Pub Date : 2022-07-27DOI: 10.48550/arXiv.2207.13410
K. Khabarlak
Many applications require high accuracy of neural networks as well as low latency and user data privacy guaranty. Face anti-spoofing is one of such tasks. However, a single model might not give the best results for different device performance categories, while training multiple models is time consuming. In this work we present Post-Train Adaptive (PTA) block. Such a block is simple in structure and offers a drop-in replacement for the MobileNetV2 Inverted Residual block. The PTA block has multiple branches with different computation costs. The branch to execute can be selected on-demand and at runtime; thus, offering different inference times and configuration capability for multiple device tiers. Crucially, the model is trained once and can be easily reconfigured after training, even directly on a mobile device. In addition, the proposed approach shows substantially better overall performance in comparison to the original MobileNetV2 as tested on CelebA-Spoof dataset. Different PTA block configurations are sampled at training time, which also decreases overall wall-clock time needed to train the model. While we present computational results for the anti-spoofing problem, the MobileNetV2 with PTA blocks is applicable to any problem solvable with convolutional neural networks, which makes the results presented practically significant.
{"title":"Post-Train Adaptive MobileNet for Fast Anti-Spoofing (keynote)","authors":"K. Khabarlak","doi":"10.48550/arXiv.2207.13410","DOIUrl":"https://doi.org/10.48550/arXiv.2207.13410","url":null,"abstract":"Many applications require high accuracy of neural networks as well as low latency and user data privacy guaranty. Face anti-spoofing is one of such tasks. However, a single model might not give the best results for different device performance categories, while training multiple models is time consuming. In this work we present Post-Train Adaptive (PTA) block. Such a block is simple in structure and offers a drop-in replacement for the MobileNetV2 Inverted Residual block. The PTA block has multiple branches with different computation costs. The branch to execute can be selected on-demand and at runtime; thus, offering different inference times and configuration capability for multiple device tiers. Crucially, the model is trained once and can be easily reconfigured after training, even directly on a mobile device. In addition, the proposed approach shows substantially better overall performance in comparison to the original MobileNetV2 as tested on CelebA-Spoof dataset. Different PTA block configurations are sampled at training time, which also decreases overall wall-clock time needed to train the model. While we present computational results for the anti-spoofing problem, the MobileNetV2 with PTA blocks is applicable to any problem solvable with convolutional neural networks, which makes the results presented practically significant.","PeriodicalId":251021,"journal":{"name":"International Workshop on Intelligent Information Technologies & Systems of Information Security","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116809752","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}
V. Morkun, N. Morkun, A. Pikilnyak, S. Semerikov, O. Serdiuk, I. Gaponenko
It is proposed to carry out the spatial effect of high-energy ultrasound dynamic effects with controlled characteristics on the solid phase particles of the ore pulp in the deslimer input product to increase the efficiency of thickening and desliming processes of iron ore beneficiation products. The above allows predicting the characteristics of particle gravitational sedimentation based on an assessment of the spatial dynamics of pulp solid- phase particles under the controlled action of high-energy ultrasound and fuzzy logical inference. The object of study is the assessment of the characteristics and the process of control the operations of thickening and deslaming of iron ore beneficiation products in the conditions of the technological line of the ore beneficiation plant. The subject of study is a cyber-physical system based on the use of high-energy ultrasound radiation pressure effects on iron-containing beneficiation products in the technological processes of thickening and desliming. The working hypothesis of the project is that there is a relationship between the physical-mechanical and chemical-mineralogical characteristics of the iron ore pulp solid- phase particles and their behavior in technological flows under the influence of controlled ultrasonic radiation, based on which the imitation modeling of the gravitational sedimentation process of the iron ore pulp solid-phase particles can be performed directly in the technological process. Also, the optimal control actions concerning the processes of thickening and desliming can be determined.
{"title":"The Cyber-Physical System for Increasing the Efficiency of the Iron Ore Desliming Process","authors":"V. Morkun, N. Morkun, A. Pikilnyak, S. Semerikov, O. Serdiuk, I. Gaponenko","doi":"10.31812/123456789/4373","DOIUrl":"https://doi.org/10.31812/123456789/4373","url":null,"abstract":"It is proposed to carry out the spatial effect of high-energy ultrasound dynamic effects with controlled characteristics on the solid phase particles of the ore pulp in the deslimer input product to increase the efficiency of thickening and desliming processes of iron ore beneficiation products. The above allows predicting the characteristics of particle gravitational sedimentation based on an assessment of the spatial dynamics of pulp solid- phase particles under the controlled action of high-energy ultrasound and fuzzy logical inference. The object of study is the assessment of the characteristics and the process of control the operations of thickening and deslaming of iron ore beneficiation products in the conditions of the technological line of the ore beneficiation plant. The subject of study is a cyber-physical system based on the use of high-energy ultrasound radiation pressure effects on iron-containing beneficiation products in the technological processes of thickening and desliming. The working hypothesis of the project is that there is a relationship between the physical-mechanical and chemical-mineralogical characteristics of the iron ore pulp solid- phase particles and their behavior in technological flows under the influence of controlled ultrasonic radiation, based on which the imitation modeling of the gravitational sedimentation process of the iron ore pulp solid-phase particles can be performed directly in the technological process. Also, the optimal control actions concerning the processes of thickening and desliming can be determined.","PeriodicalId":251021,"journal":{"name":"International Workshop on Intelligent Information Technologies & Systems of Information Security","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130667738","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}
I. Pilkevych, O. Boychenko, N. Lobanchykova, T. Vakaliuk, S. Semerikov
Modern types of internal threats and methods of counteracting these threats are analyzed. It is established that increasing the competence of the staff of the institution through training (education) is the most effective method of counteracting internal threats to information. A method for assessing the influence of personnel competence on institutional information security is proposed. This method takes into account violator models and information threat models that are designed for a specific institution. The method proposes to assess the competence of the staff of the institution by three components: the level of knowledge, skills, and character traits (personal qualities). It is proposed to assess the level of knowledge based on the results of test tasks of different levels of complexity. Not only the number of correct answers is taken into account, but also the complexity of test tasks. It is proposed to assess the assessment of the level of skills as the ratio of the number of correctly performed practical tasks to the total number of practical tasks. It is assumed that the number of practical tasks, their complexity is determined for each institution by the direction of activity. It is proposed to use a list of character traits for each position to assess the character traits (personal qualities) that a person must have to effectively perform the tasks assigned to him. This list should be developed in each institution. It is proposed to establish a quantitative assessment of the state of information security, defining it as restoring the amount of probability of occurrence of a threat from the relevant employee to the product of the general threat and employees of the institution. An experiment was conducted, the results of which form a particular institution show different values of the level of information security of the institution for different values of the competence of the staff of the institution. It is shown that with the increase of the level of competence of the staff of the institution the state of information security in the institution increases.
{"title":"Method of Assessing the Influence of Personnel Competence on Institutional Information Security","authors":"I. Pilkevych, O. Boychenko, N. Lobanchykova, T. Vakaliuk, S. Semerikov","doi":"10.31812/123456789/4374","DOIUrl":"https://doi.org/10.31812/123456789/4374","url":null,"abstract":"Modern types of internal threats and methods of counteracting these threats are analyzed. It is established that increasing the competence of the staff of the institution through training (education) is the most effective method of counteracting internal threats to information. A method for assessing the influence of personnel competence on institutional information security is proposed. This method takes into account violator models and information threat models that are designed for a specific institution. The method proposes to assess the competence of the staff of the institution by three components: the level of knowledge, skills, and character traits (personal qualities). It is proposed to assess the level of knowledge based on the results of test tasks of different levels of complexity. Not only the number of correct answers is taken into account, but also the complexity of test tasks. It is proposed to assess the assessment of the level of skills as the ratio of the number of correctly performed practical tasks to the total number of practical tasks. It is assumed that the number of practical tasks, their complexity is determined for each institution by the direction of activity. It is proposed to use a list of character traits for each position to assess the character traits (personal qualities) that a person must have to effectively perform the tasks assigned to him. This list should be developed in each institution. It is proposed to establish a quantitative assessment of the state of information security, defining it as restoring the amount of probability of occurrence of a threat from the relevant employee to the product of the general threat and employees of the institution. An experiment was conducted, the results of which form a particular institution show different values of the level of information security of the institution for different values of the competence of the staff of the institution. It is shown that with the increase of the level of competence of the staff of the institution the state of information security in the institution increases.","PeriodicalId":251021,"journal":{"name":"International Workshop on Intelligent Information Technologies & Systems of Information Security","volume":"104 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132867648","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：