In this paper, a 2-D noncausal Markov model is proposed for passive digital image-splicing detection. Different from the traditional model, approach models an image as signal and captures underlying dependencies between current node its neighbors. The parameters are treated discriminative features to differentiate spliced images natural ones. We apply in block discrete cosine transformation domain Meyer wavelet transform domain, cross-domain final classification. support vector machine which...

Software-defined industrial Internet of things (SD-IIoT) exploits federated learning to process the sensitive data at edges, while adaptive poisoning attacks threat security SD-IIoT. To address this problem, article proposes a multi-tentacle (MTFL) framework, which is essential guarantee trustness training in In MTFL, participants with similar tasks are assigned same tentacle group. identify attacks, distribution-based efficient attack detection (TD-EPAD) algorithm presented. And also,...

Abstract Perovskite photodetectors with bipolar photoresponse characteristics are expected to be applied in the field of secure optical communication (SOC). However, how realize perovskite photodetector response remains challenging. Herein, by introducing bismuth iodide (BiI 3 ) into Sn‐Pb mixed precursor solution, 2D FA Bi 2 I 9 is spontaneously formed at bottom a wide‐narrow bandgap‐laminated film. Wavelength‐dependent realized based on absorption difference photoactive region different...

Recently, detecting the traces introduced by content-preserving image manipulations has received a great deal of attention from forensic analyzers. It is well known that median filter widely used nonlinear denoising operator. Therefore, detection filtering important realistic significance in forensics. In this letter, novel local texture operator, named second-order ternary pattern (LTP), proposed for detection. The operator encodes derivative direction variations using 3-valued coding...

Spectrum sensing based dynamic spectrum sharing is one of the key innovative techniques in future 5G communications. When realistic mobile scenarios are concerned, location primary user (PU) great significance to reliable detections and cognitive network enhancements. Given disappearance its emission signals, passive locations tracking PU, nevertheless, remains dramatically different from existing positioning problems. In this investigation, a new joint estimation paradigm, namely deep...

The machine-to-machine (M2M) communications, which achieve the implementation of Internet Things (IoT), can be carried over wireless cellular networks. massive random access (RA) in M2M communications will cause radio network congestion base station (BS), leading to sharp deterioration delay and probability. Access class barring (ACB) that directly control flow machine-type communication (MTC) devices by an ACB factor is efficient scheme prevent BS from traffic overload. In networks, RA...

Semi-Supervised Learning (SSL) is a powerful derivative for humans to discover the hidden knowledge, and will be great substitute data taggers. Although availability of unlabeled rises up huge passion SSL, untrustness leads many unknown security risks. In this paper, we first identify an insidious backdoor threat SSL where training are poisoned by methods migrated from supervised settings. Then, further exploit threat, Deep Neural Backdoor (DeNeB) scheme proposed, which requires less...

This paper considers the stochastic point location (SPL) problem as a learning mechanism trying to locate on real line via interacting with random environment. Compared environment in literatures that confines moving two directions, i.e., left or right, this introduces general triple level which not only tells go but also informs it stay unmoved. It is easy understand, we will prove paper, reported previous just special case of And new algorithm, named walk-based proposed an unknown under...

The threat of data-poisoning backdoor attacks on learning algorithms typically comes from the labeled data. However, in deep semi-supervised (SSL), unknown threats mainly stem unlabeled In this paper, we propose a novel hidden (DeHiB) attack scheme for SSL-based systems. contrast to conventional attacking methods, DeHiB can inject malicious training data learner so as enable SSL model output premeditated results. particular, robust adversarial perturbation generator regularized by unified...

Adversarial attacks and defenses in machine learning deep neural network have been gaining significant attention due to the rapidly growing applications of Internet relevant scenarios. This survey provides a comprehensive overview recent advancements field adversarial attack defense techniques, with focus on network-based classification models. Specifically, we conduct methods state-of-the-art techniques based principles, present them visually appealing tables tree diagrams. is rigorous...

Abstract Flexible perovskite photodetectors (FPDs) are promising for novel wearable devices in bionics, robotics and health care. However, their performance degradation instability during operations remain a grand challenge. Superior flexibility spontaneous functional repair of without the need any external drive or intervention ideal goals FPDs. Herein, by using phenyl disulfide instead alkyl as crosslinking agent, bonds with lower bond energy introduced, thus endowing polyurethane network...

We focus on a full-duplex multi-user MIMO system where the base station (BS) serves multiple uplink and downlink users simultaneously. Both BS mobile stations (MSs) are equipped with antennas. The performance of is limited by self-interference at interference caused users. To address this issue, we propose to jointly design beamformers MSs. An optimization problem formulated maximize weighted sum data rate subject maximum power constraints. Although non-convex, it can be solved via iterative...

Deep neural network are one of the most powerful model for machine learning, which can learn underlying patterns automatically from a large amount data. So it be extensively used in more and Internet-of-Things (IoT) applications. However, training deep models is difficult, suffering overfitting gradient vanishing problem. Besides, parameters multiplication operations make impractical learning to directly execute on target hardware. In this paper, we propose method gradually pruning weakly...

Many factors influence the connection states between nodes of wireless sensor networks, such as physical distance, and network load, making network's edge length dynamic in abundant scenarios. This property makes essentially form a graph with stochastic lengths. In this paper, we study shortest path problem on directional lengths, using reinforcement learning algorithms. regard each random variable following unknown probability distribution aim to find graph. We evaluate performance...

Influence diffusion estimation is a crucial problem in social network analysis. Most prior works mainly focus on predicting the total influence spread, i.e., expected number of influenced nodes given an initial set active (aka. seeds). However, accurate susceptibility, probability being for each individual, more appealing and valuable real-world applications. Previous methods generally adopt Monte Carlo simulation or heuristic rules to estimate influence, resulting high computational cost...