Search

The ubiquitous adoption of Internet-of-Things (IoT) based applications has resulted in the emergence of the Fog computing paradigm, which allows seamlessly harnessing both mobile-edge and cloud resources. Efficient scheduling of application tasks in such environments is challenging due to constrained resource capabilities, mobility factors in IoT, resource heterogeneity, network hierarchy, and stochastic behaviors. xisting heuristics and Reinforcement Learning based approaches lack generalizability and quick adaptability, thus failing to tackle this problem optimally. They are also unable to utilize the temporal workload patterns and are suitable only for centralized setups. However, Asynchronous-Advantage-Actor-Critic (A3C) learning is known to quickly adapt to dynamic scenarios with less data and Residual Recurrent Neural Network (R2N2) to quickly update model parameters. Thus, we propose an A3C based real-time scheduler for stochastic Edge-Cloud environments allowing decentralized learning, concurrently across multiple agents. We use the R2N2 architecture to capture a large number of host and task parameters together with temporal patterns to provide efficient scheduling decisions. The proposed model is adaptive and able to tune different hyper-parameters based on the application requirements. We explicate our choice of hyper-parameters through sensitivity analysis. The experiments conducted on real-world data set show a significant improvement in terms of energy consumption, response time, Service-Level-Agreement and running cost by 14.4%, 7.74%, 31.9%, and 4.64%, respectively when compared to the state-of-the-art algorithms., Comment: Accepted in IEEE Transaction on Mobile Computing

Scalable data processing platforms built on cloud computing becomes increasingly attractive as infrastructure for supporting big data applications. But privacy concerns are one of the major obstacles to making use of public cloud platforms. Multidimensional anonymisation, a global-recoding generalisation scheme for privacy-preserving data publishing, has been a recent focus due to its capability of balancing data obfuscation and usability. Existing multidimensional anonymisation methods suffer from scalability problems when handling big data due to the impractical serial I/O cost. Given the recursive feature of multidimensional anonymisation, parallelisation is an ideal solution to scalability issues. However, it is still a challenge to use existing distributed and parallel paradigms directly for recursive computation. In this paper, we propose a scalable approach for big data multidimensional anonymisation based on MapReduce, a state-of-the-art data processing paradigm. Our basic idea is to partition a data set recursively into smaller partitions using MapReduce until all partitions can fit in the memory of a computing node. A tree indexing structure is proposed to achieve recursive computation. Moreover, we show the applicability of our approach to differential privacy. Experimental results on real-life data demonstrate that our approach can significantly improve the scalability of multidimensional anonymisation over existing methods.

Today's vehicles are advancing from stand-alone transportation means to vehicle-to-vehicle, and vehicle-to-infrastructure communications enabled devices which are able to exchange data through the transportation communication infrastructure. As the IoT and data remain intrinsically linked together, the fast-changing mobility landscape of intent-based networking for the Internet of connected vehicles comes with a great risk of data security and privacy violations. This paper considers the privacy issues in the distributed edge computing, in which the data is communicated between a number of vehicles in the IoT layer and potentially untrusted edge controllers at the edge of the network. The sensory data communicated by the vehicles contain sensitive information, such as location and speed, which could violate the users' privacy if they are leaked with no perturbation. Recent studies suggest mechanisms for randomizing the stream of data to ensure individuals' privacy. Although the past works on differential privacy provide a strong privacy guarantee, they are limited to applications where communication parties are trusted and/or there is no correlation between the users or the featured of sensory data. In this paper, we address this gap by proposing a differentially private data streaming system that adds a correlated noise in the vehicle's side (IoT layer) rather than the transportation infrastructure. Also, our system is able to ensure a strong privacy level over time. The proposed mechanism is data-adaptive and scales the noise with respect to the data correlation. Our extensive experiments demonstrate that the utility of the output generated by our method outperforms the recent approaches.

The virtualization concept and elasticity feature of cloud computing enable users to request resources on-demand and in the pay-as-you-go model. However, the high flexibility of the model makes the on-time resource scaling problem more complex. A variety of techniques such as threshold-based rules, time series analysis, or control theory are utilized to increase the efficiency of dynamic scaling of resources. However, the inherent dynamicity of cloud-hosted applications requires autonomic and adaptable systems that learn from the environment in real-time. Reinforcement Learning (RL) is a paradigm that requires some agents to monitor the surroundings and regularly perform an action based on the observed states. RL has a weakness to handle high dimensional state space problems. Deep-RL models are a recent breakthrough for modeling and learning in complex state space problems. In this article, we propose a Hybrid Anomaly-aware Deep Reinforcement Learning-based Resource Scaling (ADRL) for dynamic scaling of resources in the cloud. ADRL takes advantage of anomaly detection techniques to increase the stability of decision-makers by triggering actions in response to the identified anomalous states in the system. Two levels of global and local decision-makers are introduced to handle the required scaling actions. An extensive set of experiments for different types of anomaly problems shows that ADRL can significantly improve the quality of service with less number of actions and increased stability of the system.

Cloud service providers (CSP) and cloud consumers often need to forecast the cloud price to optimize their business strategy. However, pricing of cloud services is a challenging task due to its services complexity and dynamic nature of the ever-changing environment. Moreover, the cloud pricing based on consumers’ willingness to pay (W2P) becomes even more challenging due to the subjectiveness of consumers’ experiences and implicit values of some non-marketable features, such as burstable CPU, dedicated server, and cloud data center global footprints. Unfortunately, many existing pricing models often cannot support value-based pricing. In this paper, we propose a novel solution based on value-based pricing, which does not only consider how much does the service cost (or intrinsic values) to a CSP but also how much a customer is willing to pay (or extrinsic values) for the service. We demonstrate that the cloud extrinsic values would not only become one of the competitive advantages for CSPs to lead the cloud market but also increase the profit margin. Our approach is often referred to as a hedonic pricing model. We show that our model can capture the value of non-marketable features. This value is about 43.4 percent on average above the baseline, which is often ignored by many traditional cloud pricing models. We also show that Average Annual Growth Rate (AAGR) of Amazon Web Services’ (AWS) is about -20.0 percent per annum between 2008 and 2017, ceteris paribus. In comparison with Moore's law (-50 percent per annum), it is at a far slower pace. We argue this value is Moore's law equivalent in the cloud. The primary goal of this research is to provide a less biased pricing model for cloud decision makers to develop their optimizing investment strategy.

The fourth industrial revolution, widely known as Industry 4.0, is realizable through widespread deployment of Internet of Things (IoT) devices across the industrial ambiance. Due to communication latency and geographical distribution, Cloud-centric IoT models often fail to satisfy the Quality of Service requirements of different IoT applications assisting Industry 4.0 in real time. Therefore, Fog computing focuses on harnessing edge resources to place and execute these applications in the proximity of data sources. Since most of the Fog nodes are heterogeneous, distributed, and resource-constrained, it is challenging to place Industry 4.0-oriented applications (I4OAs) over them ensuring time-optimized service delivery. Diversified data sensing frequency of different industrial IoT devices and their data size further intensify the application placement problem. To address this issue, in this article we propose a context-aware application placement policy for Fog environments. Our policy coordinates the IoT device-level contexts with the capacity of Fog nodes and minimizes the service delivery time of various I4OAs such as image processing and robot navigation applications. It also ensures that the streams of input data flowing toward the placed applications neither congest the network nor increase the computing overhead of host Fog nodes significantly. Performance of the proposed policy is evaluated in both real-world and simulated Fog environments and compared with the existing placement policies. The experiment results show that our policy offers overall 16% improvement in service latency, network relaxation, and computing overhead management compared to other placement policies.

A common cause of traffic congestions is the concentration of intersecting vehicle routes. It can be difficult to reduce the intersecting routes in existing traffic systems where the routes are decided independently from vehicle to vehicle. The development of connected autonomous vehicles provides the opportunity to address the intersecting route problem as the route of vehicles can be coordinated globally. We prototype a traffic management system for optimizing traffic with connected autonomous vehicles. The system allocates routes to the vehicles based on streaming traffic data. We develop two route assignment algorithms for the system. The algorithms can help to mitigate traffic congestions by reducing intersecting routes. Extensive experiments are conducted to compare the proposed algorithms and two state-of-the-art route assignment algorithms with both synthetic and real road networks in a simulated traffic management system. The experimental results show that the proposed algorithms outperform the competitors in terms of the travel time of the vehicles.

Big sensing data is commonly encountered from various surveillance or sensing systems. Sampling and transferring errors are commonly encountered during each stage of sensing data processing. How to recover from these errors with accuracy and efficiency is quite challenging because of high sensing data volume and unrepeatable wireless communication environment. While Cloud provides a promising platform for processing big sensing data, however scalable and accurate error recovery solutions are still need. In this paper, we propose a novel approach to achieve fast error recovery in a scalable manner on cloud. This approach is based on the prediction of a recovery replacement data by making multiple data sources based approximation. The approximation process will use coverage information carried by data units to limit the algorithm in a small cluster of sensing data instead of a whole data spectrum. Specifically, in each sensing data cluster, a Euclidean distance based approximation is proposed to calculate a time series prediction. With the calculated time series, a detected error can be recovered with a predicted data value. Through the experiment with real world meteorological data sets on cloud, we demonstrate that the proposed error recovery approach can achieve high accuracy in data approximation to replace the original data error. At the same time, with MapReduce based implementation for scalability, the experimental results also show significant efficiency on time saving.

Modeling a utility function for cloud business customers is one of the critical challenges facing many cloud service providers (CSPs) for their pricing strategy. It concerns how to measure various subjective experiences of the business customers and how to translate their cloud service experiences into a quantifiable unit, which can be determined by a utility function that reflects cloud resource consumption. The aim of this modeling process is to set up a pricing foundation so that CSPs can target a broader range of customers from various market segments and identify the optimal price point of their various pricing models. Previous studies have either focused on simple theoretical proof or drifted the meaning of utility between demand and supply or proposed a solution based on a uniform cloud market assumption. This paper proposes a novel and practical solution to define multiple utility functions based on a scenario of six cloud market segments, which are analyzed by three analytic approaches that are known as Markov chains analysis, queueing theory, and risk assessment. The entire pricing strategy emphasizes value co-creation between CSP and cloud business customers. In comparison with other methods, such as calibrated, price-quality, resource-based, simple linear, and capacity-aware, this method provides both internal and external rationalities for CSP to capture the subjective value of cloud business customers. Consequently, our experiment results show that this modeling method can increase a profit margin by 51% and decrease a unit cost by 22% for a CSP.

The marketplace for Internet of Things (IoT)-enabled smart systems is rapidly expanding. The integration of Fog and Cloud paradigm aims at harnessing both edge device and remote datacentre-based computing resources to meet Quality of Service (QoS) requirements of these smart systems. Due to lack of instance pricing and revenue maximizing techniques, it becomes difficult for service providers to make comprehensive profit from such integration. This problem further intensifies when associated expenses and allowances are charged from the revenue. Conversely, the rigid revenue maximizing intention of providers affects user’s budget and system’s service quality. To address these issues, we propose a profit-aware application placement policy for integrated Fog–Cloud environments. It is formulated using constraint Integer Linear Programming model that simultaneously enhances profit and ensures QoS during application placement on computing instances. Furthermore, it provides compensation to users for any violation of Service Level Agreement (SLA) and sets the price of instances according to their ability of reducing service delivery time. The performance of proposed policy is evaluated in a simulated Fog–Cloud environment using iFogSim and the results demonstrate that it outperforms other placement policies in concurrently increasing provider’s profit and user’s QoS satisfaction rate.

Cloud price modeling is the major challenge facing many cloud computing practitioners and researchers in the field of cloud economics, which is also known as “Cloudonomics.” Previous attempts mainly focused on a uniform market and used existing price models to explain the issue of revenue maximization for cloud service providers (CSPs) from a cost or internal rationality perspective but paid less attention to the cloud market segmentation for cloud business customers from a surplus value or external rationality perspective. This study considers both aspects of the value proposition. Based on the assumptions of the customers’ utility values for different market segments, we establish a framework of value-based pricing strategy and demystify the process of modeling and optimizing cloud prices for CSP to maximize its profits. This framework is built upon the theory of value co-creation for both customers and CSPs to form a business partnership. We show how to create four cloud pricing models, namely: on-demand, bulk-selling, reserved, and bulk + reserved. We also demonstrate how to identify the optimal price point of each model to maximize CSP’s profit by genetic algorithm. We exhibit that reserved, bulk + reserved, on-demand, and bulk-selling can deliver a profit margin of 203%, 183%, 166%, and 157% for CSPs respectively. While the reserved model provides the highest profit margin, it does not necessarily mean that CSPs should adopt one model only. We provide a novel solution that allows CSPs to achieve the maximum profit by offering multiple pricing models simultaneously to various customers in the segmented market. We argue CSPs should capitalize on cloud pricing rather than price to gain more cloud market share and profit. Thus, we present state-of-the-art cloud pricing for segmented business to business cloud market.

White matter lesion (WML) is one of the common cerebral abnormalities, it indicates changes in the white matter of human brain and have shown significant association with stroke, dementia and deaths. Magnetic resonance imaging (MRI) of the brain is frequently used to diagnose white matter lesion (WML) volume. Regular screening can detect WML in early stage and save from severe consequences. Current option of MRI based diagnosis is impractical for regular screening because of its high expense and unavailability. Thus, earlier screening and prediction of the WML volume/load specially in the rural and remote areas becomes extremely difficult. Research has shown that changes in the retinal micro vascular system reflect changes in the cerebral micro vascular system. Using this information, we have proposed a retinal image based WML volume and severity prediction model which is very convenient and easy to operate.Our proposed model can help the physicians to detect the patients who need immediate and further MRI based detail diagnosis of WML. Our model uses quantified measurement of retinal micro-vascular signs (such as arteriovenular nicking (AVN), Opacity (OP) and focal arteriolar narrowing (FAN)) as input and estimate the WML volume/load and classify its severity. We evaluate our proposed model on a dataset of 111 patients taken from the ENVISion study which have retinal and MRI images for each patient. Our model shows high accuracy in estimating the WML volume, mean square error (MSE) between our predicted WML load and manually annotated WML load is 0.15. The proposed model achieves an F1 score of 0.92 in classifying the patients having mild and severe WML load.The preliminary results of our study indicate that quantified measurement of retinal micro-vascular features (AVN, OP and FAN) can more accurately identify the patients who have high risk of cardio-vascular diseases and dementia.

Traffic forecasting plays a vital role in traffic management systems. Recently, deep learning models have been applied to citywide traffic forecasting. However, the existing work models and predicts traffic at a single (dense) resolution, making it challenging to capture long-range spatial dependencies or high-level traffic dynamics. This shortcoming limits the accuracy of prediction and results in computationally expensive models. We propose a traffic forecasting model based on deep convolutional networks to improve the accuracy of citywide traffic forecasting. Our model uses a hierarchical architecture that captures traffic dynamics at multiple spatial resolutions. Based on this architecture, we apply a multi-task learning scheme, which trains the model to predict traffic at different resolutions. Our model helps provide a coherent understanding of traffic dynamics by capturing spatial dependencies between different regions of a city. Experimental results on multiple real datasets show that our model can achieve competitive results compared to complex state-of-the-art approaches while being more computationally efficient.

Diffusion MRI tractography is the most widely used macroscale method for mapping connectomes in vivo. However, tractography is prone to various errors and biases, and thus tractography-derived connectomes require careful validation. Here, we critically review studies that have developed or utilized phantoms and tracer maps to validate tractography-derived connectomes, either quantitatively or qualitatively. We identify key factors impacting connectome reconstruction accuracy, including streamline seeding, propagation and filtering methods, and consider the strengths and limitations of state-of-the-art connectome phantoms and associated validation studies. These studies demonstrate the inherent limitations of current fiber orientation models and tractography algorithms and their impact on connectome reconstruction accuracy. Reconstructing connectomes with both high sensitivity and high specificity is challenging, given that some tractography methods can generate an abundance of spurious connections, while others can overlook genuine fiber bundles. We argue that streamline filtering can minimize spurious connections and potentially improve the biological plausibility of connectomes derived from tractography. We find that algorithmic choices such as the tractography seeding methodology, angular threshold, and streamline propagation method can substantially impact connectome reconstruction accuracy. Hence, careful application of tractography is necessary to reconstruct accurate connectomes. Improvements in diffusion MRI acquisition techniques will not necessarily overcome current tractography limitations without accompanying modeling and algorithmic advances.

This article provides a systematic review of cloud pricing in an interdisciplinary approach. It examines many historical cases of pricing in practice and tracks down multiple roots of pricing in research. The aim is to help both cloud service provider (CSP) and cloud customers to capture the essence of cloud pricing when they need to make a critical decision either to achieve competitive advantages or to manage cloud resource effectively. Currently, the number of available pricing schemes in the cloud market is overwhelming. It is an intricate issue to understand these schemes and associated pricing models clearly due to involving several domains of knowledge, such as cloud technologies, microeconomics, operations research, and value theory. Some earlier studies have introduced this topic unsystematically. Their approaches inevitably lead to much confusion for many cloud decision-makers. To address their weaknesses, we present a comprehensive taxonomy of cloud pricing, which is driven by a framework of three fundamental pricing strategies that are built on nine cloud pricing categories. These categories can be further mapped onto a total of 60 pricing models. Many of the pricing models have been already adopted by CSPs. Others have been widespread across in other industries. We give descriptions of these model categories and highlight both advantages and disadvantages. Moreover, this article offers an extensive survey of many cloud pricing models that were proposed by many researchers during the past decade. Based on the survey, we identify four trends of cloud pricing and the general direction, which is moving from intrinsic value per physical box to extrinsic value per serverless sandbox. We conclude that hyper-converged cloud resources pool supported by cloud orchestration, virtual machine, Open Application Programming Interface, and serverless sandbox will drive the future of cloud pricing.

Fog computing aims at offering Cloud like services at the network edge for supporting Internet of Things (IoT) applications with low latency response requirements. Hierarchical, distributed and heterogeneous nature of computational instances make application placement in Fog a challenging task. Diversified user expectations and different features of IoT devices also intensify the application placement problem. Placement of applications to compatible Fog instances based on user expectations can enhance Quality of Experience (QoE) regarding the system services. In this paper, we propose a QoE-aware application placement policy that prioritizes different application placement requests according to user expectations and calculates the capabilities of Fog instances considering their current status. In Fog computing environment, it also facilitates placement of applications to suitable Fog instances so that user QoE is maximized in respect of utility access, resource consumption and service delivery. The proposed policy is evaluated by simulating a Fog environment using iFogSim. Experimental results indicate that the policy significantly improves data processing time, network congestion, resource affordability and service quality.

Users need to understand the predictions of a classifier, especially when decisions based on the predictions can have severe consequences. The explanation of a prediction reveals the reason why a classifier makes a certain prediction, and it helps users to accept or reject the prediction with greater confidence. This paper proposes an explanation method called Pattern Aided Local Explanation (PALEX) to provide instance-level explanations for any classifier. PALEX takes a classifier, a test instance and a frequent pattern set summarizing the training data of the classifier as inputs, and then outputs the supporting evidence that the classifier considers important for the prediction of the instance. To study the local behavior of a classifier in the vicinity of the test instance, PALEX uses the frequent pattern set from the training data as an extra input to guide generation of new synthetic samples in the vicinity of the test instance. Contrast patterns are also used in PALEX to identify locally discriminative features in the vicinity of a test instance. PALEX is particularly effective for scenarios where there exist multiple explanations. In our experiments, we compare PALEX to several state-of-the-art explanation methods over a range of benchmark datasets and find that it can identify explanations with both high precision and high recall.

Cloud computing as a model to deliver distributed resource and services on the pay-as-you-go policy has become increasingly popular for all organizations including industry. However, the inherent dynamicity in this environment makes it prone to various types of performance problems which introduce many challenges in the area of distributed resource management. Advances in the big data learning approaches can bring the opportunity for a data aware dynamic management of resources in the cloud. The collected data from the performance indicators of the system can be a valuable source of information to identify unusual behaviors in the resource consumptions or application performance. Different types of problems can cause the performance degradations at VM or system level. System administrators are overwhelmed with the huge amount of data to be analyzed to find the problems and overall health of the system. In this paper, we argue that a better selection of dynamic resource scaling policies can be employed for better performance by predicting the anomalies in the system and narrowing down the possible cause of the anomaly to one of the attributes of the system. Therefore, we propose a 2-level cause aware auto-scaling framework which leverages two types of resource management solutions, horizontal and vertical, as the corrective actions when the performance is degraded. We show the effectiveness of vertical scaling strategy as a quick solution for cases that a VM is exposed to some type of the local anomaly, while the horizontal scaling solutions can be used for system wide anomaly to add new VMs in the system. Moreover, our data analysis module can predict anomalies to give sufficient time to the scaling system to make an effective scaling decision. The proposed unsupervised anomaly detection module leverages a new updating strategy for renewing the models which considers the changes in the state of the system to reduce the overhead of recurrent model trainings. We have performed a comparison of the proposed framework with an approach which is used by several popular cloud providers to show the advantage of mixing the multi-level auto-scaling with the knowledge of anomaly detection analysis in resolving performance problems in the cloud.

Thermal management in the hyper-scale cloud data centers is a critical problem. Increased host temperature creates hotspots which significantly increases cooling cost and affects reliability. Accurate prediction of host temperature is crucial for managing the resources effectively. Temperature estimation is a non-trivial problem due to thermal variations in the data center. Existing solutions for temperature estimation are inefficient due to their computational complexity and lack of accurate prediction. However, data-driven machine learning methods for temperature prediction is a promising approach. In this regard, we collect and study data from a private cloud and show the presence of thermal variations. We investigate several machine learning models to accurately predict the host temperature. Specifically, we propose a gradient boosting machine learning model for temperature prediction. The experiment results show that our model accurately predicts the temperature with the average RMSE value of 0.05 or an average prediction error of 2.38 degree Celsius, which is 6 degree Celsius less as compared to an existing theoretical model. In addition, we propose a dynamic scheduling algorithm to minimize the peak temperature of hosts. The results show that our algorithm reduces the peak temperature by 6.5 degree Celsius and consumes 34.5% less energy as compared to the baseline algorithm., Under submission at IEEE Transactions on Parallel and Distributed Systems (TPDS)

The cost of traffic congestions has been significantly high in many countries. Traffic congestion can be minimized by coordinated route allocation to maximize the traffic efficiency in the whole road network. Unfortunately, the existing traffic management systems cannot achieve this type of optimization as vehicles tend to follow the shortest/fastest routes to their destinations. Such individually optimized routes may cause significant congestions in a road network. In the coming era of connected autonomous vehicles, traffic management systems can have access to a huge volume of prior temporal traffic data that depicts the historical traffic conditions collected at regular time intervals. This type of data provides great opportunities for traffic optimization at the network level. We propose a new route assignment algorithm for the era of connected autonomous vehicles. Our algorithm optimizes traffic based on real-time traffic conditions and prior temporal traffic data. Our experiments show that the proposed algorithm can improve traffic efficiency by up to 10% over the state-of-the-art algorithm.

While the function of most biological systems is tightly constrained by their structure, current evidence suggests that coupling between the structure and function of brain networks is relatively modest. We aimed to investigate whether the modest coupling between connectome structure and function is a fundamental property of nervous systems or a limitation of current brain network models. We developed a new deep learning framework to predict an individual's brain function from their structural connectome, achieving prediction accuracies that substantially exceeded state-of-the-art biophysical models (group: R=0.9±0.1, individual: R=0.55±0.1). Crucially, brain function predicted from an individual's structural connectome explained significant inter-individual variation in cognitive performance. Our results suggest that structure-function coupling in human brain networks is substantially tighter than previously suggested. We establish the margin by which current brain network models can be improved and demonstrate how deep learning can facilitate investigation of relations between brain function and behavior.

Ensemble selection is a very active research topic in machine learning area. It aims to achieve a better performance by selecting a proper subset of the original ensemble, which is essentially a searching problem in large combinatorial spaces. In this paper, we propose an instance-based reinforcement learning (IBRL) model, that selects distinct subsets for different instances. Specifically, we use deep Q-network to approximate the optimal policy. Rather than considering the overall performance of each classifier, the network learns from the feedback of classifiers on individual instance, so that it generates non-static subsets for different instances. Experiments are conducted to compare our model against state-of-the-art approaches for both selection and combination. The proposed method generates promising results and it shows exceptional advantage in large scale distributed environment. Due to the environment-free characteristic of reinforcement learning, our model is adaptable to various real world tasks with minimal changes.

The Internet of Things (IoT) paradigm is being rapidly adopted for the creation of smart environments in various domains. The IoT-enabled cyber-physical systems associated with smart city, healthcare, Industry 4.0 and Agtech handle a huge volume of data and require data processing services from different types of applications in real time. The Cloud-centric execution of IoT applications barely meets such requirements as the Cloud datacentres reside at a multi-hop distance from the IoT devices. Fog computing , an extension of Cloud at the edge network, can execute these applications closer to data sources. Thus, Fog computing can improve application service delivery time and resist network congestion. However, the Fog nodes are highly distributed and heterogeneous, and most of them are constrained in resources and spatial sharing. Therefore, efficient management of applications is necessary to fully exploit the capabilities of Fog nodes. In this work, we investigate the existing application management strategies in Fog computing and review them in terms of architecture, placement and maintenance. Additionally, we propose a comprehensive taxonomy and highlight the research gaps in Fog-based application management. We also discuss a perspective model and provide future research directions for further improvement of application management in Fog computing.

We propose a new framework to map structural connectomes using deep learning and diffusion MRI. We show that our framework not only enables connectome mapping with a convolutional neural network (CNN), but can also be straightforwardly incorporated into conventional connectome mapping pipelines to enhance accuracy. Our framework involves decomposing the entire brain volume into overlapping blocks. Blocks are sufficiently small to ensure that a CNN can be efficiently trained to predict each block's internal connectivity architecture. We develop a block stitching algorithm to rebuild the full brain volume from these blocks and thereby map end-to-end connectivity matrices. To evaluate our block decomposition and stitching (BDS) framework independent of CNN performance, we first map each block's internal connectivity using conventional streamline tractography. Performance is evaluated using simulated diffusion MRI data generated from numerical connectome phantoms with known ground truth connectivity. Due to the redundancy achieved by allowing blocks to overlap, we find that our block decomposition and stitching steps per se can enhance the accuracy of probabilistic and deterministic tractography algorithms by up to 20-30%. Moreover, we demonstrate that our framework can improve the strength of structure-function coupling between in vivo diffusion and functional MRI data. We find that structural brain networks mapped with deep learning correlate more strongly with functional brain networks (r ​= ​0.45) than those mapped with conventional tractography (r ​= ​0.36). In conclusion, our BDS framework not only enables connectome mapping with deep learning, but its two constituent steps can be straightforwardly incorporated as part of conventional connectome mapping pipelines to enhance accuracy.

Semantic segmentation is one of the key problems in the field of computer vision, as it enables computer image understanding. However, most research and applications of semantic segmentation focus on addressing unique segmentation problems, where there is only one gold standard segmentation result for every input image. This may not be true in some problems, e.g., medical applications. We may have non-unique segmentation annotations as different surgeons may perform successful surgeries for the same patient in slightly different ways. To comprehensively learn non-unique segmentation tasks, we propose the reward-penalty Dice loss (RPDL) function as the optimization objective for deep convolutional neural networks (DCNN). RPDL is capable of helping DCNN learn non-unique segmentation by enhancing common regions and penalizing outside ones. Experimental results show that RPDL improves the performance of DCNN models by up to 18.4% compared with other loss functions on our collected surgical dataset.

The fog computing paradigm has drawn significant research interest as it focuses on bringing cloud-based services closer to Internet of Things (IoT) users in an efficient and timely manner. Most of the physical devices in the fog computing environment, commonly named fog nodes, are geographically distributed, resource constrained, and heterogeneous. To fully leverage the capabilities of the fog nodes, large-scale applications that are decomposed into interdependent Application Modules can be deployed in an orderly way over the nodes based on their latency sensitivity. In this article, we propose a latency-aware Application Module management policy for the fog environment that meets the diverse service delivery latency and amount of data signals to be processed in per unit of time for different applications. The policy aims to ensure applications’ Quality of Service (QoS) in satisfying service delivery deadlines and to optimize resource usage in the fog environment. We model and evaluate our proposed policy in an iFogSim-simulated fog environment. Results of the simulation studies demonstrate significant improvement in performance over alternative latency-aware strategies.

Purpose Human connectomics necessitates high-throughput, whole-brain reconstruction of multiple white matter fiber bundles. Scaling up tractography to meet these high-throughput demands yields new fiber tracking challenges, such as minimizing spurious connections and controlling for gyral biases. The aim of this study is to determine which of the two broadest classes of tractography algorithms-deterministic or probabilistic-is most suited to mapping connectomes. Methods This study develops numerical connectome phantoms that feature realistic network topologies and that are matched to the fiber complexity of in vivo diffusion MRI (dMRI) data. The phantoms are utilized to evaluate the performance of tensor-based and multi-fiber implementations of deterministic and probabilistic tractography. Results For connectome phantoms that are representative of the fiber complexity of in vivo dMRI, multi-fiber deterministic tractography yields the most accurate connectome reconstructions (F-measure = 0.35). Probabilistic algorithms are hampered by an abundance of false-positive connections, leading to lower specificity (F = 0.19). While omitting connections with the fewest number of streamlines (thresholding) improves the performance of probabilistic algorithms (F = 0.38), multi-fiber deterministic tractography remains optimal when it benefits from thresholding (F = 0.42). Conclusions Multi-fiber deterministic tractography is well suited to connectome mapping, while connectome thresholding is essential when using probabilistic algorithms.

Fog computing overcomes the limitations of executing Internet of Things (IoT) applications in remote Cloud datacentres by extending the computation facilities closer to data sources. Since most of the Fog nodes are resource constrained, accommodation of every IoT application within Fog environments is very challenging. Hence, we need to efficiently identify which set of applications should be deployed in Fog. It becomes even more complicated when the application characteristics in terms of urgency, size and flow of inputs are considered simultaneously. The necessity of time-optimized execution further intensifies the application management problem. In this work, we propose a policy for Fog environments that distributes application management tasks across the gateway and the infrastructure level. It classifies and places applications according to their Edge affinity. Edge affinity of an application denotes the relative intensity of different attributes coherent with its characteristics such as user-defined deadline, amount of data per input and sensing frequency of IoT devices, which are required to be addressed within Fog environments to meet its Quality of Service (QoS). The proposed policy also minimizes the service delivery time of applications in Fog infrastructure. Its performance is compared with existing application management policies in both iFogSim-simulated and FogBus-based real environments. The experiment results show that our policy outperforms others in combined QoS enhancement, network relaxation and resource utilization.

Summary Debugging is crucial for producing reliable software. One of the effective bug localization techniques is spectral-based fault localization. It tries to locate a buggy statement by applying an evaluation metric to program spectra and ranking program components on the basis of the score it computes. Here, we propose a restricted class of “hyperbolic” metrics, with a small number of numeric parameters. This class of functions is based on past theoretical and empirical results. We show that optimization methods such as genetic programming and simulated annealing can reliably discover effective metrics over a wide range of data sets of program spectra. We evaluate the performance for both real programs and model programs with single bugs, multiple bugs, “deterministic” bugs, and nondeterministic bugs and find that the proposed class of metrics performs as well as or better than the previous best-performing metrics over a broad range of data.

Car occupancy rates (travelers per vehicle) are currently very low in most developed countries, for example, on average between 1.15 and 1.25 in Australia. Enabling shared rides on short notice can be an effective solution to counter the problem of increasing traffic through the use of the untapped transportation capacity. Common inhibitors for the uptake of ride sharing services are privacy and safety concerns. We present an approach to ride sharing where the pick up/drop off locations for passengers are selected from a fixed set, which has the advantage of increased safety through video surveillance. We present a scheme that chooses optimally fixed locations of Pick up Points (PuPs) and aims to maximize the car occupancy rates while preserving user privacy and safety. Our method enhances privacy as the users do not need to provide their precise home/work locations. We have extended the well studied 1-coverage problem, i.e., to cover an area with the minimum number of circles of a given radius [1] , to road networks. The challenges for road networks are the varying population densities of suburbs which requires circles of different radii. The aim is to ensure that every point of a city's area is covered by at least one PuP while minimizing the total number of PuPs. By ensuring that we have different circle radii for PuPs the anonymity of individuals is the same throughout. Using Voronoi diagrams we present a k -anonymity model that guarantees a minimum number of individuals covered by every PuP. Our problem is a multi objective problem where we aim to maximize coverage, k -anonymity and privacy provided by the system to its users while facilitating ride sharing. Through greedy randomized adaptive search procedure (GRASP) we find out the Pareto front of solutions and evaluate their impact on ride sharing.

Summary Debugging is crucial for producing reliable software. One of the effective bug localization techniques is spectral-based fault localization (SBFL). It helps to locate a buggy statement by applying an evaluation metric to program spectra and ranking program components on the basis of the score it computes. SBFL is an example of a dynamic analysis – an analysis of computer program that is performed by executing it with sufficient number of test cases. Static analysis, on the other hand, is performed in a non-runtime environment. We introduce a weighting technique by combining these two kinds of program analysis. Static analysis is performed to categorize program statements into different classes and giving them weights based on the likelihood of being buggy statement. Statements are finally ranked on the basis of the weights computed by statements' categorization (static analysis) and scores computed by SBFL metrics (dynamic analysis). We evaluate the performance of our technique on Siemens test suite and Flex (having seeded bugs seeded by expert developers), Sed (having mixture of real and seeded bugs), and Space (having real bugs). In our evaluation, proposed weighting technique improves the performance of a wide variety of fault localization metrics up to 20% on single bug datasets and up to 42% on multi-bug datasets. Copyright © 2017 John Wiley & Sons, Ltd.

Retinal vascular network pattern is unique to each individual which can be used for person identification in biometric authentication. In this study, the authors have proposed a novel biometric authentication method using retinal vascular branch, bifurcation and crossover points (i.e. feature points). The method automatically extracts the vascular network from colour retinal images and identifies these feature points. The major blood vessels characterised by width and length are identified from the segmented vascular network. For this, a novel vessel width measurement method is applied and vessels more than certain widths are selected as major vessels following an established protocol. The geometric hashing technique is developed to compute the invariant features from these feature points. They consider the feature points from major vessels which will be less susceptible to noise for modelling a basis pair and all other points together for locations in the hash table entries. The models are invariant to rotation, translation and scaling as inherited from geometric hashing. For each person, the system is trained with the models to accept or reject a claimed identity. They have tested their method on 3010 retinal images and achieved 96.64% precision and 100% recall.

We use traffic simulations to quantify the impact of autonomous vehicles in various traffic scenarios, where vehicles at higher automation levels behave more opportunistically in car-following and lane-changing and can react to road situations more quickly. Our experimental results show that an increased automation level can improve traffic efficiency but may lead to more potential conflicts between vehicles, which should not be neglected if human drivers still need to take part in the driving.

Minimizing the energy consumption of servers within cloud computing systems is of upmost importance to cloud providers toward reducing operational costs and enhancing service sustainability by consolidating services onto fewer active servers. Moreover, providers must also provision high levels of availability and reliability, hence cloud services are frequently replicated across servers that subsequently increases server energy consumption and resource overhead. These two objectives can present a potential conflict within cloud resource management decision making that must balance between service consolidation and replication to minimize energy consumption whilst maximizing server availability and reliability, respectively. In this paper, we propose a cuckoo optimization-based energy-reliability aware resource scheduling technique (CRUZE) for holistic management of cloud computing resources including servers, networks, storage, and cooling systems. CRUZE clusters and executes heterogeneous workloads on provisioned cloud resources and enhances the energy-efficiency and reduces the carbon footprint in datacenters without adversely affecting cloud service reliability. We evaluate the effectiveness of CRUZE against existing state-of-the-art solutions using the CloudSim toolkit. Results indicate that our proposed technique is capable of reducing energy consumption by 20.1% whilst improving reliability and CPU utilization by 17.1% and 15.7% respectively without affecting other Quality of Service parameters.

Image inpainting aims at restoring missing regions of corrupted images, which has many applications such as image restoration and object removal. However, current GAN-based generative inpainting models do not explicitly exploit the structural or textural consistency between restored contents and their surrounding contexts.To address this limitation, we propose to enforce the alignment (or closeness) between the local data submanifolds (or subspaces) around restored images and those around the original (uncorrupted) images during the learning process of GAN-based inpainting models. We exploit Local Intrinsic Dimensionality (LID) to measure, in deep feature space, the alignment between data submanifolds learned by a GAN model and those of the original data, from a perspective of both images (denoted as iLID) and local patches (denoted as pLID) of images. We then apply iLID and pLID as regularizations for GAN-based inpainting models to encourage two levels of submanifold alignment: 1) an image-level alignment for improving structural consistency, and 2) a patch-level alignment for improving textural details. Experimental results on four benchmark datasets show that our proposed model can generate more accurate results than state-of-the-art models., Comment: accepted by IJCAI 2019

Cloud computing is a model for on-demand access to shared resources based on the pay-per-use policy. In order to efficiently manage the resources, a continuous analysis of the operational state of the system is required to be able to detect the performance degradations and malfunctioned resources as soon as possible. Every change in the workload, hardware condition, or software code can change the state of the system from normal to abnormal, which causes the performance and quality of service degradations. These changes or anomalies vary from a simple gradual increase in the load to flash crowds, hardware faults, software bugs, etc. In this paper, we propose Isolation-Forest based anomaly detection (IFAD) framework based on the unsupervised Isolation technique for anomaly detection in a multi-attribute space of performance indicators for web-based applications. Unsupervised nature of the algorithm and its fast execution make this algorithm most suitable for the environments with dynamic nature where the patterns of data change frequently. The experiment results demonstrate that IFAD can achieve good detection accuracy especially in terms of precision for multiple types of the anomaly. Moreover, we show the importance of validating the accuracy of anomaly detection algorithms with regard to both Area Under the Curve (AUC) and Precision-Recall AUC (PRAUC) in an extensive set of comparisons including multiple unsupervised algorithms. The demonstration of the effectiveness of each algorithm shown by PRAUC results indicates the importance of PRAUC in selecting suitable anomaly detection algorithm, which is largely ignored in the literature.

Trip recommendation is an important location-based service that helps relieve users from the time and efforts for trip planning. It aims to recommend a sequence of places of interest (POIs) for a user to visit that maximizes the user's satisfaction. When adding a POI to a recommended trip, it is essential to understand the context of the recommendation, including the POI popularity, other POIs co-occurring in the trip, and the preferences of the user. These contextual factors are learned separately in existing studies, while in reality, they jointly impact on a user's choice of POI visits. In this study, we propose a POI embedding model to jointly learn the impact of these contextual factors. We call the learned POI embedding a context-aware POI embedding. To showcase the effectiveness of this embedding, we apply it to generate trip recommendations given a user and a time budget. We propose two trip recommendation algorithms based on our context-aware POI embedding. The first algorithm finds the exact optimal trip by transforming and solving the trip recommendation problem as an integer linear programming problem. To achieve a high computation efficiency, the second algorithm finds a heuristically optimal trip based on adaptive large neighborhood search. We perform extensive experiments on real datasets. The results show that our proposed algorithms consistently outperform state-of-the-art algorithms in trip recommendation quality, with an advantage of up to 43% in F_1-score.

We present a novel method for the quantification of focal arteriolar narrowing (FAN) in human retina, a precursor for hypertension, stroke and other cardiovascular diseases. A reliable and robust arteriolar boundary mapping method is proposed where intensity, gradient and spatial prior knowledge about the arteriolar shape is incorporated into a graph based optimization method to obtain the arteriolar boundary. Following the mapping of the arteriolar boundaries, arteriolar widths are analysed to quantify the severity of focal arteriolar narrowing (FAN). We evaluate our proposed method on a dataset of 116 retinal arteriolar segments which are manually graded by two expert graders. The experimental results indicate a strong correlation between the quantified FAN measurement scores provided by our method and two experts graded FAN severity levels. Our proposed FAN measurement score: percent narrowing (PN) shows high correlation (Spearman correlation coefficient of 0.82 ( p < 0.0001 ) for grader-1 and 0.84 ( p < 0.0001 ) for grade-2) with the manually graded FAN severity levels provided by two expert graders. In addition to that, the proposed method shows better reproducibility (Spearman correlation coefficient ź = 0.92 ( p < 0.0001 ) ) compared to two expert graders ( ź grader 1 = 0.81 ( p < 0.0001 ) and ź grader 2 = 0.75 ( p < 0.0001 ) ) in two successive sessions. The quantitative measurements provided by the proposed method can help us to establish a more reliable link between FAN and known systemic and eye diseases.

Background Relation extraction from biomedical publications is an important task in the area of semantic mining of text. Kernel methods for supervised relation extraction are often preferred over manual feature engineering methods, when classifying highly ordered structures such as trees and graphs obtained from syntactic parsing of a sentence. Tree kernels such as the Subset Tree Kernel and Partial Tree Kernel have been shown to be effective for classifying constituency parse trees and basic dependency parse graphs of a sentence. Graph kernels such as the All Path Graph kernel (APG) and Approximate Subgraph Matching (ASM) kernel have been shown to be suitable for classifying general graphs with cycles, such as the enhanced dependency parse graph of a sentence. In this work, we present a high performance Chemical-Induced Disease (CID) relation extraction system. We present a comparative study of kernel methods for the CID task and also extend our study to the Protein-Protein Interaction (PPI) extraction task, an important biomedical relation extraction task. We discuss novel modifications to the ASM kernel to boost its performance and a method to apply graph kernels for extracting relations expressed in multiple sentences. Results Our system for CID relation extraction attains an F-score of 60%, without using external knowledge sources or task specific heuristic or rules. In comparison, the state of the art Chemical-Disease Relation Extraction system achieves an F-score of 56% using an ensemble of multiple machine learning methods, which is then boosted to 61% with a rule based system employing task specific post processing rules. For the CID task, graph kernels outperform tree kernels substantially, and the best performance is obtained with APG kernel that attains an F-score of 60%, followed by the ASM kernel at 57%. The performance difference between the ASM and APG kernels for CID sentence level relation extraction is not significant. In our evaluation of ASM for the PPI task, ASM performed better than APG kernel for the BioInfer dataset, in the Area Under Curve (AUC) measure (74% vs 69%). However, for all the other PPI datasets, namely AIMed, HPRD50, IEPA and LLL, ASM is substantially outperformed by the APG kernel in F-score and AUC measures. Conclusions We demonstrate a high performance Chemical Induced Disease relation extraction, without employing external knowledge sources or task specific heuristics. Our work shows that graph kernels are effective in extracting relations that are expressed in multiple sentences. We also show that the graph kernels, namely the ASM and APG kernels, substantially outperform the tree kernels. Among the graph kernels, we showed the ASM kernel as effective for biomedical relation extraction, with comparable performance to the APG kernel for datasets such as the CID-sentence level relation extraction and BioInfer in PPI. Overall, the APG kernel is shown to be significantly more accurate than the ASM kernel, achieving better performance on most datasets.

Manifold landmarks can approximately represent the low-dimensional nonlinear manifold structure embedded in high-dimensional ambient feature space. Due to the quadratic complexity of many learning algorithms in the number of training samples, selecting a sample subset as manifold landmarks has become an important issue for scalable learning. Unfortunately, state-of-the-art Gaussian process methods for selecting manifold landmarks themselves are not scalable to large datasets. In an attempt to speed up learning manifold landmarks, uniformly selected minibatch stochastic gradient descent is used by the state-of-the-art approach. Unfortunately, this approach only goes part way to making manifold learning tractable. We propose two adaptive sample selection approaches for gradient-descent optimization, which can lead to better performance in accuracy and computational time. Our methods exploit the compatibility of locality-sensitive hashing (via LSH and DBH) and the manifold assumption, thereby limiting expensive optimization to relevant regions of the data. Landmarks selected by our methods achieve superior accuracy than training the state-of-the-art learner with randomly selected minibatch. We also demonstrate that our methods can be used to find manifold landmarks without learning Gaussian processes at all, which leads to orders-of-magnitude speed up with only minimal decrease in accuracy.

The topics of cloud pricing models and resources management have been receiving enormous attention recently. However, very few studies have considered the importance of cloud market segmentation. Moreover, there is no a better, practical and quantifiable solution for a cloud service providers (CSP) to segment cloud market. We propose a novel solution that combines both hierarchical clustering and time series forecasting on the basis of the classical theory of market segmentation. In comparison with some traditional approaches, such as nested, analytic, Delphi, and strategy-based approaches, our method is much more effective, flexible, measurable and practical for CSPs to implement their cloud market strategies by rolling out different pricing models. Our tested results and empirical analysis show that our solution can efficiently segment cloud markets and also predict the market demands. Our primary goal is to offer a new solution so that CSPs can tail its limited cloud resources for its targeted market or cloud customers.

In the past decade, positioning system-enabled devices such as smartphones have become most prevalent. This functionality brings the increasing popularity of location-based services in business as well as daily applications such as navigation, targeted advertising, and location-based social networking. Continuous spatial queries serve as a building block for location-based services. As an example, an Uber driver may want to be kept aware of the nearest customers or service stations. Continuous spatial queries require updates to the query result as the query or data objects are moving. This poses challenges to the query efficiency, which is crucial to the user experience of a service. A large number of approaches address this efficiency issue using the concept of safe region . A safe region is a region within which arbitrary movement of an object leaves the query result unchanged. Such a region helps reduce the frequency of query result update and hence improves query efficiency. As a result, safe region-based approaches have been popular for processing various types of continuous spatial queries. Safe regions have interesting theoretical properties and are worth in-depth analysis. We provide a comparative study of safe region-based approaches. We describe how safe regions are computed for different types of continuous spatial queries, showing how they improve query efficiency. We compare the different safe region-based approaches and discuss possible further improvements.

In recommendation systems, items of interest are often classified into categories such as genres of movies. Existing research has shown that diversified recommendations can improve real user experience. However, most existing methods do not consider the fact that users’ levels of interest (i.e., user preferences) in different categories usually vary, and such user preferences are not reflected in the diversified recommendations. We propose an algorithm that considers user preferences for different categories when recommending diversified results, and refer to this problem as personalized recommendation diversification. In the proposed algorithm, a model that captures user preferences for different categories is optimized jointly toward both relevance and diversity. To provide the proposed algorithm with informative training labels and effectively evaluate recommendation diversity, we also propose a new personalized diversity measure. The proposed measure overcomes limitations of existing measures in evaluating recommendation diversity: existing measures either cannot effectively handle user preferences for different categories, or cannot evaluate both relevance and diversity at the same time. Experiments using two real-world datasets confirm the superiority of the proposed algorithm, and show the effectiveness of the proposed measure in capturing user preferences.

Monitoring and predicting traffic conditions are of utmost importance in reacting to emergency events in time and for computing the real-time shortest travel-time path. Mobile sensors, such as GPS devices and smartphones, are useful for monitoring urban traffic due to their large coverage area and ease of deployment. Many researchers have employed such sensed data to model and predict traffic conditions. To do so, we first have to address the problem of associating GPS trajectories with the road network in a robust manner. Existing methods rely on point-by-point matching to map individual GPS points to a road segment. However, GPS data is imprecise due to noise in GPS signals. GPS coordinates can have errors of several meters and, therefore, direct mapping of individual points is error prone. Acknowledging that every GPS point is potentially noisy, we propose a radically different approach to overcome inaccuracy in GPS data. Instead of focusing on a point-by-point approach, our proposed method considers the set of relevant GPS points in a trajectory that can be mapped together to a road segment. This clustering approach gives us a macroscopic view of the GPS trajectories even under very noisy conditions. Our method clusters points based on the direction of movement as a spatial-linear cluster, ranks the possible route segments in the graph for each group, and searches for the best combination of segments as the overall path for the given set of GPS points. Through extensive experiments on both synthetic and real datasets, we demonstrate that, even with highly noisy GPS measurements, our proposed algorithm outperforms state-of-the-art methods in terms of both accuracy and computational cost.

Hitchhiking is the oldest ridesharing process without prior arrangements by the ride sharers. It usually involves uncertain waiting times and various combinations of lifts on roads. For this way of traveling, the problem of finding an optimal route is extremely important and has not been studied. We propose the concept of a hitchhiking graph to represent all possible decisions that a hitchhiker can consider on a road network. We develop an efficient pruning technique for a faster computation of the optimal route with the least expected journey time. The effectiveness of our methods is evaluated on road networks of selected countries.

Display Omitted A sequential contrast patterns based septic shock prediction approach is proposed.Reconstructs a patient instance as an ordered sequence of contrast patterns.Coupled HMMs are applied to multiple channels of contrast pattern sequences.Coupled HMM results are compared against single variable HMM and SVM models.Meta-information about patterns, like sequence order, improve model performance. Background and objectiveCritical care patient events like sepsis or septic shock in intensive care units (ICUs) are dangerous complications which can cause multiple organ failures and eventual death. Preventive prediction of such events will allow clinicians to stage effective interventions for averting these critical complications. MethodsIt is widely understood that physiological conditions of patients on variables such as blood pressure and heart rate are suggestive to gradual changes over a certain period of time, prior to the occurrence of a septic shock. This work investigates the performance of a novel machine learning approach for the early prediction of septic shock. The approach combines highly informative sequential patterns extracted from multiple physiological variables and captures the interactions among these patterns via coupled hidden Markov models (CHMM). In particular, the patterns are extracted from three non-invasive waveform measurements: the mean arterial pressure levels, the heart rates and respiratory rates of septic shock patients from a large clinical ICU dataset called MIMIC-II. Evaluation and resultsFor baseline estimations, SVM and HMM models on the continuous time series data for the given patients, using MAP (mean arterial pressure), HR (heart rate), and RR (respiratory rate) are employed. Single channel patterns based HMM (SCP-HMM) and multi-channel patterns based coupled HMM (MCP-HMM) are compared against baseline models using 5-fold cross validation accuracies over multiple rounds. Particularly, the results of MCP-HMM are statistically significant having a p-value of 0.0014, in comparison to baseline models. Our experiments demonstrate a strong competitive accuracy in the prediction of septic shock, especially when the interactions between the multiple variables are coupled by the learning model. ConclusionsIt can be concluded that the novelty of the approach, stems from the integration of sequence-based physiological pattern markers with the sequential CHMM model to learn dynamic physiological behavior, as well as from the coupling of such patterns to build powerful risk stratification models for septic shock patients.

During the recent years, workflows have emerged as an important abstraction for collaborative research and managing complex large-scale distributed data analytics. Workflows are increasingly becoming prevalent in various distributed environments, such as clusters, grids, and clouds. These environments provide complex infrastructures that aid workflows in scaling and parallel execution of their components. However, they are prone to performance variations and different types of failures. Thus, workflow management systems need to be robust against performance variations and tolerant against failures. Numerous research studies have investigated fault-tolerant aspect of the workflow management system in different distributed systems. In this study, we analyze these efforts and provide an in-depth taxonomy of them. We present the ontology of faults and fault-tolerant techniques then position the existing workflow management systems with respect to the taxonomies and the techniques. In addition, we classify various failure models, metrics, tools, and support systems. Finally, we identify and discuss the strengths and weaknesses of the current techniques and provide recommendations on future directions and open areas for the research community.

Regularization plays an important role in machine learning systems. We propose a novel methodology for model regularization using random projection. We demonstrate the technique on neural networks, since such models usually comprise a very large number of parameters, calling for strong regularizers. It has been shown recently that neural networks are sensitive to two kinds of samples: (i) adversarial samples, which are generated by imperceptible perturbations of previously correctly-classified samples—yet the network will misclassify them; and (ii) fooling samples, which are completely unrecognizable, yet the network will classify them with extremely high confidence. In this paper, we show how robust neural networks can be trained using random projection. We show that while random projection acts as a strong regularizer, boosting model accuracy similar to other regularizers, such as weight decay and dropout, it is far more robust to adversarial noise and fooling samples. We further show that random projection also helps to improve the robustness of traditional classifiers, such as Random Forrest and Gradient Boosting Machines.

Data centers consume an enormous amount of energy to meet the ever‐increasing demand for cloud resources. Computing and Cooling are the two main subsystems that largely contribute to energy consumption in a data center. Dynamic Virtual Machine (VM) consolidation is a widely adopted technique to reduce the energy consumption of computing systems. However, aggressive consolidation leads to the creation of local hotspots that has adverse effects on energy consumption and reliability of the system. These issues can be addressed through efficient and thermal‐aware consolidation methods. We propose an Energy and Thermal‐Aware Scheduling (ETAS) algorithm that dynamically consolidates VMs to minimize the overall energy consumption while proactively preventing hotspots. ETAS is designed to address the trade‐off between time and the cost savings and it can be tuned based on the requirement. We perform extensive experiments by using the real‐world traces with precise power and thermal models. The experimental results and empirical studies demonstrate that ETAS outperforms other state‐of‐the‐art algorithms by reducing overall energy without any hotspot creation.

Selecting appropriate services for task execution in workflows should not only consider budget and deadline constraints, but also ensure the best probability that workflow will succeed and minimize the potential loss in case of exceptions. This requirement is more critical for data-intensive applications in grids or clouds since any failure is costly. Therefore, we design a fine-grained risk evaluation model customized for workflows to precisely compute the cost of failure for selected services. In comparison with current course-grained model, ours takes the relation of task dependency into consideration and assigns higher impact factor to tasks at the end. Thereafter, we design the utility function with the model and apply a genetic algorithm to find the optimized service allocations, thereby maximizing the robustness of the workflow while minimizing the possible risk of failure. Experiments and analysis show that the application of customized risk evaluation model into service selection can generally improve the successful probability of a workflow while reducing its exposure to the risk.