The Internet of things (IoT), which integrates a variety devices into networks to provide advanced and intelligent services, has protect user privacy address attacks such as spoofing attacks, denial service (DoS) jamming, eavesdropping. We investigate the attack model for IoT systems review security solutions based on machine-learning (ML) techniques including supervised learning, unsupervised reinforcement learning (RL). ML-based authentication, access control, secure offloading, malware...

Mobile edge computing usually uses caching to support multimedia contents in 5G mobile Internet reduce the overhead and latency. (MEC) systems are vulnerable various attacks such as denial of service rogue attacks. This article investigates attack models MEC systems, focusing on both offloading procedures. In this article, we propose security solutions that apply reinforcement learning (RL) techniques provide secure nodes against jamming We also present lightweight authentication...

Mobile edge computing helps healthcare Internet of Things (IoT) devices with energy harvesting provide satisfactory quality experiences for computation intensive applications. We propose a reinforcement learning (RL)-based privacy-aware offloading scheme to help IoT protect both the user location privacy and usage pattern privacy. More specifically, this enables device choose rate that improves performance, protects privacy, saves without being aware leakage, consumption, model. This uses...

Physical (PHY)-layer authentication systems can exploit channel state information of radio transmitters to detect spoofing attacks in wireless networks. The use multiple landmarks each with antennas enhances the spatial resolution transmitters, and thus improves detection accuracy PHY-layer authentication. Unlike most existing schemes that apply hypothesis tests rely on known model, we propose a logistic regression-based remove assumption be applicable more generic Frank-Wolfe algorithm is...

Mobile edge computing systems help improve the performance of computational-intensive applications on mobile devices and have to resist jamming attacks heavy interference. In this paper, we present a reinforcement learning based offloading scheme for against interference, which uses safe avoid choosing risky policy that fails meet computational latency requirements tasks. This enables device choose device, transmit power rate its utility including sharing gain, latency, energy consumption...

The dense deployment of small cells in 5G cellular networks raises the issue controlling downlink inter-cell interference under time-varying channel states. In this paper, we propose a reinforcement learning based power control scheme to suppress and save energy for ultra-dense cells. This enables base stations schedule transmit without knowing distribution states neighboring A deep algorithm is designed further accelerate with large number active users. Analytical convergence performance...

Internet of things (IoT) that integrate a variety devices into networks to provide advanced and intelligent services have protect user privacy address attacks such as spoofing attacks, denial service jamming eavesdropping. In this article, we investigate the attack model for IoT systems, review security solutions based on machine learning techniques including supervised learning, unsupervised reinforcement learning. We focus authentication, access control, secure offloading malware detection...

In this letter, we present an anti-jamming underwater transmission framework that applies reinforcement learning to control the transmit power and uses transducer mobility address jamming in acoustic networks. The deep Q-networks-based scheme can achieve optimal node without knowing model channel dynamic game. Experiments performed with transducers a non-anechoic pool show our proposed reduce bit error rate of against reactive compared Q-learning based scheme.

In this letter, we propose a physical (PHY)-layer authentication framework to detect spoofing attacks in underwater sensor networks. This scheme exploits the power delay profile of acoustic channel discriminate sensors and applies reinforcement learning (RL) choose parameter without being aware network model. We an RL-based provide light-weight detection deep further improve accuracy for sinks that support learning. Experiment results show improves increases utility compared with benchmark...

Cloud-based malware detection improves the performance for mobile devices that offload their tasks to security servers with much larger database and powerful computational resources. In this paper, we investigate competition of radio transmission bandwidths data sharing server in dynamic game, which each device chooses its offloading rate application traces server. As Q-learning technique has a slow learning game high dimension, have designed based on hotbooting-Q techniques, initiates...

In this paper, we propose a physical (PHY)-layer authentication system that exploits the channel state information of radio transmitters to detect spoofing attacks in wireless networks. By using multiple landmarks and antennas estimation, enhances spatial resolution thus improves detection accuracy. Unlike most existing hypothesis test based PHY-layer schemes rely on known model, our proposed uses logistic regression remove assumption model is applicable more generic The Frank-Wolfe...

Edge computing for mobile devices in vehicular ad hoc networks (VANETs) has to address rogue edge attacks, which a node claims be the serving vehicle steal user secrets and help launch other attacks such as man-in-the-middle attacks. Rogue detection VANETs is more challenging than spoofing indoor wireless due high mobility of onboard units (OBUs) large-scale network infrastructure with roadside (RSUs). In this paper, we propose physical (PHY)- layer scheme according shared ambient radio...

Safe reinforcement learning is important for the safety critical applications especially network security, as exploration of some dangerous actions can result in huge short-term losses such failure or large scale privacy leakage. In this paper, we propose a algorithm with safe and uses transfer to reduce initial random exploration. A blacklist maintained record most state-action pairs constraint. deep version convolutional neural estimate risk levels thus further improves accelerates speed...

Mobile edge computing usually uses cache to support multimedia contents in 5G mobile Internet reduce the overhead and latency. caching (MEC) systems are vulnerable various attacks such as denial of service rogue attacks. This article investigates attack models MEC systems, focusing on both offloading procedures. In this paper, we propose security solutions that apply reinforcement learning (RL) techniques provide secure nodes against jamming We also present light-weight authentication...

Malicious URLs result in malware installation, privacy leakage and illegal funding of mobile devices computers. However, attackers frequently change domain names to avoid static detection the malicious URL has address variance structure names, which seriously degrades accuracy fixed policy selection impedes optimal with theoretical analysis. In this paper, we propose an accurate protect Internet users from accessing URLs, designs a multi-feature analysis architecture exploit lexical...

The rapid development of multi-view 3D human pose estimation (HPE) is attributed to the maturation monocular 2D HPE and geometry reconstruction. However, detection outliers in occluded views due neglect view consistency, implausible poses lack coherence, remain challenges. To solve this, we introduce a Multi-View Fusion module refine results by establishing correlations. Then, Holistic Triangulation proposed infer whole as an entirety, anatomy prior injected maintain coherence improve...