Design and Evaluation of CRDMAC： Circular RTR Directional MAC Protocol for WANETs

Using directional antennas in Wireless Ad hoc Networks (WANETs) offers great potential of reducing the radio interference, and improving the communication throughput. Directional antennas, however, introduces new problems in the wireless Media Access Control (MAC), that is, the deafness and new hidden terminal problem, which may cause severe performance degradation. To solve the problems, we propose an effective Circular RTR Directional MAC (CRDMAC) protocol for WANETs by using a sub-transmission channel and Ready to Receive (RTR) packets, which modifies the IEEE 802.11 Distributed Coordinated Function (DCF). The sub-channel avoids collisions to other ongoing transmission, and the RTR packets notify the neighbor nodes that the mutual transmission has been finished. We evaluate the CRDMAC protocol through simulations and the results show that the proposed protocol outperforms existing DMAC (directional MAC) protocol and the CRCM (Circular RTS and CTS MAC) protocol in terms of throughput and packet drop rate.

Edge Computing-Based Joint Client Selection and Networking Scheme for Federated Learning in Vehicular IoT

In order to support advanced vehicular Internet-of-Things(IoT)applications,information exchanges among different vehicles are required to find efficient solutions for catering to different application requirements in complex and dynamic vehicular environments.Federated learning(FL),which is a type of distributed learning technology,has been attracting great interest in recent years as it performs knowledge exchange among different network entities without a violation of user privacy.However,client selection and networking scheme for enabling FL in dynamic vehicular environments,which determines the communication delay between FL clients and the central server that aggregates the models received from the clients,is still under-explored.In this paper,we propose an edge computing-based joint client selection and networking scheme for vehicular IoT.The proposed scheme assigns some vehicles as edge vehicles by employing a distributed approach,and uses the edge vehicles as FL clients to conduct the training of local models,which learns optimal behaviors based on the interaction with environments.The clients also work as forwarder nodes in information sharing among network entities.The client selection takes into account the vehicle velocity,vehicle distribution,and the wireless link connectivity between vehicles using a fuzzy logic algorithm,resulting in an efficient learning and networking architecture.We use computer simulations to evaluate the proposed scheme in terms of the communication overhead and the information covered in learning.

双标签监督的几何约束对抗训练

近年来的研究表明,对抗训练是一种有效的防御对抗样本攻击的方法.然而,现有的对抗训练策略在提升模型鲁棒性的同时会造成模型的泛化能力下降.现阶段主流的对抗训练方法通常都是独立地处理每个训练样本,而忽略了样本之间的关系,这使得模型无法充分挖掘样本间的几何关系来学习更鲁棒的模型,以便更好地防御对抗攻击.因此,重点研究如何在对抗训练过程中保持样本间的几何结构稳定性,达到提升模型鲁棒性的目的.具体而言,在对抗训练中,设计了一种新的几何结构约束方法,其目的是保持自然样本与对抗样本的特征空间分布一致性.此外,提出了一种基于双标签的监督学习方法,该方法同时采用自然样本和对抗样本的标签对模型进行联合监督训练.最后,分析了双标签监督学习方法的特性,试图从理论上解释对抗样本的工作机理.多个基准数据集上的实验结果表明:相比于已有方法,该方法有效地提升了模型的鲁棒性且保持了较好的泛化精度.相关代码已经开源:https://github.com/SkyKuang/DGCAT.

A Cooperative Forwarding Scheme for VANET Routing Protocols

Providing efficient packet delivery in vehicular ad hoc networks （VANETs） is particularly challenging due to the vehicle move- ment and lossy wireless channels. A data packet can be lost at a forwarding node even when a proper node is selected as the for- warding node. In this paper, we propose a loss-tolerant scheme for unicast routing protocols in VANETs. The proposed scheme employs multiple forwarding nodes to improve the packet reception ratio at the forwarding nodes. The scheme uses network coding to reduce the number of required transmissions, resulting in a significant improvement in end-to-end packet delivery ratio with low message overhead. The effectiveness of the proposed scheme is evaluated by using both theoretical analysis and computer sim-

An Instance-Learning-Based Intrusion-Detection System for Wireless Sensor Networks

This paper proposes an instance-learning-based intrusion-detection system （IL-IDS） for wireless sensor networks （WSNs）. The goal of the proposed system is to detect routing attacks on a WSN. Taking an existing instance-learning algorithm for wired networks as our basis, we propose IL-IDS for handling routing security problems in a WSN. Attacks on a routing protocol for a WSN include black hole attack and sinkhole attack. The basic idea of our system is to differentiate the changes between secure instances and attack instances. Considering the limited resources of sensor nodes, the existing algorithm cannot be used directly in a WSN. Our system mainly comprises four parts： feature vector selection, threshold selection, instance data processing, and instance determination. We create a feature vector form composed of the attributes that changes obviously when an attack occurs within the network. For the data processing in resource-constrained sensor nodes, we propose a data-reduction scheme based on the clustering algorithm. For instance determination, we provide a threshold-selection scheme and describe the concrete-instance-determination mechanism of the system. Finally, we simulate and evaluate the proposed IL-IDS for different types of attacks.

Theoretical Treatment of Target Coverage in Wireless Sensor Networks

The target coverage is an important yet challenging problem in wireless sensor networks, especially when both coverage and energy constraints should be taken into account. Due to its nonlinear nature, previous studies of this problem have mainly focused on heuristic algorithms; the theoretical bound remains unknown. Moreover, the most popular method used in the previous literature, i.e., discretization of continuous time, has yet to be justified. This paper fills in these gaps with two theoretical results. The first one is a formal justification for the method. We use a simple example to illustrate the procedure of transforming a solution in time domain into a corresponding solution in the pattern domain with the same network lifetime and obtain two key observations. After that, we formally prove these two observations and use them as the basis to justify the method. The second result is an algorithm that can guarantee the network lifetime to be at least （1 - ε） of the optimal network lifetime, where ε can be made arbitrarily small depending on the required precision. The algorithm is based on the column generation （CG） theory, which decomposes the original problem into two sub-problems and iteratively solves them in a way that approaches the optimal solution. Moreover, we developed several constructive approaches to further optimize the algorithm. Numerical results verify the efficiency of our CG-based algorithm.

Design and Implementation of NAREGI SuperScheduler Based on the OGSA Architecture

NAREGI is a 5-year Japanese National Grid Project during 2003--2007, whose chief aim is to develop a set of grid middleware to serve as a basis for future e-Science. NAREGI also aims to lead the way in standardization of grid middleware, based on the OGSA architecture. Its super-scheduler is based on the proposed OGSA-EMS Architecture, in that it becomes the first working implementation that implements the documented component relationships within the OGSA-EMS architecture document v.l.0. Through the efforts and experience in the design and implementation, it has been confirmed that the documented OGSA-EMS architecture is quite feasible, but will require significant amount of refinement and speed improvements to finalize its detailed specifications. The super-scheduler also supports co-allocation across multiple sites to support automated execution of grid-based MPIs that execute across machines. Such a resource allocation requires sophisticated interactions between the OGSA-EMS components not covered in the current OGSA-EMS architecture, some of which are non-trivial. Overall, job scheduling with OGSA-EMS has proven to not only work, but also that its job allocation and execution time is within reasonable bounds.

Near real-time characterisation of urban environments: a holistic approach for monitoring dengue fever risk areas

Despite frequent use of digital devices in everyday life,cost-effective measurement of public health issues in urban areas is still challenging.This study was,therefore,planned to extract land-use types using object-based and spatial metric approaches to explore the dengue incidence in relation to the surrounding environment in near real-time using Google and Advanced Land Observation Satellite images.The characterised image showed useful classification of an urban areawith 77%accuracy and 0.68 kappa.Geospatial analysis on public health data indicated that most of the dengue cases were found in densely populated areas surrounded by dense vegetation.People living in independent houses having sparsely vegetated surroundings were found to be less vulnerable.Disease incidence was more prevalent in people of 5-24 years of age(67%);while in terms of occupation,mostly students,the unemployed,labourers and farmers(88%)were affected.In general,males were affected slightly more(10%)than females.Proximity analyses indicated that most of the dengue cases were around institutions(40%),religious places(18%)and markets(15%).Thus,usage of Digital Earth scalable tools for monitoring health issues would open new ways for maintaining a healthy and sustainable society in the years ahead.

Iterative Method for Constructing Complete Complementary Sequences with Lengths of 2^mN

Complete complementary sequences are widely used in spectrum spread communications because of their ideal correlation functions. A previous method generates complete complementary sequences with lengths of N^nN （n,N ∈ Z^＋）. This paper presents a new iterative method to construct complete complementary sequences with lengths of 2^mN （m,N ∈ Z^＋）. The analysis proves that this method can produce many sequence sets that do not appear in sequence sets generated by the former method, especially shorter sequence sets. The result will certainly increase the application of complete complementary sequences in communication engineering and related fields.

基于采样和加权损失函数的模型窃取攻击方法

模型窃取攻击旨在获得一个和目标受害模型功能相似的替代模型.现有的方法主要采用数据生成或数据选择方法和交叉熵损失函数去获得一个较好的攻击效果.据此,本文着重研究了攻击过程中这两个极为重要的模块:数据采样和损失函数.同时,本文提出了一个新颖的模型窃取攻击方法S&W,其包含了一种新的采样策略和一个精心设计的加权损失函数.首先,新的采样策略更加关注于从受害者模型中获得更多信息的重要样本.与此同时,本文通过引入k-Center算法达到选择样本的多样性的目的.其次,受到经典Focal损失函数的启发,本文设计了一种新的加权损失函数.该损失函数主要关注于受害者模型和替代模型对于相同输入所给出的输出之间的差异,从而促使替代模型模拟受害者模型.在4个常用的数据集上,我们通过实验证明了本文提出的方法的有效性.相比于之前最好的方法,本文方法最高有5.03%的性能提升.

在61比特可编程超导量子处理器上对量子多体态进行量子神经元感知

对具有不同性质和物相的多体量子态进行分类是量子多体物理学中最基本的任务之一.然而,由于巨大数量的相互作用的粒子所产生的指数级的复杂性,大规模量子态的分类对于经典的方法来说极具挑战性.本文提出了一种新的方法,称为量子神经元感知.利用一个61比特的超导量子处理器作为演示,作者表明该方案可以有效地对两种不同类型的多体现象,即遍历相和局域相,进行分类.量子神经元感知过程使他们能够通过只测量一个量子比特来区分这些多体物相,并提供比传统方法(如测量不平衡度)更好的分辨率.本研究证明了量子神经元感知在近期量子处理器应用的可行性和扩展性,并为探索更大规模系统中的量子多体现象开辟了新的途径.

Supporting feature model refinement with updatable view

In the research of software reuse, feature models have been widely adopted to capture, organize and reuse the requirements of a set of similar applications in a software do- main. However, the construction, especially the refinement, of feature models is a labor-intensive process, and there lacks an effective way to aid domain engineers in refining feature models. In this paper, we propose a new approach to support interactive refinement of feature models based on the view updating technique. The basic idea of our approach is to first extract features and relationships of interest from a possibly large and complicated feature model, then organize them into a comprehensible view, and finally refine the feature model through modifications on the view. The main characteristics of this approach are twofold： a set of powerful rules （as the slicing criterion） to slice the feature model into a view auto- matically, and a novel use of a bidirectional transformation language to make the view updatable. We have successfully developed a tool, and a nontrivial case study shows the feasi- bility of this approach.

A New Approach to Graph Recognition and Applications to Distance-Hereditary Graphs

Algorithms used in data mining and bioinformatics have to deal with huge amount of data efficiently. In many applications, the data are supposed to have explicit or implicit structures. To develop efficient algorithms for such data, we have to propose possible structure models and test if the models are feasible. Hence, it is important to make a compact model for structured data, and enumerate all instances efficiently. There are few graph classes besides trees that can be used for a model. In this paper, we investigate distance-hereditary graphs. This class of graphs consists of isometric graphs and hence contains trees and cographs. First, a canonical and compact tree representation of the class is proposed. The tree representation can be constructed in linear time by using prefix trees. Usually, prefix trees are used to maintain a set of strings. In our algorithm, the prefix trees are used to maintain the neighborhood of vertices, which is a new approach unlike the lexicographically breadth-first search used in other studies. Based on the canonical tree representation, efficient algorithms for the distance-hereditary graphs are proposed, including linear time algorithms for graph recognition and graph isomorphism and an efficient enumeration algorithm. An efficient coding for the tree representation is also presented; it requires [3.59n] bits for a distance-hereditary graph of n vertices and 3n bits for a cograph. The results of coding improve previously known upper bounds （both are 2^O（nlogn）） of the number of distance-hereditary graphs and cographs to 2^[3.59n] and 2^3n, respectively.

Interactive Inconsistency Fixing in Feature Modeling

Feature models have been widely adopted to reuse the requirements of a set of similar products in a domain. In feature models＇ construction, one basic task is to ensure the consistency of feature models, which often involves detecting and fixing of inconsistencies in feature models. While many approaches have been proposed, most of them focus on detecting inconsistencies rather than fixing inconsistencies. In this paper, we propose a novel dynamic-priority based approach to interactively fixing inconsistencies in feature models, and report an implementation of a system that not only automatically recommends a solution to fixing inconsistencies but also supports domain analysts to gradually reach the desirable solution by dynamically adjusting priorities of constraints. The key technical contribution is, as far as we are aware, the first application of the constraint hierarchy theory to feature modeling, where the degree of domain analysts＇ confidence on constraints is expressed by using priority and inconsistencies are resolved by deleting one or more lower-priority constraints. Two case studies demonstrate the usability and scalability （efficiency） of our new approach.

Spatially adaptive long-term semi-Lagrangian method for accurate velocity advection

We introduce a new advection scheme for fluid animation.Our main contribution is the use of long-term temporal changes in pressure to extend the commonly used semi-Lagrangian scheme further back along the time axis.Our algorithm starts by tracing sample points along a trajectory following the velocity field backwards in time for many steps.During this backtracing process,the pressure gradient along the path is integrated to correct the velocity of the current time step.We show that our method effectively suppresses numerical diffusion,retains small-scale vorticity,and provides better long-term kinetic energy preservation.

Facial-sketch Synthesis:A New Challenge

This paper aims to conduct a comprehensive study on facial-sketch synthesis(FSS).However,due to the high cost of obtaining hand-drawn sketch datasets,there is a lack of a complete benchmark for assessing the development of FSS algorithms over the last decade.We first introduce a high-quality dataset for FSS,named FS2K,which consists of 2104 image-sketch pairs spanning three types of sketch styles,image backgrounds,lighting conditions,skin colors,and facial attributes.FS2K differs from previous FSS datasets in difficulty,diversity,and scalability and should thus facilitate the progress of FSS research.Second,we present the largest-scale FSS investigation by reviewing 89 classic methods,including 25 handcrafted feature-based facial-sketch synthesis approaches,29 general translation methods,and 35 image-to-sketch approaches.In addition,we elaborate comprehensive experiments on the existing 19 cutting-edge models.Third,we present a simple baseline for FSS,named FSGAN.With only two straightforward components,i.e.,facialaware masking and style-vector expansion,our FSGAN surpasses the performance of all previous state-of-the-art models on the proposed FS2K dataset by a large margin.Finally,we conclude with lessons learned over the past years and point out several unsolved challenges.Our code is available at https://github.com/DengPingFan/FSGAN.