Machine learning has largely applied to malware detection and classification, due the ineffectiveness of signature-based method toward rapid proliferation. Although state-of-the-art machine models tend achieve high performances, they require a large number training samples. It is infeasible train with sufficient samples while facing newly appeared variants. Therefore, it important for security protectors model given small set data, which can identify variants based on similarity function. In...

Traditional voting and bidding systems largely rely on paperwork human resources throughout the process, which can incur high costs in terms of both time money. Electronic electronic be used to reduce costs, many new have been introduced. However, most require a powerful trusted third party guarantee system integrity security. With developments blockchain technology, research has begun highlight core concept decentralization. In this study, we introduce first decentralized based smart...

Domain adaptive semantic segmentation methods commonly utilize stage-wise training, consisting of a warm-up and self-training stage. However, this popular approach still faces several challenges in each stage: for warm-up, the widely adopted adversarial training often results limited performance gain, due to blind feature alignment; self-training, finding proper categorical thresholds is very tricky. To alleviate these issues, we first propose replace stage by novel symmetric knowledge...

With the rapid development of cloud computing, an increasing number companies are adopting storage technology to reduce overhead. However, ensure privacy sensitive data, uploaded data need be encrypted before being outsourced cloud. The concept public-key encryption with keyword search (PEKS) was introduced by Boneh et al. provide flexible usage data. Unfortunately, most PEKS schemes not secure against inside guessing attacks (IKGA), so information trapdoor may leaked adversary. To solve...

In this paper, we propose a novel loop closure detection algorithm that uses graph attention neural networks to encode semantic graphs perform place recognition and then use registration estimate the 6 DoF relative pose constraint. Our has two key modules, namely, encoder module comparison module. The employs efficiently spatial, geometric information from of input point cloud. We self-attention mechanism in both node-embedding graph-embedding steps create distinctive vectors. vectors...

Deep learning has achieved significant success in the field of remote sensing image change detection (CD), yet two major challenges remain: scarcity sub-meter, all-inclusive open-source CD datasets, and difficulty achieving consistent satisfactory results across images with varying areas. To address these issues, we introduce JL1-CD dataset, which contains 5,000 pairs 512 x pixel a resolution 0.5 to 0.75 meters. Additionally, propose multi-teacher knowledge distillation (MTKD) framework for...

Abstract The industrial Internet of Things (IIoT) integrates sensors, instruments, equipment and applications, enabling traditional industries to automate intelligently process data. To reduce the cost demand required service equipment, IIoT relies on cloud computing further store Public-key encryption with keyword search (PEKS) plays an important role, due its functionality, ensure privacy confidentiality outsourced data maintenance flexibility in use Recently, Huang Li proposed ‘public-key...

In the era of cloud computing, massive quantities data are encrypted and uploaded to realize a variety applications services while protecting user confidentiality. Accordingly, formulation methods for efficiently searching has become critical problem. Public-key encryption with keyword search is an efficient solution that allows owner generate keywords given document also allowing corresponding trapdoor searching. Huang Li proposed public-key authenticated (PAEKS) scheme resist guessing...

Bandwidth estimation and congestion control for real-time communications (i.e., audio video conferencing) remains a difficult problem, despite many years of research. Achieving high quality experience (QoE) end users requires continual updates due to changing network architectures technologies. In this paper, we apply reinforcement learning the first time problem (RTC), where seek optimize user-perceived quality. We present initial proof-of-concept results, learn an agent sending rate in RTC...

Blockchain is the core technology underlying first decentralized cryptocurrency, Bitcoin, introduced by Nakamoto in 2008. Since then, blockchain has many more advancements that are being developed and experimented. In particular, recent platforms such as Ethereum offer general executable scripts, namely smart contracts, can be employed to develop applications (DApps) domains beyond payment. However, transparency of data raises concerns for require high privacy level. Therefore, enhancing...

Model checking is an automated formal verification method to verify whether epistemic multi-agent systems adhere property specifications.Although there extensive literature on qualitative properties such as safety and liveness, still a lack of quantitative uncertain verifications for these systems.In environments, agents must make judicious decisions based subjective epistemic.To measurable in systems, this paper extends fuzzy computation tree logic by introducing modalities proposing new...

The application of digital imaging to color measurement is more and extensive, which also provides new ideas for gemology. In this paper, the single-factor experiment method orthogonal design are used study influence parameters system (DIS) on gem measurement. Among them, illuminance has most significant impact results, followed by exposure time, change distance an insignificant results; through range analysis variance, suitable red jadeite 1600 lx, 12,500 μs, 25 cm.

Due to the lack of training labels and difficulty annotating, dealing with adverse driving conditions such as nighttime has posed a huge challenge perception system autonomous vehicles. Therefore, adapting knowledge from labelled daytime domain an unlabelled been widely researched. In addition datasets, existing datasets usually provide images corresponding reference captured at nearby locations for reference. The key is minimize performance gap between two domains. this paper, we propose...

The verification of reachability properties fuzzy systems is usually based on the Kripke structure or possibilistic structure. However, not enough to describe nondeterministic and concurrent in real life. In this paper, firstly, we propose generalized decision process as model systems, deduce possibilities sets paths relying defining schedulers. Then, give matrices calculation methods maximal minimal eventual reachability, always constrained repeated persistent reachability. Finally, a...

Nowadays, RFID is used to identify a wide range of work pieces or individual products for tracking their movements during the product life cycle. For future purposes idea storing only single ID must be extended "Digital Product Memory". A Digital Memory provides digital diary complete cycle and embedded in itself. It records all relevant ambient parameters form. Sensors note where when an interaction with takes place. People can access these diaries at any time. The autonomous behavior...

Blind multisignature (BMS), first introduced by Horster et al, constitutes a crucial primitive that allows user to generate signature of message from multiple signers, while the signers cannot obtain any information about message. With these useful properties, blind is suitable for electronic payments and voting. However, most current BMS schemes may be attacked quantum computers in future because they are based on traditional number theories, such as discrete logarithm assumption large...

Since machine learning and deep are largely used for image recognition in real-world applications, how to avoid adversarial attacks become an important issue. It is common that attackers add perturbation a normal order fool the models. The N-pixel attack one of recently popular methods by simply changing few pixels image. We observe leads obvious difference with its neighboring pixels. Therefore, this research aims defend based on reconstruction. develop three-staged reconstructing algorithm...

In public key infrastructure, a certificate, issued by certificate authority (CA), is used to guarantee the connection between user and her/his key. order improve efficiency, concept of implicit protocol introduced Girault Gönther. existing protocol, must issue request CA for each pair. However, in certain applications (e.g., IoT, sensor networks, cryptocurrency), (or device) will have multiple public/private pairs that are related same identity. Therefore, communication cost be linearly...