Zhiping Cai

Extreme Learning Machines [Trends & Controversies]

This special issue includes eight original works that detail the further developments of ELMs in theories, applications, and hardware implementation. In Representational Learning with ELMs for Big Data, Liyanaarachchi Lekamalage Chamara Kasun, Hongming Zhou, Guang-Bin Huang, and Chi Man Vong propose using the ELM as an auto-encoder for learning feature representations using singular values. In A Secure and Practical Mechanism for Outsourcing ELMs in Cloud Computing, Jiarun Lin, Jianping Yin, Zhiping Cai, Qiang Liu, Kuan Li, and Victor C.M. Leung propose a method for handling large data applications by outsourcing to the cloud that would dramatically reduce ELM training time. In ELM-Guided Memetic Computation for Vehicle Routing, Liang Feng, Yew-Soon Ong, and Meng-Hiot Lim consider the ELM as an engine for automating the encapsulation of knowledge memes from past problem-solving experiences. In ELMVIS: A Nonlinear Visualization Technique Using Random Permutations and ELMs, Anton Akusok, Amaury Lendasse, Rui Nian, and Yoan Miche propose an ELM method for data visualization based on random permutations to map original data and their corresponding visualization points. In Combining ELMs with Random Projections, Paolo Gastaldo, Rodolfo Zunino, Erik Cambria, and Sergio Decherchi analyze the relationships between ELM feature-mapping schemas and the paradigm of random projections. In Reduced ELMs for Causal Relation Extraction from Unstructured Text, Xuefeng Yang and Kezhi Mao propose combining ELMs with neuron selection to optimize the neural network architecture and improve the ELM ensembles computational efficiency. In A System for Signature Verification Based on Horizontal and Vertical Components in Hand Gestures, Beom-Seok Oh, Jehyoung Jeon, Kar-Ann Toh, Andrew Beng Jin Teoh, and Jaihie Kim propose a novel paradigm for hand signature biometry for touchless applications without the need for handheld devices. Finally, in An Adaptive and Iterative Online Sequential ELM-Based Multi-Degree-of-Freedom Gesture Recognition System, Hanchao Yu, Yiqiang Chen, Junfa Liu, and Guang-Bin Huang propose an online sequential ELM-based efficient gesture recognition algorithm for touchless human-machine interaction.

RRED: robust RED algorithm to counter low-rate denial-of-service attacks

The existing Random Early Detection (RED) algorithm and its variants are found vulnerable to emerging attacks, especially the Low-rate Denial-of-Service (LDoS) attacks. In this letter we propose a Robust RED (RRED) algorithm to improve the TCP throughput against LDoS attacks. The basic idea behind the RRED is to detect and filter out attack packets before a normal RED algorithm is applied to incoming flows. We conduct a set of simulations to evaluate the performance of the proposed RRED algorithm. The results show that, compared to existing RED-like algorithms, the RRED algorithm nearly fully preserves the TCP throughput in the presence of LDoS attacks.

Flow level detection and filtering of low-rate DDoS

The recently proposed TCP-targeted Low-rate Distributed Denial-of-Service (LDDoS) attacks send fewer packets to attack legitimate flows by exploiting the vulnerability in TCPs congestion control mechanism. They are difficult to detect while causing severe damage to TCP-based applications. Existing approaches can only detect the presence of an LDDoS attack, but fail to identify LDDoS flows. In this paper, we propose a novel metric - Congestion Participation Rate (CPR) - and a CPR-based approach to detect and filter LDDoS attacks by their intention to congest the network. The major innovation of the CPR-base approach is its ability to identify LDDoS flows. A flow with a CPR higher than a predefined threshold is classified as an LDDoS flow, and consequently all of its packets will be dropped. We analyze the effectiveness of CPR theoretically by quantifying the average CPR difference between normal TCP flows and LDDoS flows and showing that CPR can differentiate them. We conduct ns-2 simulations, test-bed experiments, and Internet traffic trace analysis to validate our analytical results and evaluate the performance of the proposed approach. Experimental results demonstrate that the proposed CPR-based approach is substantially more effective compared to an existing Discrete Fourier Transform (DFT)-based approach - one of the most efficient approaches in detecting LDDoS attacks. We also provide experimental guidance to choose the CPR threshold in practice.

Applying a new localized generalization error model to design neural networks trained with extreme learning machine

High accuracy and low overhead are two key features of a well-designed classifier for different classification scenarios. In this paper, we propose an improved classifier using a single-hidden layer feedforward neural network (SLFN) trained with extreme learning machine. The novel classifier first utilizes principal component analysis to reduce the feature dimension and then selects the optimal architecture of the SLFN based on a new localized generalization error model in the principal component space. Experimental and statistical results on the NSL-KDD data set demonstrate that the proposed classifier can achieve a significant performance improvement compared with previous classifiers.

Multicast Service-Oriented Virtual Network Embedding in Wireless Mesh Networks

Multicast is used in numerous real-time applications which have high demand on quality of service (QoS). When a number of multicast applications are deployed in a wireless mesh network (WMN), network virtualization technology can be used to guarantee the QoS for each application. Since wireless links are unreliable, packet loss is inevitable when the multicast service-oriented virtual networks (VNs) are embedded into a WMN. Although multicast allows the occurrence of packet loss, it is still important to ensure the packet loss rate is below a certain level for QoS guarantee. In this letter, we propose a novel approach, referred to as WELL, to settle the problem of embedding multicast VNs with reliability constraints into a WMN with lossy links. Through opportunistic rebroadcast, WELL minimizes the activation time of the VNs while satisfying their reliability constraints. Simulation results reveal that WELL dramatically outperforms both pure broadcast and unicast based solutions.

FADE: Forwarding Assessment Based Detection of Collaborative Grey Hole Attacks in WMNs

Data security, which is concerned with the confidentiality, integrity and availability of data, is still challenging the application of wireless mesh networks (WMNs). In this paper, we focus on a special type of denial-of-service attack, called selective forwarding or grey hole attack. When this attack is launched at the gateways of a WMN where data tend to aggregate, it could lead to severe damages due to loss of sensitive data. Most existing proposals that focus on detecting stand-alone attackers via channel overhearing are ineffective against collusive attackers. In this paper, we propose a forwarding assessment based detection (FADE) scheme to mitigate collaborative grey hole attacks. Specifically, FADE detects sophisticated attacks by means of forwarding assessments aided by two-hop acknowledgement monitoring. Moreover, FADE can coexist with contemporary link security techniques. We analyze the optimal detection threshold that minimizes the sum of false positive rate and false negative rate of FADE, considering the network dynamics due to degraded channel quality or medium access collisions. Extensive simulation results are presented to demonstrate the adaptability of FADE to network dynamics and its effectiveness in detecting collaborative grey hole attacks.

Security-aware virtual network embedding

Network virtualization is a promising technology to enable multiple architectures to run on a single network. However, virtualization also introduces additional security vulnerabilities that may be exploited by attackers. It is necessary to ensure that the security requirements of virtual networks are met by the physical substrate, which however has not received much attention thus far. This paper represents an early attempt to consider the security issue in virtual network embedding, the process of mapping virtual networks onto physical nodes and links. We model the security demands of virtual networks by proposing a simple taxonomy of abstractions, which is enough to meet the variations of security requirements. Based on the abstraction, we formulate security-aware virtual network embedding as an optimization problem, proposing objective functions and mathematical constraints which involve both resource and security restrictions. Then a heuristic algorithm is developed to solve this problem. Our simulation results indicate its high efficiency and effectiveness.

Using multiple unmanned aerial vehicles to maintain connectivity of MANETs

Unmanned Aerial Vehicles (UAVs) have emerged as promising relay platforms to improve the connectivity of ground Mobile Ad Hoc Networks (MANETs). Due to the relatively high cost of UAVs, a lot of efforts have been made to optimize the deployment of UAVs so that the number of UAVs needed in maintaining the connectivity of ground nodes can be minimized. However, existing work on optimization of UAVs deployment hasnt considered the situation that there are already some UAVs deployed in the field. In this paper, we study the problem of deploying minimum number of UAVs to maintain the connectivity of ground MANETs under the condition that some UAVs have already been deployed in the field. We formulate this problem as a Minimum Steiner Tree problem with Existing Mobile Steiner points under Edge Length Bound constraints (MST-EMSELB) and prove that the problem is NP-Complete. We also propose an Existing UAVs Aware (EUA) polynomial time approximate algorithm for the MST-EMSELB problem that uses a maximum match heuristic to compute new positions for existing UAVs. Simulation results demonstrate that the proposed EUA method has bester performance than a non-EUA method in the term of needed new UAV numbers. Compared with the non-EUA method, the EUA method can reduce at most 60% of the new UAVs number.

ISAR: Improved Situation-Aware Routing Method for Wireless Mesh Backbones

Due to the increasing performance requirements of wireless access services, adaptive interference resistant routing is crucial for providing reliable access and high bandwidth services via wireless mesh networks. In this letter, we propose an improved method for situation-aware routing, which extends the existing routing protocols via a refined situation-aware routing metric and a greedy load-balancing routing scheme. Simulation results demonstrate that the proposed method can achieve a higher throughput with lower transmission delay compared to previous routing schemes.

A Novel Threat Assessment Method for DDoS Early Warning Using Network Vulnerability Analysis

Distributed Denial of Service (DDoS) attack is one of main threats to Internet security. Due to the spatio-temporal properties of the attack, it is possible to detect the attack at its early stage. In this paper, we propose a novel method of DDoS threat assessment based on network vulnerability analysis. Both the multi-phase character in the temporal dimension and the impacts in the spatial dimension are concerned in our method. We use three metrics to assess threat, namely the ratio of progress, botnet size, and bots distribution. Experimental results show that our method is sensitive to the changes of attack states, and is easy to be implemented in an early warning system because of its simplicity.