To leverage rich data distributed at the network edge, a new machine-learning paradigm, called edge learning, has emerged where learning algorithms are deployed for providing intelligent services to mobile users. While computing speeds advancing rapidly, communication latency is becoming bottleneck of fast learning. address this issue, work focused on designing low-latency multi-access scheme end, we consider popular privacy-preserving framework, federated (FEEL), global AI-model an...

In this paper, we propose to extend the recently introduced model-agnostic meta-learning algorithm (MAML, Finn, et al., 2017) for low-resource neural machine translation (NMT). We frame as a problem where learn adapt languages based on multilingual high-resource language tasks. use universal lexical representation (Gu 2018b) overcome input-output mismatch across different languages. evaluate proposed strategy using eighteen European (Bg, Cs, Da, De, El, Es, Et, Fr, Hu, It, Lt, Nl, Pl, Pt,...

High peak-to-average power ratio (PAPR) of the transmitted signal is one limitations to employing orthogonal frequency division multiplexing (OFDM) system. In this paper, we propose a new nonlinear companding algorithm that transforms OFDM signals into desirable statistics form defined by linear piecewise function. By introducing variable slopes and an inflexion point in target probability density function, more flexibility effective trade-off between PAPR bit error rate performances can be...

An analog photonic link with the compensation of dispersion-induced power fading is proposed and demonstrated based on phase modulation to intensity conversion in a Sagnac loop. Due velocity mismatch modulator, only incident light wave along clockwise direction effectively modulated by radio frequency signals, while counterclockwise not modulated. After combining two waves polarizer, an optical signal generated, which can be directly detected. In addition, difference between adjusted through...

This article elucidates the critical importance of sustainable land resource management in safeguarding human well-being amid globalization and urbanization, with a particular emphasis on optimal utilization cultivated resources mountainous regions North China. The study focuses within specific village these areas, encompassing open-air cropland, conventional greenhouses, smart greenhouses. Data is sourced from actual plot classifications crop cultivation conditions. A Markov chain model...

Many concerns have been raised regarding the privacy issues in social media. The risks, as well security and of media business, public policy, legislation need to be evaluated studied. However, there is lack effective practical way quantify, measure, evaluate privacy. In this paper, we propose a network model, SONET, for monitoring ranking. model provides novel, effective, Further, proposed also flexible built on real data from risk indicator, PIDX, can calculated time value used control.

In this paper, we propose to extend the recently introduced model-agnostic meta-learning algorithm (MAML) for low-resource neural machine translation (NMT). We frame as a problem, and learn adapt languages based on multilingual high-resource language tasks. use universal lexical representation~\citep{gu2018universal} overcome input-output mismatch across different languages. evaluate proposed strategy using eighteen European (Bg, Cs, Da, De, El, Es, Et, Fr, Hu, It, Lt, Nl, Pl, Pt, Sk, Sl, Sv...

Recently, discriminant locality preserving projection based on L1-norm (DLPP-L1) was developed for robust subspace learning and image classification. It obtains vectors by greedy strategy, i.e., all are optimized individually through maximizing the objective function. Thus, obtained solution does not necessarily best optimize corresponding trace ratio optimization algorithm, which is essential function general dimensionality reduction. results in insufficient recognition accuracy. To tackle...

Mobile devices (e.g., smart phones) and location based services are very popular with today's users. Due to this popularity, there has been a tremendous increase in use of these as result, leakages both query contents is common. In case where users constrained by road network, an attacker can follow users' trajectory ease. Most the current privacy preserving solutions focus on temporal spatial cloaking methods protect privacy. However, vulnerable when subjected continuous environments....

Photonic microwave and millimeter wave (mm-wave) frequency conversion is a promising technique for future electronic systems due to its large bandwidth, electromagnetic immunity, compatibility with other optical systems. However, the gain of photonic mixer often low may be further reduced by fiber dispersion. In this paper, mm-wave using an integrated polarization division multiplexing Mach-Zehnder modulator (PDM-MZM) proposed, which exhibits merits high gain, purity, compensation power...

A wideband microwave Doppler frequency shift (DFS) measurement approach based on a single dual-parallel Mach-Zehnder modulator (DPMZM) is proposed and experimentally demonstrated. The light wave modulated by the transmitted echo signals in DPMZM to generate carrier suppressed double sidebands signals. Based phase difference powers of upper bands lower signal, DFS value, as well sign, can be measured. An experiment carried out. from -100 +100 kHz at frequencies 10, 20, 30, 39 GHz accurately...

In this paper, an efficient Iterative Companding Transform and Filtering (ICTF) technique is proposed for reducing the Peak-to-Average Power Ratio (PAPR) of Orthogonal Frequency Division Multiplexing (OFDM) signal. By means a specially designed iterative procedure, ICTF able to obtain both improved Bit Error Rate (BER) minimized Out-of-Band Interference (OBI) while PAPR significantly. A comprehensive theoretical analysis presented, some important results such as achievable gain, impact...

After providing a brief overview of routing protocols for next-generation sensor networks (NGSNs), the authors propose Bee-Sensor-C, an energy-efficient, swarm intelligence-based, and scalable multipath protocol that integrates dynamic clustering, routing, bee-inspired to meet performance requirements NGSNs. A evaluation is also provided.

Network slicing is a key technology in 5G. Generally, 5G networks employ to provide the isolated and customizable network services for different scenarios (e.g., vertical industries, customers, businesses etc.) form of virtual industry private networks. has potential meet individual requirements users services, terms bandwidth, delay, reliability, mobility. This paper gives an overall introduction management, end-to-end processes, wireless capabilities. Then, this carries on algorithm...

Privacy measurement is critical to evaluate privacy risks for business, public policy and legislation. However, there a lack of effective practical way quantify, measure, privacy. In this paper, we propose three indexes (ranking) social network actor model, i.e., weighted index (w-PIDX), maximum (m-PIDX), composite (c-PIDX). We also introduce novel virtual attribute model describe the combined attributes' behavior. further demonstrate effectiveness these PIDXes various user groups in...

In order to maximize system energy efficiency (EE) under user quality of service (QoS) restraints in Long Term Evolution-Advanced (LTE-A) networks, a constrained joint resource optimization allocation scheme is presented, which NP-hard. Hence, we divide it into three sub-problems reduce computation complexity, i.e., the block (RB) allocation, power distribution, and modulation coding (MCS) assignment for codewords. Then an enhanced heuristic approach GAPSO proposed adopted RB respectively...