Publishing datasets is a key part of data mining and analysis. Handling datasets containing a large number of attributes is a major challenge for analyzing these datasets. Ensuring the privacy of personal and sensitive information is also represent another main challenge. In this paper, we will show how concepts from algebraic topology, and in particular persistent homology can be used for publishing differentially private datasets. We will propose a sampling-based framework to explore the dependencies among all attributes and subsequently build a dependency graph. From the dependency graph, we will regenerate the sub-graphs privately and publish an anonymized, synthetic datasets. Evaluation results demonstrate that our method achieves superior performance when compared to other methods in the literature.

Distributed ledger technologies (DLTs) based on Directed Acyclic Graphs (DAGs) have been gaining much attention due to their performance advantage over the traditional blockchain. IOTA is an example of DAG-based DLT that has shown its significance in the Internet of Things (IoT) environment. Despite that, IOTA is vulnerable to double-spend attacks, which threaten the immutability of the ledger. In this paper, we propose an efficient yet simple method for detecting a parasite chain, which is one form of attempting a double-spend attack in the IOTA network. In our method, a score function measuring the importance of each transaction in the IOTA network is employed. Any abrupt change in the importance of a transaction is reflected in the 1st and 2nd order derivatives of this score function, and therefore used in the calculation of an anomaly score. Due to how the score function is formulated, this anomaly score can be used in the detection of a particular type of parasite chain, characterized by sudden changes in the in-degree of a transaction in the IOTA graph. The experimental results demonstrate that the proposed method is accurate and linearly scalable in the number of edges in the network.

Access to data at large scales expedites the progress of research in medical fields. Nevertheless, accessibility to patients' data faces significant challenges on regulatory, organizational and technical levels. In light of this, we present a novel approach based on the crowdsourcing paradigm to solve this data scarcity problem. Utilizing the infrastructure that blockchain provides, our decentralized platform enables researchers to solicit contributions to their well-defined research study from a large crowd of volunteers. Furthermore, to overcome the challenge of breach of privacy and mutual trust, we employed the cryptographic primitive of Zero-knowledge Argument of Knowledge (zk-SNARK). This not only allows participants to make contributions without exposing their privacy-sensitive health data, but also provides a means for a distributed network of users to verify the validity of the contributions in an efficient manner. Finally, since without an incentive mechanism in place, the crowdsourcing platform would be rendered ineffective, we incorporated smart contracts to ensure a fair reciprocal exchange of data for reward between patients and researchers.

Abstract Label manipulation attacks are a subclass of data poisoning attacks in adversarial machine learning used against different applications, such as malware detection. These types of attacks represent a serious threat to detection systems in environments having high noise rate or uncertainty, such as complex networks and Internet of Thing (IoT). Recent work in the literature has suggested using the K -nearest neighboring algorithm to defend against such attacks. However, such an approach can suffer from low to miss-classification rate accuracy. In this paper, we design an architecture to tackle the Android malware detection problem in IoT systems. We develop an attack mechanism based on silhouette clustering method, modified for mobile Android platforms. We proposed two convolutional neural network-type deep learning algorithms against this Silhouette Clustering-based Label Flipping Attack . We show the effectiveness of these two defense algorithms— label-based semi-supervised defense and clustering-based semi-supervised defense —in correcting labels being attacked. We evaluate the performance of the proposed algorithms by varying the various machine learning parameters on three Android datasets: Drebin, Contagio, and Genome and three types of features: API, intent, and permission. Our evaluation shows that using random forest feature selection and varying ratios of features can result in an improvement of up to 19% accuracy when compared with the state-of-the-art method in the literature.

Solid-State Drives (SSDs) play a crucial role in today's storage systems. They are appended into the Hard-Disk Drives (HDDs) storage systems to improve performance. They provide high IO rate and low latency, which makes them a perfect candidate for analytic-based workloads such as MapReduce. Defining an efficient SSD deployment strategies for MapReduce workloads is a challenging task: SSDs are costly and have limited capacity, the workloads have a big process data size, and the platform has a unique workflow nature. The goal of the work is to establish performance and cost relationship between SSD approaches and MapReduce workloads. In our setup, MapReduce workloads were executed with two SSD approaches of tiering and caching each with two setups: compress and uncompress. Our results showed that by using SSD as a tier, MapReduce workload performs better by up to 66% and increased SSD lifespan by around 20% when comparing with cache approach. We also observed that applying compression on the tier approach enhanced the lifespan by 60% but reduced lifespan of cache tier by 50%.

In this paper, we took a two-stage approach to address privacy issues in blockchain-based applications. Using smart contracts, our approach automatically enforces access policies through encryption and is in contrast to most access control models that solely rely on smart contract policies. These contracts are prone to many security vulnerabilities and in some cases written by developers whose trustworthiness is dubious. A cryptically enforced access control, on the other hand, gives more of the sovereignty promised by the blockchain technology, back to users as showcased in our medical data management framework.

Intrusion Detection Systems (IDSs) can identify the malicious activities and anomalies in networks and present robust protection for these systems. Clustering of attacks plays an important role in defining IDS defense policies. A key challenge in clustering has been finding the optimal value for the number of clusters. In this paper, we propose an automatic clustering algorithm as part of an IDS architecture. This algorithm is based on concepts of coherence and separation. Our automatic clustering algorithms find clusters with the most similarity between the proposed cluster elements and the least similarity with other clusters. The proposed clustering is further optimized by considering two types of objective index functions, and Artificial Bee Colony (ABC), Particle Swarm Optimization (PSO), and Differential Evolution (DE) methods. Comparison of the results obtained with other work in the literature shows improvements in terms of the low average number of evaluations functions, high accuracy, and low computation cost.

The permission system in the SmartThings platform governs how apps access devices. The system was designed to protect devices from third-party apps, by forcing apps to access devices through their capabilities. Design flaws in the system result in apps being over-privileged with unauthorized capabilities. This vulnerability represents serious security challenges to this platform and its users. In this paper, we present an automated tool that can identify over-privilege vulnerability in SmartThings apps. We have identified common patterns, and we have used this knowledge to design our automated over-privilege detection tool. We have evaluated the effectiveness of our tool on 222 official and third-party apps, and we have found that approximately 5.5% of defined devices were misused with 76 identified instances of over-privilege.

In large scale data centers Virtual Machines and Tasks (VMs/tasks) scheduling, VMs allocation, workload predictions and monitoring are a vital concern in any cloud-based data center. In all these fields, clustering algorithms are very useful to a group of workload components that have similar behaviors characteristics. Effective clustering is to select an appropriate clustering technique for a specific application. This choice is very necessary, especially when there are wide choices that provide different results. To address such issue, this paper proposes a novel systematic framework to select the suitable VMs/tasks clustering method in large-scale data centers based on clustering purpose, validation indices and results comparison.

In today's ultra-connected world, the production and consumption of digital data has become immensely huge in volume. Differential privacy is a relatively new approach which attempts to provide strong privacy protection to users' data, while still maintaining outside access to data. However, the comparison of these methods in terms of performance and database suitability remains an open question. In this paper, we will provide a comprehensive comparison of input and output data perturbations for differentially private databases and evaluate the results in terms of accuracy, privacy, efficiency and scalability.

Cloud computing has generated much interest in the research community in recent years for its many advantages, but has also raise security and privacy concerns. The storage and access of confidential documents have been identified as one of the central problems in the area. In particular, many researchers investigated solutions to search over encrypted documents stored on remote cloud servers. While many schemes have been proposed to perform conjunctive keyword search, less attention has been noted on more specialized searching techniques. In this paper, we present a phrase search technique based on Bloom filters that is significantly faster than existing solutions, with similar or better storage and communication cost. Our technique uses a series of n-gram filters to support the functionality. The scheme exhibits a trade-off between storage and false positive rate, and is adaptable to defend against inclusion-relation attacks. A design approach based on an application's target false positive rate is also described.

Compressed sensing (CS) schemes have been used in a wide number of applications in practice. Recently, they have been proposed for use in encryption algorithms because of their properties. In this paper, we present an empirical security analysis of compressed sensing-based encryption. Using a neural network model, we will show that the security of this type of encryption can be compromised. We consider at least three different scenarios in which an attack could occur causing partial information about the plaintext to be revealed without knowledge of the CS secret key.

In recent years, the Internet of Things (IoT) has been an inseparable part of our lives. IoT is typically heterogeneous in nature and requires interconnection with different types of devices or “things”. Being able to secure such a distributed environment is an onerous task. The heterogeneity of IoT, along with other factors, poses a challenge when it comes to securing communication between these devices. In this paper, we propose a novel IoT trust and reputation model that employs distributed probabilistic neural networks (PNNs) to classify trustworthy nodes from malicious ones. Our model tackles the cold start problem in IoT environments by predicting ratings for newly joined devices based on their characteristics and learns over time. Processing is completely distributed and is handled by the nodes themselves. This guarantees better availability, since there is no single point of failure. Moreover, our model can accommodate the various capabilities and types of IoT devices. Unlike other proposed models in the literature, our model provides different levels of security depending on the sensitivity of the data being transmitted.

Data generated and stored worldwide is increasing multifold every year, with images and media content accounting for a large portion of this data. In addition to volume, ensuring adequate security is an important challenge that needs to be addressed. In this paper, we propose an efficient, secure data-storage approach based on compressive sensing. Our approach uses data deduplication to remove identical copies of data. Our experimental results show significant storage savings, while providing strong level security.

In a cloud-based service selection system, for a given request, there could be a large number of software services matching the functional requirements. The selection should then be done based on their QoS values. Since in a cloud environment, a software service might need collaboration from other types of cloud services (e.g., A software service delivered through an infrastructure service) to offer a complete solution to an end user, the selection system should have a way to measure the QoS values of the whole solution, instead of QoS of software services alone. This kind of end-to-end QoS values of cloud-based software solutions may or may not be available in recorded history logs. In this paper, we propose a model for predicting end-to-end QoS values of cloud-based software solutions composed of services from multiple cloud layers. It relies on the internal features of services and end users such as locations, configurations, functionality, and user profiles to calculate service similarity and then predict QoS values. The experiments demonstrate the accuracy of our approach. We also studied the impact of the proposed internal features on QoS prediction accuracy.

There have been increasing interest in the area of privacy-protected searching as industries continue to adopt cloud technologies. Much of the recent efforts have been towards incorporating more advanced searching techniques. Although many have proposed solutions for conjunctive keyword search, it is only recently that researchers began exploring phrase search over encrypted data. In this paper, we present a scheme that incorporates both functionalities. Our solution makes use of symmetric encryption, which provides computational and storage efficiency over schemes based on public key encryption. By considering the statistical properties of natural languages, we were able to design indexes that significantly reduce storage cost when compared to existing solutions. Our solution allows for simple ranking of results and requires a low storage cost while providing document and keyword security. By using both the index and the encrypted documents to performs searches, our scheme is also currently the only phrase search scheme capable of searching for non-indexed keywords.

Data centre prediction models can be used to forecast future loads for a given centre in terms of CPU, memory, VM requests, and other parameters. An effective and efficient model can not only be used to optimize resource allocation, but can also be used as part of a strategy to conserve energy, improve performance and increase profits for both clients and service providers. In this paper, we have developed a prediction model, which combines k-means clustering techniques and Extreme Learning Machines (ELMs). We have shown the effectiveness of our proposed model by using it to estimate future VM requests in a data centre based on its historical usage. We have tested our model on real Google traces that feature over 25 million tasks collected over a 29-day time period. Experimental results presented show that our proposed system outperforms other models reported in the literature.

In the cloud computing paradigm, Infrastructure-as-a-Service (IaaS) providers can provision virtualized hardware and resources to users, removing the need for users to own and operate these resources, which can lead to lower costs and improved performance. This paper gives a general description of most commonly used open source IaaS service platforms. It includes descriptions and comparisons of OpenNebula, Eucalyptus, Nimbus, OpenStack and CloudStack platforms, and it should be accessible to a wide audience.

This paper presents an Extreme Learning Machine (ELM) time series prediction strategy to estimate the current and voltage behaviour of an Electric Arc Furnace (EAF). The proposed ELM predictor is designed for both long and short term predictions of the v-i characteristics of an EAF. The proposed predictor is evaluated using two real sensors' outputs collected over different time periods with a rate of 2000 samples per second, and its performance is compared against Feed-Forward Neural Networks (FFNN), Radial Basis Functions (RBF) and Adaptive Neuro-Fuzzy Inference Systems (ANFIS) algorithms. Experimental results obtained show the proposed ELM predictor to have superior speed and stability behaviour, while obtaining similar error values to comparable techniques.

Despite the many benefits of cloud technologies, there have also been significant concerns regarding its security and privacy. To address the issues, much efforts have been made towards development of an encrypted cloud system. One of the key features being investigated is the ability to search over encrypted data. Although many have proposed solutions for conjunctive keyword search, few have considered phrase searching techniques over encrypted data. Due to the increased amount of information required to identify phrases, existing phrase search algorithms require significantly more storage than conjunctive keyword search schemes. In this paper, we propose a phrase search scheme, which takes advantage of the space efficiency of Bloom filters, for applications requiring a low storage cost. It makes use of symmetric encryption, which provides computational and storage efficiency over schemes based on public key encryption. The scheme provides simple ranking capability, can be adapted to non-keyword search and is suitable against inclusion-relation attack.

The major challenge in current online social networks (OSNs) is privacy violation by OSN providers or unauthorised users. OSN providers collect unprecedented amounts of personal information for targeted advertising. Moreover, users are not able to share their social data with their friends with complete access control. Peer‐to‐peer (P2P) infrastructure is an interesting solution for a big‐brother‐free alternative to current OSN designs. However, the fundamental nature of P2P systems has dynamic peer turn‐over which results in data unavailability. Additionally, users’ data must be available in the OSN when authorised data audiences want to access them. For these reasons, we propose a P2P‐OSN architecture which is composed of a privacy enabled setup for users’ social communications and an adaptive replica placement strategy for ensuring availability for users’ shared data. The proposed framework correlates the availability of shared content in the P2P‐OSN to the access control assigned to them. Our evaluations show the proposed P2P‐OSN has considerable improvements in providing data privacy and availability compared with the existing approaches.

On behalf of the Organizing Committee, I would like to welcome you to the Twelfth Annual Conference on Privacy, Security, and Trust (PST 2014) held at Ryerson University, Toronto, Canada. The objective of the conference is to provide a forum for researchers worldwide to unveil their latest work in privacy, security and trust and to show how this research can be used to enable innovation. Topics covered this year include Privacy for Mobile and Social Applications, Privacy-Preserving Cryptography and Data Analysis, Business security and Security software Evaluation, Web and Mobile Application Security, Malware and Malicious Detection, Encryption and Secret Sharing, Trust Models, and Reputation Systems. This year's conference received 161 submissions from 46 countries, out of which 47 papers were selected for regular oral presentation, and 7 for short presentations. All submissions went through a careful anonymous review process (3 or more reviews per submission) aided by 89 Technical Program Committee members and 55 external reviewers. This year's program includes two keynote addresses by Dr. Ann Cavoukian (Information and Privacy Commissioner, Ontario, Canada) and Dr. Rei Safavi-Naini (AITF Strategic Chair in Information Security at the University of Calgary, Canada). I would like to thank everyone who has given his or her time, energy and ideas to assist in organizing this event, including all the members of the organizing committee, the TPC Co-Chairs, TPC members and all the reviewers, and our two distinguished keynote speakers, Dr. Cavoukian and Dr. Safavi-Naini who have agreed to address the conference attendees. In particular, I would like to highlight and acknowledge the tremendous efforts of Dr. Joaquin Garcia-Alfaro (Publicity Co-chair, and Publication Chair) who worked tirelessly on various conference-related tasks. I also wish to thank all of our sponsors who have made this event possible. It is through the collective efforts of these individuals and organizations that we are able to bring you a great event!

With the advent of cloud computing as a new paradigm and technology, and the increased tendency of decision makers to envision a staged migration to cloud services, most enterprises are choosing to outsource their data to cloud storage providers, for better management of their IT resources, in terms of security, control, space and storage costs. In this context, assuming that the cloud service provider may not be trustworthy (i.e. is honest but curious), ensuring data privacy in all operations performed on enterprise data while these data reside in the Cloud is still a challenge. This paper proposes a novel twolevel data deduplication framework that can be used in cloud storage by enterprises. At the enterprise level, the enterprise performs cross-user data deduplication and outsources its data to the Cloud. At the cloud storage provider level, cross enterprise data deduplication is performed by the cloud service provider to further remove duplicates, resulting in cost and space savings. We argue that our framework will allow the enterprise to facilitate operations such as searching over encrypted data, sharing data within the enterprise, and downloading data from the Cloud directly, in a secure and efficient manner without the need to trust the cloud service provider.

In order to select and rank the best services in a cloud computing environment, the end-to-end quality of service (QoS) values of cloud services have to be computed. For a new SaaS provider, the deployment of its software application in the cloud is a challenging job. It has to find a hosting service (IaaS) that hosts its service. The primary goal of the SaaS provider is to make its service at the top of the ranked list of cloud services returned to end users through satisfying their QoS requirements. In this paper, we propose a mechanism to map the users' QoS requirements of cloud services to the right QoS specifications of SaaS then map them to best IaaS service that offers the optimal QoS guarantees. Then together SaaS and IaaS services can provide the best service offer to end users. As a result of the mapping, the end-to-end QoS values can be calculated. We propose a set of rules to perform the mapping process. We hierarchically model the QoS specifications of cloud services using the Analytic Hierarchy Process (AHP) method. The AHP based model helps to facilitate the mapping process across the cloud layers, and to rank the candidate cloud services for end users. We use a case study to illustrate and validate our solution approach.

ABSTRACT The relay attack is a simple yet effective attack against most radio‐frequency identification (RFID) authentication systems. Because of the minimalist design of an RFID system, a lightweight authentication scheme must be designed to provide a strong level of security for low‐cost RFID tags. In this work, we propose a simple but secure masking scheme that counters the relay attack in RFID systems. Our scheme is the first solution based on the chaos suppression theory. We have exploited the chaotic characteristics of a dynamic Lorenz controller to distinguish a legitimate RFID reader from a proxy reader in the relay attack. We also show that the proposed approach is practical using simulation results. Copyright © 2012 John Wiley & Sons, Ltd.

ABSTRACT In recent years, online social networks (OSNs) have had explosive growth in numbers and popularity. In an OSN, users communicate with each other and share information about themselves. However, limiting the flow of private information across OSNs is very important especially because most OSNs provide insufficient privacy settings to control information leakage. In this paper, we propose a mediated architecture for OSNs that protects users' information from both the OSN provider and unauthorized OSN users. Our proposed approach delegates most of the computation tasks to a semi‐trusted proxy server. We exploit a simplified broadcast encryption method in order to design a dynamic, efficient, flexible, and fine‐grained (DEFF) control system. In the proposed DEFF system, users are allowed to cryptographically categorize their friends into different relations and to share data with arbitrary groups of them. The results of our analysis indicate that the DEFF system fully protects users' privacy and is very efficient in terms of communication and computation complexities. Copyright © 2012 John Wiley & Sons, Ltd.

This paper presents a novel chaos-based algorithm for image encryption. A 2D chaotic map is used to shuffle the image pixel positions. Substitution (confusion) and permutation (diffusion) operations on every block are combined using two perturbed chaotic PWLCM maps in multiple rounds. Our proposed algorithm uses a new chaotic substitution method based on DNA coding and the complementary rule. Experimental results and comparisons to existing image encryption algorithms are also included that show the high level of security that is obtained.

A lightweight mutual authentication protocol is proposed for RFID systems in which both the tags and the reader can be authenticated to each other. The proposed protocol is based on the McEliece cryptosystem without requiring Radio Frequency Identification (RFID) tags to store the large matrices needed in the McEliece cryptosystem. Complex computational operations in the McEliece cryptosystem are removed from the RFID tags, as they only perform simple binary operations on short vectors. The size of the memory needed in the RFID tag is trivial and suitable for low-cost tags. Readers perform most of the encryption and decryption involved in the authentication protocol using McEliece functions. After every authentication, the content of the RFID tag is securely refreshed making it ready for a new round of authentication. This will ensure that the tags cannot be traced by unauthorized readers, thereby protecting the privacy of the RFID tags.

Cloud computing has empowered the individual user by providing seemingly unlimited storage space and availability and accessibility of data anytime and anywhere. Cloud service providers are able to maximize data storage space by incorporating data deduplication into cloud storage. Although data deduplication removes data redundancy and data replication, it also introduces major data privacy and security issues for the user. In this paper, a new privacy-preserving framework that addresses this issue is proposed. Our framework uses an efficient deduplication algorithm to divide a given file into smaller units. These units are then encrypted by the user using the combination of a secure hash function and a block encryption algorithm. An index tree of hash values of these units is also generated and encrypted using an asymmetric search encryption scheme by the user. This index tree will enable the cloud service provider to search through the index and return the requested units. We will show that our proposed framework will allow cloud service and storage providers to employ data deduplication techniques without giving them access to either the users' plaintexts or the users' decryption keys.

The performance of RFID MAC algorithms is expressed in terms of the total time it takes by a reader to read all the tags in the reading range. One of the major MAC approaches for RFID systems is the framed slotted ALOHA based approach. It is proven that this class of algorithms achieve the smallest total reading time when, at each reading round, the frame length is set equal to the actual number of remaining unread tags. An important design challenge to increase the performance of this class of algorithms is to estimate the remaining tag count before each reading round. In this paper, we address this remaining tag count estimation problem for such RFID systems. The introduced algorithm is an a posteriori tag count estimation scheme from the collision statistics of the previous reading round. We compare the algorithm with the optimum, lower bound and other a posteriori estimation algorithms and demonstrate its efficiency. The performance of the algorithm is better than the current a posteriori tag count estimation algorithms. Another improvement brought by the algorithm is the reduction in reader side computational overhead existing in the previous schemes.

We propose a new identity-based authenticated group key agreement protocol that provides both perfect forward secrecy and key integrity, while offering low communication and computational complexity. We prove the security of our protocol under the decisional bilinear Diffie-Hellman assumption in the random oracle model. In addition, we propose efficient auxiliary key agreement protocols that maintain perfect forward secrecy, while allowing members to discard their ephemeral keys in the same session in which they were created.

Using Boolean AND and OR functions to combine the responses of multiple one- or two-class classifiers in the ROC space may significantly improve performance of a detection system over a single best classifier. However, techniques found in literature assume that the classifiers are conditionally independent, and that their ROC curves are convex. These assumptions are not valid in most real-world applications, where classifiers are designed using limited and imbalanced training data. A new Iterative Boolean Combination (IBC) technique applies all Boolean functions to combine the ROC curves produced by multiple classifiers without prior assumptions, and its time complexity is linear according to the number of classifiers. The results of computer simulations conducted on synthetic and real-world host-based intrusion detection data indicate that combining the responses from multiple HMMs with IBC can achieve a significantly higher level of performance than with the AND and OR combinations, especially when training data is limited and imbalanced.

An efficient authentication mechanism is proposed for RFID tags, based on error-correcting codes. The proposed scheme is an asymmetric design benefiting from advantages of the public-key infrastructure. The security of our protocol is guaranteed based on the difficulty of the syndrome decoding problem. Our protocol is rearranged in a novel way to meet the practical requirements of RFID tags. Complex computational operations are removed from the RFID tags, and they only perform simple binary operations on short vectors. The proposed protocol is considered zero-knowledge, as tags can be easily authenticated by readers without sharing the tag's secret with the readers.

On behalf of the Organizing Committee, we would like welcome you to the Sixth Annual IEEE International Conference on Wireless and Mobile Computing, Networking and Communications (WiMob). The tremendous advances in wireless communications and mobile computing, combined with the rapid evolution in smart appliances and devices have generated new challenges and problems requiring solutions that rely on interactions between different network layers and applications in order to offer advanced mobile services. WiMob'2010 addresses three main areas: Wireless Communications, Mobile Networking, Ubiquitous Computing and Applications. IEEE WiMob has grown to become one of the major events in these areas, and it aims to stimulate interactions among participants and enable them to exchange new ideas and practical experiences in their respective areas.

Hidden Markov Models (HMMs) have been shown to provide a high level performance for detecting anomalies in intrusion detection systems. Since incomplete training data is always employed in practice, and environments being monitored are susceptible to changes, a system for anomaly detection should update its HMM parameters in response to new training data from the environment. Several techniques have been proposed in literature for on-line learning of HMM parameters. However, the theoretical convergence of these algorithms is based on an infinite stream of data for optimal performances. When learning sequences with a finite length, on-line incremental versions of these algorithms can improve discrimination by allowing for convergence over several training iterations. In this paper, the performance of these techniques is compared for learning new sequences of training data in host-based intrusion detection. The discrimination of HMMs trained with different techniques is assessed from data corresponding to sequences of system calls to the operating system kernel. In addition, the resource requirements are assessed through an analysis of time and memory complexity. Results suggest that the techniques for online incremental learning of HMM parameters can provide a higher level of discrimination than those for on-line learning, yet require significantly fewer resources than with batch training. On-line incremental learning techniques may provide a promising solution for adaptive intrusion detection systems.

Association rule mining provides useful knowledge from raw data in different applications such as health, insurance, marketing and business systems. However, many real world applications are distributed among two or more parties, each of which wants to keep its sensitive information private, while they collaboratively gaining some knowledge from their data. Therefore, secure and distributed solutions are needed that do not have a central or third party accessing the parties' original data. In this paper, we present a new protocol for privacy-preserving association rule mining to overcome the security flaws in existing solutions, with better performance, when data is vertically partitioned among two or more parties. Two sub-protocols for secure binary dot product and cardinality of set intersection for binary vectors are also designed which are used in the main protocols as building blocks.

In host-based intrusion detection systems (HIDS), anomaly detection involves monitoring for significant deviations from normal system behavior. Hidden Markov Models (HMMs) have been shown to provide a high level performance for detecting anomalies in sequences of system calls to the operating system kernel. Although the number of hidden states is a critical parameter for HMM performance, it is often chosen heuristically or empirically, by selecting the single value that provides the best performance on training data. However, this single best HMM does not typically provide a high level of performance over the entire detection space. This paper presents a multiple-HMMs approach, where each HMM is trained using a different number of hidden states, and where HMM responses are combined in the receiver operating characteristics (ROC) space according to the maximum realizable ROC (MRROC) technique. The performance of this approach is compared favorably to that of a single best HMM and to a traditional sequence matching technique called STIDE, using different synthetic HIDS data sets. Results indicate that this approach provides a higher level of performance over a wide range of training set sizes with various alphabet sizes and irregularity indices, and different anomaly sizes, without a significant computational and storage overhead.

This paper introduces a new medium access control (MAC) protocol for passive Radio Frequency Identification (RFID) systems. The protocol is designed as an enhancement to framed slotted ALOHA MAC protocols in which tags randomly select a slot number on a given frame size. As shown in this paper, the completely random slot selection in the framed slotted ALOHA systems is not the optimum approach to the slot selection problem. To minimize the collision probability, our protocol, named Distance Based Slot Selection (DiSEL), uses a cross- layer approach for tags to select the most appropriate time slot in a given frame. A tag in DiSEL uses the maximum and minimum received power levels of the reader-tag communications to choose a slot number. A resonant boosting network to increase the received RF power granularity and an efficient rectifier to convert the RF signal into DC introduced for the power level measurements at the tags. We test DiSEL under various tag deployment and density scenarios and show that DiSEL decreases the tag collision probability in both random uniform and evenly spaced dense tag deployments.

A full diversity constellation, that is, a set of unitary matrices whose differences have nonzero determinant, is a design criterion for codes with good performance using differential unitary space-time modulation. Fixed-point free groups and the infinite groupSU(2) have been used to produce full diversity unitary group constellations. In this paper, we present a new Bruhat decomposition design for constructing full diversity unitary space-time constellations for any number of antennas. They are constructed from cosets of a unitary diagonal subgroupD, and our design has a particularly simple structure for the case where the number of transmitter antennas is prime. We also consider the extension of these constellation designs for an even number of transmitter antennas by replacingDwith a Hamiltonian constellation. Some examples of proposed constellations for two to six transmitter antennas are given. Simulations show that our proposed constellations perform well in unknown Rayleigh fading channel.

The ID3 algorithm is a standard, popular, and simple method for data classification and decision tree creation. Since privacy-preserving data mining should be taken into consideration, several secure multi-party computation protocols have been presented based on this technique. Entropy and Gini Index are two protocols which compute information-gain at each step when producing a decision tree. The Gini index, however, has been less studied in privacy-preserving data mining protocols. In this paper, we show how Gini can be used in privacy-preserving ID3 algorithms to create decision tree classifications in such a way that involved parties can jointly compute the gain value of each normal attribute without revealing their own private information to each other, while the database is horizontally partitioned over two or more parties. Three secure multiparty sub-protocols are presented to evaluate the intermediate computations. The communication overhead has been kept reasonably low to make the whole protocol efficient and practical.

We present an innovative methodology for accelerating the elliptic curve point formulas over prime fields. This flexible technique uses the substitution of multiplication with squaring and other cheaper operations by exploiting the fact that field squaring is generally less costly than multiplication. Applying this substitution to the traditional formulas, we obtain faster point operations in unprotected sequential implementations. We also show the significant impact our methodology has in protecting against simple side- channel (SSCA) attacks. We modify the elliptic curve cryptography (ECC) point formulas to achieve a faster atomic structure when applying side-channel atomicity protection. In contrast to previous atomic operations that assume that squarings are indistinguishable from multiplications, our new atomic structure offers true SSCA-protection because it includes squaring in its formulation. Moreover, we extend our implementation to parallel architectures such as Single-Instruction Multiple-Data (SIMD). With the introduction of a new coordinate system and the flexibility of our methodology, we present, to our knowledge, the fastest formulas for SIMD-based schemes that are capable of executing three and four operations simultaneously. Finally, a new parallel SSCA-protected scheme is proposed for multiprocessor/parallel architectures by applying the atomic structure presented in this work. Our parallel and atomic operations are shown to be significantly faster than previous implementations.

This correspondence presents a construction of affine reflection group codes. The solution to the initial vector and nearest distance problem is presented for all irreducible affine reflection groups of rank n ges 2, for varying stabilizer subgroups. We use a detailed analysis of the geometry of affine reflection groups to produce a decoding algorithm which is equivalent to the maximum-likelihood decoder, yet whose complexity depends only on the dimension of the vector space containing the codewords, and not on the number of codewords. We give several examples of the decoding algorithm, both to demonstrate its correctness and to show how, in small rank cases, it may be further streamlined by exploiting additional symmetries of the group.

Diversity sum, which is calculated from the Frobenius norm of the difference of two distinct elements in a signal constellation, is the significant parameter to predict a unitary space–time constellation having good performance in low signal-to-noise ratio (SNR). In this correspondence, we propose a method to compute the diversity sum of a unitary group constellation using a character table. Our proposed analysis is simple, requiring only a lookup of the character table. We illustrate our method for the finite special linear groups$SL_{2}$, and compare codes with high diversity sum against fixed point free groups at low SNR. We also introduce the notion of a faithful group constellation, that is, one whose diversity sum is greater than$0$. Faithful group constellations may be obtained from any nontrivial character of a group. We describe the method to do so in this correspondence and illustrate it with the example of the finite projective special linear group${\rm PSL}_{2}$.

In this paper, distributed arithmetic (DA) has been used to implement a fully parallel LUT-based DA wavelet filterbank with interlaced input registers. In our scheme, decimation has been seamlessly integrated into the filter structure to achieve the same throughput performance as polyphase-based filterbanks. However, because partitioning of the filters is avoided, our scheme gives more flexibility to implement the LUT-DA structure, and consequently, lets designers maximize area utilization on LUT-based FPGAs. Our architecture has been designed for orthonormal and biorthogonal wavelets, and implemented on an Altera Stratix II FPGA. Significant reduction in terms of area requirements and increased throughput performance are achieved when compared to other DWT filterbanks based on DA, convolution or the lifting scheme.

There are scenarios where using production databases for testing are unavoidable. In a time/mission-critical situation where a developer is required to fix a bug immediately the only option is use production database for testing. However, this may pose a violation of privacy. In this paper, we describe a different approach utilizing a privacy filter that examines queries from an application to match with predefined privacy policy to decide the result return. The approach is illustrated using a prototype which was implemented with query modification and data randomization techniques.

A closed-form upper bound is presented for the average pairwise error probabilities (PEP) of space–time codes for a keyhole channel. It is derived from the exact conditional PEP for given fading channel coefficients using a moment generating function-based approach. Simulation results are included for varying numbers of antennas that affirm that the proposed PEP serves as a tight bound for codes in a keyhole channel.

In this correspondence, we propose new unitary space-time constellation designs with high diversity products. Our Hamiltonian and product constellations are based on Slepian's group codes, and can be used for any number of antennas and any data rate. Our Hamiltonian constellations achieve the theoretical upper bound of diversity product when the number of transmitter antennas is even and the cardinality of the signal constellation is less than or equal to 5. Many of our Hamiltonian and product constellations outperform, and have higher diversity products than, the best known designs in the literature. These include orthogonal designs, dicyclic groups, cyclic groups, parametric codes, numerical approaches, nongroup designs, Cayley codes, TAST codes and some constellations obtained from fixed-point free groups.

In a secure dot-product protocol, two parties jointly compute the dot-product of two vectors of their inputs. In our protocol, the result of a secure dot-product is known only to one party while each input in known only to its owner. This protocol is the first non-interactive secure dot-product protocol that only requires one round of interaction with O(n) communication overhead, where n is the size of the input vectors. The computation overhead of this scheme is negligible as it requires few integer multiplications and additions. In our construction, the sender and the receiver are information theoretically secure over sets E and S* respectively, where E and S* are subsets of the vector space. In our protocol, the security of the sender and the receiver can be increased by choosing large fields

In this paper, we propose new full diversity unitary space-time codes based on Hamiltonian constellation designs. Our proposed constellations can be used for any number of antennas and for any data rate. For two transmitter antennas, the constellations are constructed from cyclic group codes. For a larger number of transmitter antennas, the design employs the direct sum of 2 times 2 Hamiltonian matrices and roots of unity. We give some examples of proposed constellations, and also show that they outperform known design techniques in the literature

Oblivious transfer is such a strong cryptographic tool that it is directly used both in private data retrieval techniques and secure, multi-party computation paradigms. We propose an optimal oblivious transfer, where the communication complexity is logarithmic in size of the database with minimum computation overhead - the computational complexity in any oblivious transfer protocol is at least linear in the size of the database. Our communication-efficient oblivious transfer protocol is a non-interactive, single-database scheme that is generally based on homomorphic encryption functions.

In this paper we introduce a detection based security scheme for wireless sensor networks. Although sensor nodes have low computation and communication capabilities, they have specific properties such as their stable neighborhood information that allows for detection of anomalies in networking and transceiver behaviors of the neighboring nodes. We show that such characteristics can be exploited as key enablers for providing security to large scale sensor networks. In many attacks against sensor networks, the first step for an attacker is to establish itself as a legitimate node within the network. To make a sensor node capable of detecting an intruder a simple dynamic statistical model of the neighboring nodes is built in conjunction with a low-complexity detection algorithm by monitoring received packet power levels and arrival rates.

Efficient design of suitable public key cryptographic algorithms is one of the most important problems facing their use in communication systems. An emerging public key cryptosystem that promises to be extremely useful for devices built on embedded systems which are resource constrained in memory, space and processing power is that of hyperelliptic curve cryptosystem (HECC). This work outlines an FPGA implementation of a HEC coprocessor which is based on projective and mixed coordinate representations of the curves. A transformation of variables of the curves is also suggested to improve the operating time. Numerical results are provided that shows the improvements offered by these implementations in terms of space and operation time.

Full diversity signal constellations for any numbers of transmitter antennas and for any orders which are constructed from 2/spl times/2 Hamiltonian matrices are investigated in this paper. The diversity product of a 2/spl times/2 Hamiltonian constellation equals one half of the Euclidean distance between two points in C/sup 2/. By considering the transformation from R/sup 4/ to C/sup 2/, the idea of group codes is used to construct a high diversity product constellation for any order L. The (L,4) cyclic group codes are considered to obtain L 4-dimensional codewords for group codes. We show that our 2/spl times/2 Hamiltonian constellations have higher diversity product than orthogonal and diagonal constellation designs. We extend our construction to the general case for any numbers of transmitter antennas M>2 by using a direct sum of 2/spl times/2 Hamiltonian matrices for M even, and a direct sum of 2/spl times/2 Hamiltonian matrices with the L/sup th/ roots of unity for M odd. It is shown that these constellations outperform cyclic groups and some of those obtained using fixed-point free groups.

This paper proposes new unitary space-time constellation designs with high diversity products for any number of antennas and any rate based on Slepian's group codes. Many of our Hamiltonian and product constellations have the best known diversity products in the literature, and outperform all other constellation designs.

In this paper, we present a new design of constructing full diversity unitary constellations using differential space-time modulation for any number of transmitter and receiver antennas. These constellations are constructed from a Bruhat decomposition which allows us to choose disjoint nonoverlapping cosets of a unitary diagonal subgroup in such a way that full diversity is preserved. Simulations show that our Bruhat constellations perform well in unknown Rayleigh fading channel.

The subset difference (SD) method proposed by D. Naor et al. is one of the most efficient group key distribution techniques. Recently a polynomial based solution for key distribution was proposed by D. Liu et al., which requires a similar message size as the SD method, but also provides self-healing feature. We propose a self-healing feature for the SD method, and present some optimization techniques to reduce the overhead caused by the self-healing capability.