Federated learning is a privacy-preserving machine learning paradigm to protect the data of clients against privacy breaches. Federated learning algorithms are further reinforced with differential privacy to provide added privacy. Yet, many existing federated learning algorithms are not robust against Byzantine clients. Specifically, in the online federated learning environments, such as in real-time sensing and dynamic systems where data varies with time, coping with Byzantine clients poses a serious challenge. Byzantine clients disrupt convergence by poisoning the local models of non-faulty clients. Hence, it is important to develop an algorithm that is robust against Byzantine clients with the guarantee of convergence to the sequence of global models over time. Thus, this work proposes a robust algorithm based on online mirror descent to guarantee optimal convergence. The regret bound obtained is compared with the Federated Averaging algorithm. The regret bound shows that the proposed algorithm performs well even in the presence of Byzantine clients.

Post-translational modifications (PTMs) in proteins occur after the process of translation. PTMs account for many cellular processes such as deoxyribonucleic acid (DNA) repair, cell signaling and cell death. One of the recent PTMs is succinylation. Succinylation modifies lysine residue from $-1$ to $+1$. Locating succinylation sites using experimental methods, such as mass spectrometry is very laborious. Hence, computational methods are favored using machine learning techniques. This paper proposes a deep learning architecture to predict succinylation sites. The performance of the proposed architecture is compared to the state-of-the-art deep learning architecture and other traditional machine learning techniques for succinylation. It is shown from the performance metrics that the proposed architecture provides a good trade-off between speed of computation and classification accuracy.

The convergence of 5G architecture and deep learning has gained a lot of research interests in both the fields of wireless communication and artificial intelligence. This is because deep learning technologies have been identified to be the potential driver of the 5G technologies, that make up the 5G architecture. Hence, there have been extensive surveys on the convergence of 5G architecture and deep learning. However, most of the existing survey papers mainly focused on how deep learning can converge with a specific 5G technology, thus, not covering the full spectrum of the 5G architecture. Although there is a recent survey paper that appears to be robust, a review of that paper shows that it is not well structured to specifically cover the convergence of deep learning and the 5G technologies. Hence, this paper provides a robust overview of the convergence of the key 5G technologies and deep learning. The challenges faced by such convergence are discussed. In addition, a brief overview of the future 6G architecture, and how it can converge with deep learning is also discussed.

Pressure-sensitive smart floors deployed within homes can give great insight to the health and activity level of individuals through gait and location information. Due to the ever-changing dynamic nature of household deployments involving furniture movement, floor tile shifts, and sensor drift, challenges arise in ensuring the constant reliability of floor sensor readings over time. This paper presents a procedure to automatically calibrate a smart floor’s force sensors without specialized physical effort. The calibration algorithm automatically filters out non-human static weight while retaining weight generated by human activity. This technique is designed to correctly translate sensor values to weight units even when direct access to the force sensors is not available and when a shared tile floor sits above the sensor grid. These calibrated sensor values can then feed machine learning techniques used to extract individual contact points generated by a person’s walking cycle. Using known human weights but no knowledge of the human’s location or walking trajectory, this calibration technique resulted in small percentage differences of -7.8%, -4.8%, and -1.6% for the mean, median, and mode of calibrated smart floor walking sequences, respectively.

This paper discusses how clients in a federated learning system can collaborate with privacy guarantee in a fully decentralized setting without a central server. Most existing work includes a central server that aggregates the local updates from the clients and coordinates the training. Thus, the setting in this existing work is prone to communication and computational bottlenecks, especially when large number of clients are involved. Also, most existing federated learning algorithms do not cater for situations where the data distribution is time-varying such as in real-time traffic monitoring. To address these problems, this paper proposes a differentially-private online mirror descent algorithm. To provide additional privacy to the loss gradients of the clients, local differential privacy is introduced. Simulation results are based on a proposed differentially-private exponential gradient algorithm, which is a variant of differentially-private online mirror descent algorithm with entropic regularizer. The simulation shows that all the clients can converge to the global optimal vector over time. The regret bound of the proposed differentially-private exponential gradient algorithm is compared with the regret bounds of some state-of-the-art online federated learning algorithms found in the literature.

This paper addresses the problem of time-varying data distribution in a fully decentralized federated learning setting with budget constraints. Most existing work cover only fixed data distribution in the centralized setting, which is not applicable when the data becomes time-varying, such as in real-time traffic monitoring. More so, a lot of existing work do not address budget constraint problem common in practical federated learning settings. To address these problems, we propose an online Lagrangian descent algorithm. To provide privacy to the local model updates of the clients, local differential privacy is introduced. We show that our algorithm incurs the best regret bound when compared to other similar algorithms, while satisfying the budget constraints in the long term.

Scientific computing has advanced in the ways it deals with massive amounts of data, since the production capacities have increased significantly for the last decades. Most large science experiments require vast computing and data storage resources in order to provide results or predictions based on the data obtained. For scientific distributed computing systems with hundreds of petabytes of data and thousands of users it is important to keep track not just of how data is distributed in the system, but also of individual users’ interests in the distributed data (reveal implicit interconnection between user and data objects). This however requires the collection and use of specific statistics such as correlations between data distribution, the mechanics of data distribution, and mainly user preferences. This work focuses on user activities (specifically, data usages) and interests in such a distributed computing system, namely PanDA (Production ANd Distributed Analysis system). PanDA is a high-performance workload management system originally designed to meet production and analysis requirements for a data-driven workload at the Large Hadron Collider Computing Grid for the ATLAS Experiment hosted at CERN (the European Organization for Nuclear Research). In this work we are going to investigate whether data collection that was gathered in the past in PanDA shows any trends indicating that users could have mutual interests that would be kept for the next data usages (i.e., data usage patterns), using data mining techniques such as association analysis, sequential pattern mining, and basics of the recommender system approach. We will show that such common interests between users indeed exist and thus could be used to provide recommendations (in terms of the collaborative filtering) to help users with their data selection process.

© 2018 Advances in Science, Technology and Engineering Systems. During 2016, ASTES Journal started to publish articles under the Creative Commons Attribution License and are now using the latest version of the CC BY license, which grants authors the most extensive rights. This means that all articles published in ASTES Journal, including data, graphics, and supplements, can be linked from external sources, scanned by search engines, re-used by text mining applications or websites, blogs, etc. free of charge under the sole condition of proper accreditation of the source and original publisher. Important Note: some articles (especially Reviews) may contain figures, tables or text taken from other publications, for which ASTES Journal does not hold the copyright or the right to re-license the published material. Please note that you should inquire with the original copyright holder (usually the original publisher or authors), whether or not this material can be re-used.

Gene expression is the main process responsible for the function of every living cell. Thousands of genes expressed in a specific cell determine what that cell can do. Gene expression values can be measured by measuring the amount of messenger RNA (mRNA) molecules. There are biological methods to measure gene expression in biological samples so researchers can find genes responsible for each disease. Some example methods are Reporter gene, Microarray, and RNA sequencing. These methods however are very costly and time consuming. Computational methods have the potential to help these studies by identifying reliable directions using prediction techniques on incomplete data; so novel and efficient techniques and algorithms to predict gene expressions are in high demand. In this paper, we describe a method to recover gene expression dataset based on robust principal component analysis (RPCA). We treat the differentially expressed genes as sparse noise S and non-differentially expressed genes as low-rank matrix Y. We show how S and Y can be recovered from gene expression data using RPCA. We also used existing implementations of three other iterative optimization based matrix completion methods to provide a comparative analysis of their performances. We show that this approach consistently outperforms the other methods with reaching improvement factors beyond 7.9 in measured mean squared error.

Finding candidate genes that could cause specific diseases has been the subject of many studies. This is an important research task, however in the biological experimentation domain it can be very expensive and time consuming. So an alternative way is to find gene expression values from partial measurements and try to predict the rest. By using computational methods, we can statistically estimate these relationships faster and in a more efficient way, providing domain experts suggestions on what exploration of likely relationships they should be focusing. One common computational approach is to model the gene expression data as a matrix (where each row represents a gene and each column a subject); the entries of the matrix can then be mRNA measurements that show the extent of gene expressions. Since entries of the dataset are based on partial measurements, the dataset has missing values, and the problem is then to estimate the missing values and thus to recover the global matrix based on the known values. In this paper, we present a correlation based approach to the matrix completion task (CMC) and discuss its functionality. The CMC based algorithm is then compared to a state-of-the-art nuclear-norm minimization iterative algorithm. Our results show that the CMC based algorithm significantly outperforms the iterative algorithm and even shows a better tendency when the amount of missing values grow. We argue that the CMC based algorithm can provide good estimates for missing values, possibly guiding time consuming biological gene expression profiling as to which values should be confirmed. Although our primary focus is on gene expression prediction, the strategy discussed is applicable to any highly correlated dataset where missing values need to be estimated/recovered.

Opportunistic networks (ONs) allow wireless devices to communicate and collaborate in the absence of dedicated networking infrastructure. Recent research work exploits mobility of nodes to create delayed paths in opportunistic networks. Execution of distributed transactions that comprise a sequence of atomic operations across multiple nodes continues to be a challenge for opportunistic networks. Implementing distributed transactions poses new challenges as participation of multiple processes, both local and remote, is required.In this paper, we propose and implement a novel transaction system called DiTON, specifically tailored to operate in opportunistic networks for delay-tolerant applications. To the best of our knowledge, our work is the first attempt to implement a transaction paradigm specifically intended for opportunistic networks. DiTON is a novel transaction scheme that allows multiple nodes to collaborate on shared sets of data while providing global coherency despite network interruptions. Through experimental results, we demonstrate that DiTON can be successfully implemented in opportunistic networks. (C) 2017 Elsevier B.V. All rights reserved.

This paper introduces SmartCare, a project revolving around a smart environment especially built to enable aging in place. The paper describes the vision behind SmartCare as well as its translation into a deployed system. The physical incarnation of SmartCare is the SmartCare apartment, an actual apartment in a retirement community. We provide a description the technologies that are deployed in the SmartCare apartment. This paper does not introduce novel algorithms for smart environments, rather it provides an architectural system description of a real smart environment that was designed to enable aging in place.

The SmartCare project is to design, develop, and evaluate an intelligent sensor-driven living environment for the elderly. The core objectives are to provide emergency detection, improve quality of life, extend independence for the elderly, and detect patterns of behavior that could suggest early signs of a physical or cognitive issue, all in an unobtrusiveness manner. This paper specifically focuses on the development of the infrastructure integration component: PESTO and two of its sub components: a 3D visualization of the resident's smart apartment (VISMA) and a system to provide everyday task assistance through the Z-Wave home automation technology (ZAPS).

Wireless Sensor Networks (WSNs) are being deployed for different applications, each having its own structure, goals and requirements. Medium access control (MAC) protocols play a significant role in WSNs and hence should be tuned to the applications. However, there is no for selecting MAC protocols for different situations. Therefore, it is hard to decide which MAC protocol is good for a given situation. Having a precise model for each MAC protocol, on the other hand, is almost impossible. Using the intuition that the protocols in the same behavioral category perform similarly, our goal in this paper is to introduce a general model that selects the protocol(s) that satisfy the given requirements from the category that performs better for a given context. We define the Combined Performance Function (CPF) to demonstrate the performance of different categories protocols for different contexts. Having the general model, we then discuss the model scalability for adding new protocols, categories, requirements, and performance criteria. Considering energy consumption and delay as the initial performance criteria of the model, we focus on deriving mathematical models for them. The results extracted from CPF are the same as the well-known rule of thumb for the MAC protocols that verifies our model. We validate our models with the help of simulation study. We also implemented the current CPF model in a web page to make the model online and useful.

Opportunistic networks (ONs) exploit mobility of devices to route messages and distribute information. Unfortunately, the fleeting and intermittent connections between pairs of devices make many traditional computer collaboration paradigms, such as distributed shared memory (DSM), very difficult to realize. DSM systems, developed for traditional networks, rely on relatively stable, consistent connections among participating nodes. We propose a novel delay tolerant lazy release consistency (DTLRC) mechanism for implementing distributed shared memory in opportunistic networks. DTLRC permits mobile devices to remain independently productive while separated, and provides a mechanism for nodes to regain coherence of shared memory if and when they meet again. Simulations using experimentally determined data traces demonstrate that DTLRC is a viable concept for enhancing cooperation among mobile wireless devices in opportunistic networking environments.

This is the first study on crowdsourcing Pareto-optimal object finding, which has applications in public opinion collection, group decision making, and information exploration. Departing from prior studies on crowdsourcing skyline and ranking queries, it considers the case where objects do not have explicit attributes and preference relations on objects are strict partial orders. The partial orders are derived by aggregating crowdsourcers' responses to pairwise comparison questions. The goal is to find all Pareto-optimal objects by the fewest possible questions. It employs an iterative question-selection framework. Guided by the principle of eagerly identifying non-Pareto optimal objects, the framework only chooses candidate questions which must satisfy three conditions. This design is both sufficient and efficient, as it is proven to find a short terminal question sequence. The framework is further steered by two ideas—macro-ordering and micro-ordering. By different micro-ordering heuristics, the framework is instantiated into several algorithms with varying power in pruning questions. Experiment results using both real crowdsourcing marketplace and simulations exhibited not only orders of magnitude reductions in questions when compared with a brute-force approach, but also close-to-optimal performance from the most efficient instantiation.

xCHOKe, CHOKe and RED-PD are some of the schemes that have been proposed to control malicious flows in IP networks. Both xCHOKe and CHOKe use CHOKe hits while RED-PD uses RED drops in identifying these flows. In this paper, we are proposing a new scheme called RECHOKe (REpeatedly CHOose and Keep for responsive flows, REpeatedly CHOose and Kill for unresponsive flows) that can be used for detection, control and punishment of these malicious flows. It does this by combining the techniques used by CHOKe, xCHOKe and RED-PD. The difference is, unlike in xCHOKe and CHOKe, RECHOKe does not drop packets during CHOKe hits, thereby eliminating the complexity of dropping or marking randomly selected packets already queued. By using both the CHOKe hit and CHOKe-RED drop histories, we show using ns2 1) that RECHOKe is able to detect, control and punish these flows more accurately, when compared to RED, CHOKe and xCHOKe while providing more protection for TCP-friendly flows, 2) how accurate these histories are, when used in conjunction, in identifying malicious flows and 3) the unreliability of RED, CHOKe and xCHOKe in affecting TCP-friendly flows. Other similar proposed schemes either did not perform as well or incur too much overhead.