The German labor market relies heavily on vocational training, retraining, and continuing education. In order to match training seekers with training offers and to make the available data interoperable, we present a novel approach to automatically detect access to education and training in German training offers and advertisements and identify open research questions and areas for further research. In particular, we focus on (a) general education and school leaving certificates, (b) work experience, (c) previous apprenticeship, and (d) a list of skills provided by the German Federal Employment Agency. This novel approach combines several methods: First, we provide technical terms and classes of the education system that are used synonymously, combining different qualifications and adding obsolete terms. Second, we provide rule-based matching to identify the need for work experience or education. However, not all qualification requirements can be matched due to incompatible data schemas or non-standardized requirements such as initial tests or interviews. Although there are several shortcomings, the presented approach shows promising results for two data sets: training and retraining advertisements.
{"title":"Challenges of Automated Identification of Access to Education and Training in Germany","authors":"Jens Dörpinghaus, David Samray, Robert Helmrich","doi":"10.3390/info14100524","DOIUrl":"https://doi.org/10.3390/info14100524","url":null,"abstract":"The German labor market relies heavily on vocational training, retraining, and continuing education. In order to match training seekers with training offers and to make the available data interoperable, we present a novel approach to automatically detect access to education and training in German training offers and advertisements and identify open research questions and areas for further research. In particular, we focus on (a) general education and school leaving certificates, (b) work experience, (c) previous apprenticeship, and (d) a list of skills provided by the German Federal Employment Agency. This novel approach combines several methods: First, we provide technical terms and classes of the education system that are used synonymously, combining different qualifications and adding obsolete terms. Second, we provide rule-based matching to identify the need for work experience or education. However, not all qualification requirements can be matched due to incompatible data schemas or non-standardized requirements such as initial tests or interviews. Although there are several shortcomings, the presented approach shows promising results for two data sets: training and retraining advertisements.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135719610","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}
Bilal Naji Alhasnawi, Basil H. Jasim, Ali M. Jasim, Vladimír Bureš, Arshad Naji Alhasnawi, Raad Z. Homod, Majid Razaq Mohamed Alsemawai, Rabeh Abbassi, Bishoy E. Sedhom
The electrical demand and generation in power systems is currently the biggest source of uncertainty for an electricity provider. For a dependable and financially advantageous electricity system, demand response (DR) success as a result of household appliance energy management has attracted significant attention. Due to fluctuating electricity rates and usage trends, determining the best schedule for apartment appliances can be difficult. As a result of this context, the Improved Cockroach Swarm Optimization Algorithm (ICSOA) is combined with the Innovative Apartments Appliance Scheduling (IAAS) framework. Using the proposed technique, the cost of electricity reduction, user comfort maximization, and peak-to-average ratio reduction are analyzed for apartment appliances. The proposed framework is evaluated by comparing it with BFOA and W/O scheduling cases. In comparison to the W/O scheduling case, the BFOA method lowered energy costs by 17.75%, but the ICSA approach reduced energy cost by 46.085%. According to the results, the created ICSA algorithm performed better than the BFOA and W/O scheduling situations in terms of the stated objectives and was advantageous to both utilities and consumers.
{"title":"A Multi-Objective Improved Cockroach Swarm Algorithm Approach for Apartment Energy Management Systems","authors":"Bilal Naji Alhasnawi, Basil H. Jasim, Ali M. Jasim, Vladimír Bureš, Arshad Naji Alhasnawi, Raad Z. Homod, Majid Razaq Mohamed Alsemawai, Rabeh Abbassi, Bishoy E. Sedhom","doi":"10.3390/info14100521","DOIUrl":"https://doi.org/10.3390/info14100521","url":null,"abstract":"The electrical demand and generation in power systems is currently the biggest source of uncertainty for an electricity provider. For a dependable and financially advantageous electricity system, demand response (DR) success as a result of household appliance energy management has attracted significant attention. Due to fluctuating electricity rates and usage trends, determining the best schedule for apartment appliances can be difficult. As a result of this context, the Improved Cockroach Swarm Optimization Algorithm (ICSOA) is combined with the Innovative Apartments Appliance Scheduling (IAAS) framework. Using the proposed technique, the cost of electricity reduction, user comfort maximization, and peak-to-average ratio reduction are analyzed for apartment appliances. The proposed framework is evaluated by comparing it with BFOA and W/O scheduling cases. In comparison to the W/O scheduling case, the BFOA method lowered energy costs by 17.75%, but the ICSA approach reduced energy cost by 46.085%. According to the results, the created ICSA algorithm performed better than the BFOA and W/O scheduling situations in terms of the stated objectives and was advantageous to both utilities and consumers.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135864711","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}
Gergely Márk Csányi, Renátó Vági, Andrea Megyeri, Anna Fülöp , Dániel Nagy, János Pál Vadász, István Üveges
Few-shot learning is a deep learning subfield that is the focus of research nowadays. This paper addresses the research question of whether a triplet-trained Siamese network, initially designed for multi-class classification, can effectively handle multi-label classification. We conducted a case study to identify any limitations in its application. The experiments were conducted on a dataset containing Hungarian legal decisions of administrative agencies in tax matters belonging to a major legal content provider. We also tested how different Siamese embeddings compare on classifying a previously non-existing label on a binary and a multi-label setting. We found that triplet-trained Siamese networks can be applied to perform classification but with a sampling restriction during training. We also found that the overlap between labels affects the results negatively. The few-shot model, seeing only ten examples for each label, provided competitive results compared to models trained on tens of thousands of court decisions using tf-idf vectorization and logistic regression.
{"title":"Can Triplet Loss Be Used for Multi-Label Few-Shot Classification? A Case Study","authors":"Gergely Márk Csányi, Renátó Vági, Andrea Megyeri, Anna Fülöp , Dániel Nagy, János Pál Vadász, István Üveges","doi":"10.3390/info14100520","DOIUrl":"https://doi.org/10.3390/info14100520","url":null,"abstract":"Few-shot learning is a deep learning subfield that is the focus of research nowadays. This paper addresses the research question of whether a triplet-trained Siamese network, initially designed for multi-class classification, can effectively handle multi-label classification. We conducted a case study to identify any limitations in its application. The experiments were conducted on a dataset containing Hungarian legal decisions of administrative agencies in tax matters belonging to a major legal content provider. We also tested how different Siamese embeddings compare on classifying a previously non-existing label on a binary and a multi-label setting. We found that triplet-trained Siamese networks can be applied to perform classification but with a sampling restriction during training. We also found that the overlap between labels affects the results negatively. The few-shot model, seeing only ten examples for each label, provided competitive results compared to models trained on tens of thousands of court decisions using tf-idf vectorization and logistic regression.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135966563","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}
Learnability in Automated Driving (LiAD) is a neglected research topic, especially when considering the unpredictable and intricate ways humans learn to interact and use automated driving systems (ADS) over the sequence of time. Moreover, there is a scarcity of publications dedicated to LiAD (specifically extended learnability methods) to guide the scientific paradigm. As a result, this generates scientific discord and, thus, leaves many facets of long-term learning effects associated with automated driving in dire need of significant research courtesy. This, we believe, is a constraint to knowledge discovery on quality interaction design phenomena. In a sense, it is imperative to abstract knowledge on how long-term effects and learning effects may affect (negatively and positively) users’ learning and mental models. As well as induce changeable behavioural configurations and performances. In view of that, it may be imperative to examine operational concepts that may help researchers envision future scenarios with automation by assessing users’ learning ability, how they learn and what they learn over the sequence of time. As well as constructing a theory of effects (from micro, meso and macro perspectives), which may help profile ergonomic quality design aspects that stand the test of time. As a result, we reviewed the literature on learnability, which we mined for LiAD knowledge discovery from the experience perspective of long-term learning effects. Therefore, the paper offers the reader the resulting discussion points formulated under the Learnability Engineering Life Cycle. For instance, firstly, contextualisation of LiAD with emphasis on extended LiAD. Secondly, conceptualisation and operationalisation of the operational mechanics of LiAD as a concept in ergonomic quality engineering (with an introduction of Concepts for Applying Learnability Engineering (CALE) research based on LiAD knowledge discovery). Thirdly, the systemisation of implementable long-term research strategies towards comprehending behaviour modification associated with extended LiAD. As the vehicle industry revolutionises at a rapid pace towards automation and artificially intelligent (AI) systems, this knowledge is useful for illuminating and instructing quality interaction strategies and Quality Automated Driving (QAD).
{"title":"Learnability in Automated Driving (LiAD): Concepts for Applying Learnability Engineering (CALE) Based on Long-Term Learning Effects","authors":"Naomi Y. Mbelekani, Klaus Bengler","doi":"10.3390/info14100519","DOIUrl":"https://doi.org/10.3390/info14100519","url":null,"abstract":"Learnability in Automated Driving (LiAD) is a neglected research topic, especially when considering the unpredictable and intricate ways humans learn to interact and use automated driving systems (ADS) over the sequence of time. Moreover, there is a scarcity of publications dedicated to LiAD (specifically extended learnability methods) to guide the scientific paradigm. As a result, this generates scientific discord and, thus, leaves many facets of long-term learning effects associated with automated driving in dire need of significant research courtesy. This, we believe, is a constraint to knowledge discovery on quality interaction design phenomena. In a sense, it is imperative to abstract knowledge on how long-term effects and learning effects may affect (negatively and positively) users’ learning and mental models. As well as induce changeable behavioural configurations and performances. In view of that, it may be imperative to examine operational concepts that may help researchers envision future scenarios with automation by assessing users’ learning ability, how they learn and what they learn over the sequence of time. As well as constructing a theory of effects (from micro, meso and macro perspectives), which may help profile ergonomic quality design aspects that stand the test of time. As a result, we reviewed the literature on learnability, which we mined for LiAD knowledge discovery from the experience perspective of long-term learning effects. Therefore, the paper offers the reader the resulting discussion points formulated under the Learnability Engineering Life Cycle. For instance, firstly, contextualisation of LiAD with emphasis on extended LiAD. Secondly, conceptualisation and operationalisation of the operational mechanics of LiAD as a concept in ergonomic quality engineering (with an introduction of Concepts for Applying Learnability Engineering (CALE) research based on LiAD knowledge discovery). Thirdly, the systemisation of implementable long-term research strategies towards comprehending behaviour modification associated with extended LiAD. As the vehicle industry revolutionises at a rapid pace towards automation and artificially intelligent (AI) systems, this knowledge is useful for illuminating and instructing quality interaction strategies and Quality Automated Driving (QAD).","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136061016","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}
Takuya Nakata, Sinan Chen, Sachio Saiki, Masahide Nakamura
Software upcycling, a form of software reuse, is a concept that efficiently generates novel, innovative, and value-added development projects by utilizing knowledge extracted from past projects. However, how to integrate the materials derived from these projects for upcycling remains uncertain. This study defines a systematic model for upcycling cases and develops the Sharing Upcycling Cases with Context and Evaluation for Efficient Software Development (SUCCEED) system to support the implementation of new upcycling initiatives by effectively sharing cases within the organization. To ascertain the efficacy of upcycling within our proposed model and system, we formulated three research questions and conducted two distinct experiments. Through surveys, we identified motivations and characteristics of shared upcycling-relevant development cases. Development tasks were divided into groups, those that employed the SUCCEED system and those that did not, in order to discern the enhancements brought about by upcycling. As a result of this research, we accomplished a comprehensive structuring of both technical and experiential knowledge beneficial for development, a feat previously unrealizable through conventional software reuse, and successfully realized reuse in a proactive and closed environment through construction of the wisdom of crowds for upcycling cases. Consequently, it becomes possible to systematically perform software upcycling by leveraging knowledge from existing projects for streamlining of software development.
{"title":"SUCCEED: Sharing Upcycling Cases with Context and Evaluation for Efficient Software Development","authors":"Takuya Nakata, Sinan Chen, Sachio Saiki, Masahide Nakamura","doi":"10.3390/info14090518","DOIUrl":"https://doi.org/10.3390/info14090518","url":null,"abstract":"Software upcycling, a form of software reuse, is a concept that efficiently generates novel, innovative, and value-added development projects by utilizing knowledge extracted from past projects. However, how to integrate the materials derived from these projects for upcycling remains uncertain. This study defines a systematic model for upcycling cases and develops the Sharing Upcycling Cases with Context and Evaluation for Efficient Software Development (SUCCEED) system to support the implementation of new upcycling initiatives by effectively sharing cases within the organization. To ascertain the efficacy of upcycling within our proposed model and system, we formulated three research questions and conducted two distinct experiments. Through surveys, we identified motivations and characteristics of shared upcycling-relevant development cases. Development tasks were divided into groups, those that employed the SUCCEED system and those that did not, in order to discern the enhancements brought about by upcycling. As a result of this research, we accomplished a comprehensive structuring of both technical and experiential knowledge beneficial for development, a feat previously unrealizable through conventional software reuse, and successfully realized reuse in a proactive and closed environment through construction of the wisdom of crowds for upcycling cases. Consequently, it becomes possible to systematically perform software upcycling by leveraging knowledge from existing projects for streamlining of software development.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136129397","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 practical applications, the accuracy of domain terminology translation is an important criterion for the performance evaluation of domain machine translation models. Aiming at the problem of phrase mismatch and improper translation caused by word-by-word translation of English terminology phrases, this paper constructs a dictionary of terminology phrases in the field of electrical engineering and proposes three schemes to integrate the dictionary knowledge into the translation model. Scheme 1 replaces the terminology phrases of the source language. Scheme 2 uses the residual connection at the encoder end after the terminology phrase is replaced. Scheme 3 uses a segmentation method of combining character segmentation and terminology segmentation for the target language and uses an additional loss module in the training process. The results show that all three schemes are superior to the baseline model in two aspects: BLEU value and correct translation rate of terminology words. In the test set, the highest accuracy of terminology words was 48.3% higher than that of the baseline model. The BLEU value is up to 3.6 higher than the baseline model. The phenomenon is also analyzed and discussed in this paper.
{"title":"Machine Translation of Electrical Terminology Constraints","authors":"Zepeng Wang, Yuan Chen, Juwei Zhang","doi":"10.3390/info14090517","DOIUrl":"https://doi.org/10.3390/info14090517","url":null,"abstract":"In practical applications, the accuracy of domain terminology translation is an important criterion for the performance evaluation of domain machine translation models. Aiming at the problem of phrase mismatch and improper translation caused by word-by-word translation of English terminology phrases, this paper constructs a dictionary of terminology phrases in the field of electrical engineering and proposes three schemes to integrate the dictionary knowledge into the translation model. Scheme 1 replaces the terminology phrases of the source language. Scheme 2 uses the residual connection at the encoder end after the terminology phrase is replaced. Scheme 3 uses a segmentation method of combining character segmentation and terminology segmentation for the target language and uses an additional loss module in the training process. The results show that all three schemes are superior to the baseline model in two aspects: BLEU value and correct translation rate of terminology words. In the test set, the highest accuracy of terminology words was 48.3% higher than that of the baseline model. The BLEU value is up to 3.6 higher than the baseline model. The phenomenon is also analyzed and discussed in this paper.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"136306961","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}
Inês Oliveira, Vítor Carvalho, Filomena Soares, Paulo Novais, Eva Oliveira, Lisa Gomes
In recent years, the role of emotions in digital games has gained prominence. Studies confirm emotions’ substantial impact on gaming, influencing interactions, effectiveness, efficiency, and satisfaction. Combining gaming dynamics, Virtual Reality (VR) and the immersive Escape Room genre offers a potent avenue through which to evoke emotions and create a captivating player experience. The primary objective of this study is to explore VR game design specifically for the elicitation of emotions, in combination with the Escape Room genre. We also seek to understand how players perceive and respond to emotional stimuli within the game. Our study involved two distinct groups of participants: Nursing and Games. We employed a questionnaire to collect data on emotions experienced by participants, the game elements triggering these emotions, and their overall user experience. This study demonstrates the potential of VR technology and the Escape Room genre as a powerful means of eliciting emotions in players. “Escape VR: The Guilt” serves as a successful example of how immersive VR gaming can evoke emotions and captivate players.
近年来，情感在数字游戏中的作用越来越突出。研究证实了情绪对游戏的实质性影响，影响互动、效果、效率和满意度。结合游戏动态，虚拟现实(VR)和身临其境的密室逃脱类型提供了一个有效的途径，通过它来唤起情感和创造一个迷人的玩家体验。本研究的主要目的是结合《密室逃脱》这类游戏，探索专门用于激发情感的VR游戏设计。我们还试图理解玩家如何感知和回应游戏中的情感刺激。我们的研究涉及两组不同的参与者:护理和游戏。我们采用问卷调查的方式收集参与者的情绪体验、触发这些情绪的游戏元素以及他们的整体用户体验数据。这项研究证明了VR技术和《密室逃脱》类型作为激发玩家情感的强大手段的潜力。《Escape VR: The Guilt》是沉浸式VR游戏如何唤起情感并吸引玩家的成功范例。
{"title":"Development of a Virtual Reality Escape Room Game for Emotion Elicitation","authors":"Inês Oliveira, Vítor Carvalho, Filomena Soares, Paulo Novais, Eva Oliveira, Lisa Gomes","doi":"10.3390/info14090514","DOIUrl":"https://doi.org/10.3390/info14090514","url":null,"abstract":"In recent years, the role of emotions in digital games has gained prominence. Studies confirm emotions’ substantial impact on gaming, influencing interactions, effectiveness, efficiency, and satisfaction. Combining gaming dynamics, Virtual Reality (VR) and the immersive Escape Room genre offers a potent avenue through which to evoke emotions and create a captivating player experience. The primary objective of this study is to explore VR game design specifically for the elicitation of emotions, in combination with the Escape Room genre. We also seek to understand how players perceive and respond to emotional stimuli within the game. Our study involved two distinct groups of participants: Nursing and Games. We employed a questionnaire to collect data on emotions experienced by participants, the game elements triggering these emotions, and their overall user experience. This study demonstrates the potential of VR technology and the Escape Room genre as a powerful means of eliciting emotions in players. “Escape VR: The Guilt” serves as a successful example of how immersive VR gaming can evoke emotions and captivate players.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"197 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135015725","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 General Data Protection Regulation (GDPR) identifies consent as one of the legal bases for personal data processing and requires that it should be freely given, specific, informed, unambiguous, understandable, and easily revocable. Unfortunately, current technical mechanisms for obtaining consent often do not comply with these requirements. The conceptual consent request framework for mobile devices that is presented in this paper, addresses this issue by following the GDPR requirements on consent and offering a unified user interface for mobile apps. The proposed conceptual framework is evaluated via the development of a City Explorer app with four consent request approaches (custom, functionality-based, app-based, and usage-based) integrated into it. The evaluation shows that the functionality-based consent, which was integrated into the City Explorer app, achieved the best evaluation results and the highest average system usability scale (SUS) score. The functionality-based consent also scored the highest number of SUS points among the four consent templates when evaluated separately from the app. Additionally, we discuss the framework’s reusability and its integration into other mobile apps of different contexts.
{"title":"A Conceptual Consent Request Framework for Mobile Devices","authors":"Olha Drozd, Sabrina Kirrane","doi":"10.3390/info14090515","DOIUrl":"https://doi.org/10.3390/info14090515","url":null,"abstract":"The General Data Protection Regulation (GDPR) identifies consent as one of the legal bases for personal data processing and requires that it should be freely given, specific, informed, unambiguous, understandable, and easily revocable. Unfortunately, current technical mechanisms for obtaining consent often do not comply with these requirements. The conceptual consent request framework for mobile devices that is presented in this paper, addresses this issue by following the GDPR requirements on consent and offering a unified user interface for mobile apps. The proposed conceptual framework is evaluated via the development of a City Explorer app with four consent request approaches (custom, functionality-based, app-based, and usage-based) integrated into it. The evaluation shows that the functionality-based consent, which was integrated into the City Explorer app, achieved the best evaluation results and the highest average system usability scale (SUS) score. The functionality-based consent also scored the highest number of SUS points among the four consent templates when evaluated separately from the app. Additionally, we discuss the framework’s reusability and its integration into other mobile apps of different contexts.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135063145","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}
Mehdi Sadi, B. M. S. Bahar Talukder, Kaniz Mishty, Tauhidur Rahman
Universal adversarial perturbations are image-agnostic and model-independent noise that, when added to any image, can mislead the trained deep convolutional neural networks into the wrong prediction. Since these universal adversarial perturbations can seriously jeopardize the security and integrity of practical deep learning applications, the existing techniques use additional neural networks to detect the existence of these noises at the input image source. In this paper, we demonstrate an attack strategy that, when activated by rogue means (e.g., malware, trojan), can bypass these existing countermeasures by augmenting the adversarial noise at the AI hardware accelerator stage. We demonstrate the accelerator-level universal adversarial noise attack on several deep learning models using co-simulation of the software kernel of the Conv2D function and the Verilog RTL model of the hardware under the FuseSoC environment.
{"title":"Attacking Deep Learning AI Hardware with Universal Adversarial Perturbation","authors":"Mehdi Sadi, B. M. S. Bahar Talukder, Kaniz Mishty, Tauhidur Rahman","doi":"10.3390/info14090516","DOIUrl":"https://doi.org/10.3390/info14090516","url":null,"abstract":"Universal adversarial perturbations are image-agnostic and model-independent noise that, when added to any image, can mislead the trained deep convolutional neural networks into the wrong prediction. Since these universal adversarial perturbations can seriously jeopardize the security and integrity of practical deep learning applications, the existing techniques use additional neural networks to detect the existence of these noises at the input image source. In this paper, we demonstrate an attack strategy that, when activated by rogue means (e.g., malware, trojan), can bypass these existing countermeasures by augmenting the adversarial noise at the AI hardware accelerator stage. We demonstrate the accelerator-level universal adversarial noise attack on several deep learning models using co-simulation of the software kernel of the Conv2D function and the Verilog RTL model of the hardware under the FuseSoC environment.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"93 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135011356","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 three-dimensional MOOC analysis framework was developed, focusing on platform design, organizational mechanisms, and course construction. This framework aims to investigate the current situation of big data MOOCs in the intelligent era, particularly from the perspective of improving the mental health of college students; moreover, the framework summarizes the construction experience and areas for improvement. The construction of 525 big data courses on 16 MOOC platforms is compared and analyzed from three aspects: the platform (including platform construction, resource quantity, and resource quality), organizational mechanism (including the course opening unit, teacher team, and learning norms), and course construction (including course objectives, teaching design, course content, teaching organization, implementation, teaching management, and evaluation). Drawing from the successful practices of international big data MOOCs and excellent Chinese big data MOOCs, and considering the requirements of authoritative government documents, such as the no. 8 document (J.G. [2019]), no. 3 document (J.G. [2015]), no. 1 document (J.G. [2022]), as well as the “Educational Information Technology Standard CELTS-22—Online Course Evaluation Standard”, recommendations about the platform, organizational mechanism, and course construction are provided for the future development of big data MOOCs in China.
{"title":"Analysis of the Current Situation of Big Data MOOCs in the Intelligent Era Based on the Perspective of Improving the Mental Health of College Students","authors":"Hongfeng Sang, Liyi Ma, Nan Ma","doi":"10.3390/info14090511","DOIUrl":"https://doi.org/10.3390/info14090511","url":null,"abstract":"A three-dimensional MOOC analysis framework was developed, focusing on platform design, organizational mechanisms, and course construction. This framework aims to investigate the current situation of big data MOOCs in the intelligent era, particularly from the perspective of improving the mental health of college students; moreover, the framework summarizes the construction experience and areas for improvement. The construction of 525 big data courses on 16 MOOC platforms is compared and analyzed from three aspects: the platform (including platform construction, resource quantity, and resource quality), organizational mechanism (including the course opening unit, teacher team, and learning norms), and course construction (including course objectives, teaching design, course content, teaching organization, implementation, teaching management, and evaluation). Drawing from the successful practices of international big data MOOCs and excellent Chinese big data MOOCs, and considering the requirements of authoritative government documents, such as the no. 8 document (J.G. [2019]), no. 3 document (J.G. [2015]), no. 1 document (J.G. [2022]), as well as the “Educational Information Technology Standard CELTS-22—Online Course Evaluation Standard”, recommendations about the platform, organizational mechanism, and course construction are provided for the future development of big data MOOCs in China.","PeriodicalId":38479,"journal":{"name":"Information (Switzerland)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135201959","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：