教学课程资源库 更多>>

关键词： Computer architecture   Encoding   Task analysis   Approximation algorithms   Training   Genetic algorithms   Computer science   Convolutional variational autoencoder   evolving deep learning   genetic algorithm (GA)   neural architecture search (NAS)  

摘要： Variational autoencoders (VAEs) have demonstrated their superiority in unsupervised learning for image processing in recent years. The performance of the VAEs highly depends on their architectures, which are often handcrafted by the human expertise in deep neural networks (DNNs). However, such expertise is not necessarily available to each of the end users interested. In this article, we propose a novel method to automatically design optimal architectures of VAEs for image classification, called evolving deep convolutional VAE (EvoVAE), based on a genetic algorithm (GA). In the proposed EvoVAE algorithm, the traditional VAEs are first generalized to a more generic and asymmetrical one with four different blocks, and then a variable-length gene encoding mechanism of the GA is presented to search for the optimal network depth. Furthermore, an effective genetic operator is designed to adapt to the proposed variable-length gene encoding strategy. To verify the performance of the proposed algorithm, nine variants of AEs and VAEs are chosen as the peer competitors to perform the comparisons on MNIST, street view house numbers, and CIFAR-10 benchmark datasets. The experiments reveal the superiority of the proposed EvoVAE algorithm, which wins 21 times out of the 24 comparisons and outperforms the best competitors by 1.39%, 14.21%, and 13.03% on the three benchmark datasets, respectively.

关键词： cybersecurity   awareness   curriculum   computer science   information technology   education   framework   courses   content design  

摘要： Cybersecurity is a multifaceted global phenomenon representing complex socio-technical challenges for governments and private sectors. With technology constantly evolving, the types and numbers of cyberattacks affect different users in different ways. The majority of recorded cyberattacks can be traced to human errors. Despite being both knowledge- and environment-dependent, studies show that increasing users' cybersecurity awareness is found to be one of the most effective protective approaches. However, the intangible nature, socio-technical dependencies, constant technological evolutions, and ambiguous impact make it challenging to offer comprehensive strategies for better communicating and combatting cyberattacks. Research in the industrial sector focused on creating institutional proprietary risk-aware cultures. In contrast, in academia, where cybersecurity awareness should be at the core of an academic institution's mission to ensure all graduates are equipped with the skills to combat cyberattacks, most of the research focused on understanding students' attitudes and behaviors after infusing cybersecurity awareness topics into some courses in a program. This work proposes a conceptual Cybersecurity Awareness Framework to guide the implementation of systems to improve the cybersecurity awareness of graduates in any academic institution. This framework comprises constituents designed to continuously improve the development, integration, delivery, and assessment of cybersecurity knowledge into the curriculum of a university across different disciplines and majors;this framework would thus lead to a better awareness among all university graduates, the future workforce. This framework may be adjusted to serve as a blueprint that, once adjusted by academic institutions to accommodate their missions, guides institutions in developing or amending their policies and procedures for the design and assessment of cybersecurity awareness.

关键词： Smart connected worker   Machine learning   Real-time monitoring   Object detection   Text recognition   Energy disaggregation  

摘要： Recent advances in machine learning and computer vision brought to light technologies and algorithms that serve as new opportunities for creating intelligent and efficient manufacturing systems. In this study, the realtime monitoring system of manufacturing workflow for the Smart Connected Worker (SCW) is developed for the small and medium-sized manufacturers (SMMs), which integrates state-of-the-art machine learning techniques with the workplace scenarios of advanced manufacturing systems. Specifically, object detection and text recognition models are investigated and adopted to ameliorate the labor-intensive machine state monitoring process, while artificial neural networks are introduced to enable real-time energy disaggregation for further optimization. The developed system achieved efficient supervision and accurate information analysis in real-time for prolonged working conditions, which could effectively reduce the cost related to human labor, as well as provide an affordable solution for SMMs. The competent experiment results also demonstrated the feasibility and effectiveness of integrating machine learning technologies into the realm of advanced manufacturing systems.

摘要： We provide a self-contained introduction to finite extensive games with perfect information. In these games players proceed in turns having, at each stage, finitely many moves to their disposal, each play always ends, and in each play the players have complete knowledge of the previously made moves. Almost all discussed results are well-known, but often they are not presented in an optimal form. Also, they usually appear in the literature aimed at economists or mathematicians, so the algorithmic or logical aspects are underrepresented. © 2021, European Association for Theoretical Computer Science. All rights reserved.

关键词： Semantics   Decoding   Training   Computer science   Kernel   Logistics   Cybernetics   Common hamming space   cross-view retrieval   decoupled cross-view hashing network (DCHN)   multiview hashing   multiview representation learning  

摘要： Thanks to the low storage cost and high query speed, cross-view hashing (CVH) has been successfully used for similarity search in multimedia retrieval. However, most existing CVH methods use all views to learn a common Hamming space, thus making it difficult to handle the data with increasing views or a large number of views. To overcome these difficulties, we propose a decoupled CVH network (DCHN) approach which consists of a semantic hashing autoencoder module (SHAM) and multiple multiview hashing networks (MHNs). To be specific, SHAM adopts a hashing encoder and decoder to learn a discriminative Hamming space using either a few labels or the number of classes, that is, the so-called flexible inputs. After that, MHN independently projects all samples into the discriminative Hamming space that is treated as an alternative ground truth. In brief, the Hamming space is learned from the semantic space induced from the flexible inputs, which is further used to guide view-specific hashing in an independent fashion. Thanks to such an independent/decoupled paradigm, our method could enjoy high computational efficiency and the capacity of handling the increasing number of views by only using a few labels or the number of classes. For a newly coming view, we only need to add a view-specific network into our model and avoid retraining the entire model using the new and previous views. Extensive experiments are carried out on five widely used multiview databases compared with 15 state-of-the-art approaches. The results show that the proposed independent hashing paradigm is superior to the common joint ones while enjoying high efficiency and the capacity of handling newly coming views.

关键词： Training   Feedback loop   Uncertainty   Probabilistic logic   Hyperspectral imaging   Logistics   Computer science   Active learning (AL)   iterative training sampling (ITS)   iterative training sampling augmentation (ITSA)   spectral-spatial (SS)  

摘要： Training sample selection is a great challenge for hyperspectral image classification (HSIC), specifically when only a very limited number of labeled data samples are available for training. Two recently developed concepts for training sample selection are of particular interest. One is active learning (AL), which augments labeled training samples by including unlabeled data samples as new training samples. The other is iterative training sampling (ITS), which expands data cubes by including additional spatial classification information into the training samples. This article combines AL and ITS simultaneously to derive a joint ITS-AL spectral-spatial (SS) classification approach, to be called ITS augmentation by AL spectral-spatial classification, referred to as ITSA-AL-SS, which can improve SS classification using either AL or ITS alone. The novel idea of ITSA-AL-SS is to take advantage of ITS to expand data cubes by incorporating additional spatial classification information iteratively into the AL-augmented unlabeled data samples to update the current training samples iteration by iteration in one-shot operation. As expected, ITSA-AL-SS is benefited from both ITS and AL to not only further improve classification accuracy but also reduce classification inconsistency.

关键词： Quantum computers  

关键词： Energy system modelling   Resilience   Flexibility   Governance   Storage   Electricity system outage   Energy system architecture  

摘要： Whole energy system modelling is a valuable tool to support the development of policy to decarbonise energy systems, and has been used extensively in the UK for this purpose. However, quantitative insights produced by such models necessarily omit potentially important features of physical and engineering reality. The authors argue that important socio-technical insights can be gained by studying critical events such as the loss of 2.1 GW generation from the electricity system of Great Britain on 9th August 2019, in conjunction with literature on the behaviour of complex systems. Among these insights is the idea that models of the operation and evolution of energy systems can never be complete. Both system behaviour (operation) and the emergence and evolution of structure in such systems are formally uncomputable. This provides a starting point for a discussion of the need for additional tools, drawn from the System Architecture literature, to support the design and realisation of future, fully-decarbonised systems with high penetrations of renewable energy. Desirable properties of System Architectures, including current and future Energy System Architectures, are discussed. These include resilience and flexibility, for which there is an extensive literature. They also include the properties of comprehensibility, which helps to make complex systems easier to operate, and of evolvability, for which a working definition is offered.

关键词： Wireless sensor networks   Sensors   Heuristic algorithms   Data collection   Optimization   Computer science   Approximation algorithms   Heterogeneous wireless sensor networks (HWSNs)   the Internet of Things (IoT)   the IoT-enabled WSN   bees based algorithm   network lifetime   dominating set   energy consumption   data collection  

摘要： Wireless Sensor Networks (WSNs) lend themselves to a wide variety of applications in our daily lives, such as environmental monitoring, safety, health-care, animal monitoring, etc. However, one of the key issues in WSN is energy constraints. This makes energy-conservation one of the major keys to the efficient functioning and lifetime of WSN. In this paper, given a network of nodes with heterogeneous energy, our goal is to determine energy-aware disjoint dominating sets (DSs) that work as data collection nodes in each round, to improve overall WSN lifetime. In order to accomplish this goal, we propose an intelligent data collection technique with two phases, the collector nodes selection, and the data gathering path formation and collection phases. In the collector nodes selection phase, an energy-aware algorithm based on swarm intelligence is proposed to construct disjoint dominating sets that work as collector nodes in each round. In the data gathering path formation and collection phase, data gathering path is determined for achieving maximal data collection efficiency and reduced energy consumption. The efficiency of our proposed technique is proved mathematically and through simulations.