical features with a pre-trained BERT (bidirectional encoder representations from transformers) model and passing these, as the contextual embedding input, to an LSTM (long short-term memory) neural network. Paired cohort-directed prescriptive action is generated from learned behavior as a recommended alternative to users’ shopping steps. The practical implementation of this e-commerce personalization framework is demonstrated in this study through the empirical evaluation of experimental results.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":" 9","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135190888","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}
The article explores the implementation of digital technology in small and medium-sized Slovenian enterprises (SMEs), with a focus on understanding existing trends, obstacles, and necessary support measures during their digitalization progress. The surveyed companies mainly rely on conventional technologies like websites and teamwork platforms, emphasizing the significance of strong online communication and presence in the modern business world. The adoption of advanced technologies such as blockchain is limited due to the perceived complexity and relevance to specific sectors. This study uses variance analysis to identify potential differences in the digitalization challenges faced by companies of different sizes. The results indicate that small companies face different financial constraints and require more differentiated support mechanisms than their larger counterparts, with a particular focus on improving digital competencies among employees. Despite obtaining enhancements such as elevated operational standards and uninterrupted telecommuting via digitalization, companies still face challenges of differentiation and organizational culture change. The study emphasizes the importance of recognizing and addressing the different challenges and support needs of different-sized companies to promote comprehensive progress in digital transformation. Our findings provide important insights for policymakers, industry stakeholders, and SMEs to formulate comprehensive strategies and policies that effectively address the diverse needs and challenges of the digital transformation landscape.
{"title":"Small and Medium-Sized Enterprises in the Digital Age: Understanding Characteristics and Essential Demands","authors":"Barbara Bradač Hojnik, Ivona Huđek","doi":"10.3390/info14110606","DOIUrl":"https://doi.org/10.3390/info14110606","url":null,"abstract":"The article explores the implementation of digital technology in small and medium-sized Slovenian enterprises (SMEs), with a focus on understanding existing trends, obstacles, and necessary support measures during their digitalization progress. The surveyed companies mainly rely on conventional technologies like websites and teamwork platforms, emphasizing the significance of strong online communication and presence in the modern business world. The adoption of advanced technologies such as blockchain is limited due to the perceived complexity and relevance to specific sectors. This study uses variance analysis to identify potential differences in the digitalization challenges faced by companies of different sizes. The results indicate that small companies face different financial constraints and require more differentiated support mechanisms than their larger counterparts, with a particular focus on improving digital competencies among employees. Despite obtaining enhancements such as elevated operational standards and uninterrupted telecommuting via digitalization, companies still face challenges of differentiation and organizational culture change. The study emphasizes the importance of recognizing and addressing the different challenges and support needs of different-sized companies to promote comprehensive progress in digital transformation. Our findings provide important insights for policymakers, industry stakeholders, and SMEs to formulate comprehensive strategies and policies that effectively address the diverse needs and challenges of the digital transformation landscape.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":" 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135286570","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}
Alexandros Z. Spyropoulos, Charalampos Bratsas, Georgios C. Makris, Emmanouel Garoufallou, Vassilis Tsiantos
Nowadays, more and more sciences are involved in strengthening the work of law enforcement authorities. Scientific documentation is evidence highly respected by the courts in administering justice. As the involvement of science in solving crimes increases, so does human subjectivism, which often leads to wrong conclusions and, consequently, to bad judgments. From the above arises the need to create a single information system that will be fed with scientific evidence such as fingerprints, genetic material, digital data, forensic photographs, information from the forensic report, etc., and also investigative data such as information from witnesses’ statements, the apology of the accused, etc., from various crime scenes that will be able, through formal reasoning procedure, to conclude possible perpetrators. The present study examines a proposal for developing an information system that can be a basis for creating a forensic ontology—a semantic representation of the crime scene—through descriptive logic in the owl semantic language. The Interoperability-Enhanced information system to be developed could assist law enforcement authorities in solving crimes. At the same time, it would promote closer cooperation between academia, civil society, and state institutions by fostering a culture of engagement for the common good.
{"title":"Interoperability-Enhanced Knowledge Management in Law Enforcement: An Integrated Data-Driven Forensic Ontological Approach to Crime Scene Analysis","authors":"Alexandros Z. Spyropoulos, Charalampos Bratsas, Georgios C. Makris, Emmanouel Garoufallou, Vassilis Tsiantos","doi":"10.3390/info14110607","DOIUrl":"https://doi.org/10.3390/info14110607","url":null,"abstract":"Nowadays, more and more sciences are involved in strengthening the work of law enforcement authorities. Scientific documentation is evidence highly respected by the courts in administering justice. As the involvement of science in solving crimes increases, so does human subjectivism, which often leads to wrong conclusions and, consequently, to bad judgments. From the above arises the need to create a single information system that will be fed with scientific evidence such as fingerprints, genetic material, digital data, forensic photographs, information from the forensic report, etc., and also investigative data such as information from witnesses’ statements, the apology of the accused, etc., from various crime scenes that will be able, through formal reasoning procedure, to conclude possible perpetrators. The present study examines a proposal for developing an information system that can be a basis for creating a forensic ontology—a semantic representation of the crime scene—through descriptive logic in the owl semantic language. The Interoperability-Enhanced information system to be developed could assist law enforcement authorities in solving crimes. At the same time, it would promote closer cooperation between academia, civil society, and state institutions by fostering a culture of engagement for the common good.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":" 10","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135286573","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}
Cybersecurity in modern vehicles has received increased attention from the research community in recent years. Intrusion Detection Systems (IDSs) are one of the techniques used to detect and mitigate cybersecurity risks. This paper proposes a novel implementation of an IDS for in-vehicle security networks based on the concept of multi-scale histograms, which capture the frequencies of message identifiers in CAN-bus in-vehicle networks. In comparison to existing approaches in the literature based on a single histogram, the proposed approach widens the informative context used by the IDS for traffic analysis by taking into consideration sequences of two and three CAN-bus messages to create multi-scale dictionaries. The histograms are created from windows of in-vehicle network traffic. A preliminary multi-scale histogram model is created using only legitimate traffic. Against this model, the IDS performs traffic analysis to create a feature space based on the correlation of the histograms. Then, the created feature space is given in input to a Convolutional Neural Network (CNN) for the identification of the windows of traffic where the attack is present. The proposed approach has been evaluated on two different public data sets achieving a very competitive performance in comparison to the literature.
{"title":"In-Vehicle Network Intrusion Detection System Using Convolutional Neural Network and Multi-Scale Histograms","authors":"Gianmarco Baldini","doi":"10.3390/info14110605","DOIUrl":"https://doi.org/10.3390/info14110605","url":null,"abstract":"Cybersecurity in modern vehicles has received increased attention from the research community in recent years. Intrusion Detection Systems (IDSs) are one of the techniques used to detect and mitigate cybersecurity risks. This paper proposes a novel implementation of an IDS for in-vehicle security networks based on the concept of multi-scale histograms, which capture the frequencies of message identifiers in CAN-bus in-vehicle networks. In comparison to existing approaches in the literature based on a single histogram, the proposed approach widens the informative context used by the IDS for traffic analysis by taking into consideration sequences of two and three CAN-bus messages to create multi-scale dictionaries. The histograms are created from windows of in-vehicle network traffic. A preliminary multi-scale histogram model is created using only legitimate traffic. Against this model, the IDS performs traffic analysis to create a feature space based on the correlation of the histograms. Then, the created feature space is given in input to a Convolutional Neural Network (CNN) for the identification of the windows of traffic where the attack is present. The proposed approach has been evaluated on two different public data sets achieving a very competitive performance in comparison to the literature.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"326 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135392609","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}
Artificial intelligence is changing and influencing our world. As one of the main algorithms in the field of artificial intelligence, convolutional neural networks (CNNs) have developed rapidly in recent years. Especially after the emergence of NASNet, CNNs have gradually pushed the idea of AutoML to the public’s attention, and large numbers of new structures designed by automatic searches are appearing. These networks are usually based on reinforcement learning and evolutionary learning algorithms. However, sometimes, the blocks of these networks are complex, and there is no small model for simpler tasks. Therefore, this paper proposes POSS-CNN aiming at target recognition and detection, which employs a multi-branch CNN structure with PSNC and a method of automatic parallel selection for super parameters based on a multi-branch CNN structure. Moreover, POSS-CNN can be broken up. By choosing a single branch or the combination of two branches as the “benchmark”, as well as the overall POSS-CNN, we can achieve seven models with different precision and operations. The test accuracy of POSS-CNN for a recognition task tested on a CIFAR10 dataset can reach 86.4%, which is equivalent to AlexNet and VggNet, but the operation and parameters of the whole model in this paper are 45.9% and 45.8% of AlexNet, and 29.5% and 29.4% of VggNet. The mAP of POSS-CNN for a detection task tested on the LSVH dataset is 45.8, inferior to the 62.3 of YOLOv3. However, compared with YOLOv3, the operation and parameters of the model in this paper are reduced by 57.4% and 15.6%, respectively. After being accelerated by WRA, POSS-CNN for a detection task tested on an LSVH dataset can achieve 27 fps, and the energy efficiency is 0.42 J/f, which is 5 times and 96.6 times better than GPU 2080Ti in performance and energy efficiency, respectively.
{"title":"POSS-CNN: An Automatically Generated Convolutional Neural Network with Precision and Operation Separable Structure Aiming at Target Recognition and Detection","authors":"Jia Hou, Jingyu Zhang, Qi Chen, Siwei Xiang, Yishuo Meng, Jianfei Wang, Cimang Lu, Chen Yang","doi":"10.3390/info14110604","DOIUrl":"https://doi.org/10.3390/info14110604","url":null,"abstract":"Artificial intelligence is changing and influencing our world. As one of the main algorithms in the field of artificial intelligence, convolutional neural networks (CNNs) have developed rapidly in recent years. Especially after the emergence of NASNet, CNNs have gradually pushed the idea of AutoML to the public’s attention, and large numbers of new structures designed by automatic searches are appearing. These networks are usually based on reinforcement learning and evolutionary learning algorithms. However, sometimes, the blocks of these networks are complex, and there is no small model for simpler tasks. Therefore, this paper proposes POSS-CNN aiming at target recognition and detection, which employs a multi-branch CNN structure with PSNC and a method of automatic parallel selection for super parameters based on a multi-branch CNN structure. Moreover, POSS-CNN can be broken up. By choosing a single branch or the combination of two branches as the “benchmark”, as well as the overall POSS-CNN, we can achieve seven models with different precision and operations. The test accuracy of POSS-CNN for a recognition task tested on a CIFAR10 dataset can reach 86.4%, which is equivalent to AlexNet and VggNet, but the operation and parameters of the whole model in this paper are 45.9% and 45.8% of AlexNet, and 29.5% and 29.4% of VggNet. The mAP of POSS-CNN for a detection task tested on the LSVH dataset is 45.8, inferior to the 62.3 of YOLOv3. However, compared with YOLOv3, the operation and parameters of the model in this paper are reduced by 57.4% and 15.6%, respectively. After being accelerated by WRA, POSS-CNN for a detection task tested on an LSVH dataset can achieve 27 fps, and the energy efficiency is 0.42 J/f, which is 5 times and 96.6 times better than GPU 2080Ti in performance and energy efficiency, respectively.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"174 6","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135480459","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}
Outsourcing computation has become increasingly popular due to its cost-effectiveness, enabling users with limited resources to conduct large-scale computations on potentially untrusted cloud platforms. In order to safeguard privacy, verifiable computing (VC) has emerged as a secure approach, ensuring that the cloud cannot discern users’ input and output. Random permutation masking (RPM) is a widely adopted technique in VC protocols to provide robust privacy protection. This work presents a precise definition of the privacy-preserving property of RPM by employing indistinguishability experiments. Moreover, an innovative attack exploiting the greatest common divisor and the least common multiple of each row and column in the encrypted matrices is introduced against RPM. Unlike previous density-based attacks, this novel approach offers a significant advantage by allowing the reconstruction of matrix values from the ciphertext based on RPM. A comprehensive demonstration was provided to illustrate the failure of protocols based on RPM in maintaining the privacy-preserving property under this proposed attack. Furthermore, an extensive series of experiments is conducted to thoroughly validate the effectiveness and advantages of the attack against RPM. The findings of this research highlight vulnerabilities in RPM-based VC protocols and underline the pressing need for further enhancements and alternative privacy-preserving mechanisms in outsourcing computation.
{"title":"Enhancing Privacy Preservation in Verifiable Computation through Random Permutation Masking to Prevent Leakage","authors":"Yang Yang, Guanghua Song","doi":"10.3390/info14110603","DOIUrl":"https://doi.org/10.3390/info14110603","url":null,"abstract":"Outsourcing computation has become increasingly popular due to its cost-effectiveness, enabling users with limited resources to conduct large-scale computations on potentially untrusted cloud platforms. In order to safeguard privacy, verifiable computing (VC) has emerged as a secure approach, ensuring that the cloud cannot discern users’ input and output. Random permutation masking (RPM) is a widely adopted technique in VC protocols to provide robust privacy protection. This work presents a precise definition of the privacy-preserving property of RPM by employing indistinguishability experiments. Moreover, an innovative attack exploiting the greatest common divisor and the least common multiple of each row and column in the encrypted matrices is introduced against RPM. Unlike previous density-based attacks, this novel approach offers a significant advantage by allowing the reconstruction of matrix values from the ciphertext based on RPM. A comprehensive demonstration was provided to illustrate the failure of protocols based on RPM in maintaining the privacy-preserving property under this proposed attack. Furthermore, an extensive series of experiments is conducted to thoroughly validate the effectiveness and advantages of the attack against RPM. The findings of this research highlight vulnerabilities in RPM-based VC protocols and underline the pressing need for further enhancements and alternative privacy-preserving mechanisms in outsourcing computation.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"16 12","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135589137","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：