Search

With this position paper, we want to draw the attention of the community toward theoretical work and practical opportunities that have been overlooked so far regarding concepts, principles, and design knowledge for the user experience of Augmented and Mixed Reality content, I/O devices, interactions, applications, and systems. Despite considerable innovations in commercial products and research prototypes enabling Augmented and Mixed Reality worlds, how to design great user experiences in such worlds has been overall neglected at core, while the information and knowledge currently available to practitioners cover usability aspects mostly. Therefore, we advocate for theoretical foundations of the user experience in Augmented and Mixed Reality and propose several directions for more scientific research in this regard.

The enhanced capabilities of large scale parallel and distributed platforms produce a continuously increasing amount of data which have to be stored, exchanged and used by various tasks allocated on different nodes of the system. The management of such a huge communication demand is crucial for reaching the best possible performance of the system. Meanwhile, we have to deal with more interferences as the trend is to use a single all-purpose interconnection network whatever the interconnect (tree-based hierarchies or topology-based heterarchies). There are two different types of communications, namely, the flows induced by data exchanges during the computations, and the flows related to Input/Output operations. We propose in this paper a general model for interference-aware scheduling, where explicit communications are replaced by external topological constraints. Specifically, the interferences of both communication types are reduced by adding geometric constraints on the allocation of tasks into machines. The proposed constraints reduce implicitly the data movements by restricting the set of possible allocations for each task. This methodology has been proved to be efficient in a recent study for a restricted interconnection network (a line/ring of processors which is an intermediate between a tree and higher dimensions grids/torus). The obtained results illustrated well the difficulty of the problem even on simple topologies, but also provided a pragmatic greedy solution, which was assessed to be efficient by simulations. We are currently extending this solution for more complex topologies. This work is a position paper which describes the methodology, it does not focus on the solving part.

FT-Matrix is a typical vector-SIMD architecture that refines the cooperation between scalar and vector units. This approach is widely used in digital signal processing, high-performance computing, and artificial intelligence, among other fields. FT-Matrix currently adopts C vector extension as the main programming model, improving the utilization efficiency of SIMD by providing explicit vector extension API. Moreover, it is difficult to efficiently transplant parallel programs (OpenCL, CUDA) adopted by users. This paper proposes an automatic mapping and code optimization method for OpenCL kernels on FT-Matrix architecture. The proposed approach solves these challenges by means of work item coalescing, slicing and rotation, and instruction-level code optimization. Preliminary results show that our method can achieve high performance and good hardware utilization for OpenCL kernels, as well as decreasing the programming difficulty on FT-Matrix.

This paper introduces a context-aware system, named CAPFF, for helping people in filling paper forms, mainly in two contexts: 1) people have no idea of what should be filled in certain form fields; 2) people are not aware of the mistakes they are likely to commit in entering information, which may violate data entry constraints. In the offline phase, CAPFF provides a tool to build the knowledge about a given form, and such knowledge includes instructions, field-level examples, and constraints among form fields. In the online phase, when people set out to fill a paper form, the video camera of the system determines the position of the pen and then provides assistance, based on the user's form filling context. We evaluated CAPFF's performance through 450 paper form filling activities, and the results show that the proposed CAPFF is effective in terms of both accuracy and response time.

Deep Neural Networks (DNNs) are the algorithm of choice for image processing applications. DNNs present highly parallel workloads that lead to the emergence of custom hardware accelerators. Deep Learning (DL) models specialized in different tasks require a programmable custom hardware and a compiler/mapper to efficiently translate different DNNs into an efficient dataflow in the accelerator. The goal of this paper is to present a compiler for running DNNs on Snowflake, which is a programmable hardware accelerator that targets DNNs. The compiler correctly generates instructions for various DL models: AlexNet, VGG, ResNet and LightCNN9. Snowflake, with a varying number of processing units, was implemented on FPGA to measure the compiler and Snowflake performance properties upon scaling up. The system achieves 70 frames/s and 4.5 GB/s of off-chip memory bandwidth for AlexNet without linear layers on Xilinx’s Zynq-SoC XC7Z045 FPGA.

We present a silent self-stabilizing distributed algorithm computing a maximal p-star decomposition of the underlying communication network. Under the unfair distributed scheduler, the most general scheduler model, the algorithm converges in at most \(12\varDelta m + \mathcal {O}(m+n)\) moves, where m is the number of edges, n is the number of nodes, and \(\varDelta \) is the maximum node degree. Regarding the move complexity, our algorithm outperforms the previously known best algorithm by a factor of \(\varDelta \). While the round complexity for the previous algorithm was unknown, we show a \(5\left\lfloor \frac{n}{p+1} \right\rfloor +5\) bound for our algorithm.

We study leader election in unidirectional rings of homonym processes that have no a priori knowledge on the number of processes. We show that message-terminating leader election is impossible for any class of rings \(\mathcal K_k\) with bounded multiplicity \(k \ge 2\). However, we show that process-terminating leader election is possible in the sub-class \(\mathcal U^* \cap \mathcal K_k\), where \(\mathcal U^*\) is the class of rings which contain a process with a unique label.

Trends in processor and system architecture, driven by power and complexity, point super-computing landscape toward very high and heterogeneous core count designs. As the number of cores inevitably increase traditional ways of performing large scale computations will also need to evolve from legacy models like OpenMP & MPI. However, such models are very much wedded to a control-flow vision of parallel programs, making it difficult to express asynchrony in programs. To address these challenges, codelet model was developed which is fine-grained, event-driven asynchronous program execution model. In-spite of its initial design goals, the Codelet model is rife with opportunities for further improvements to provide high-performance for both data and control regular applications. The major inspiration behind this work is to leverage the decades of research done to exploit instruction level parallelism (ILP) for the machine instructions inside codelet while build upon the dataflow software pipelining principals at codelet graph level to further enhance performance. In this paper, we propose hardware assisted extensions to the original codelet program execution model in order to implement efficient dataflow software pipelining and extend capabilities of the codelet model. This hardware-software co-design focuses on efficient implementation of data FIFO buffers leveraging proposed optimizations like - FIFO ring buffers and multiple-head FIFO buffers and single owner FIFO buffers to further exploit advantages of dataflow software pipelining. The wide range of scientific, machine learning and specially streaming applications should be able to take advantage of techniques proposed in this paper. Identifying these kernels and bench-marking them are the next anticipated steps for us.

Nova proposes a departure from today’s complex monolithic database management systems (DBMSs) as a service using the cloud. It advocates a server-less alternative consisting of a cloud of simple components that communicate using high speed networks. Nova will monitor the workload of an application continuously, configuring the DBMS to use the appropriate implementation of a component most suitable for processing the workload. In response to load fluctuations, it will adjust the knobs of a component to scale it to meet the performance requirements of the application. The vision of Nova is compelling because it adjusts resource usage, preventing either over-provisioning of resources that sit idle or over-utilized resources that yield a low performance, optimizing total cost of ownership. In addition to introducing Nova, this vision paper presents key research challenges that must be addressed to realize Nova. We explore two challenges in detail.

The term “Microservice Architecture” has sprung up in recent years as a new style of software design that gains popularity as cloud computing prospers. In microservice-based applications, different microservices collaborate with one another via interface calls, but they may also compete for resources when an increase of users’ need renders the resources insufficient. This poses new challenges for allocating resources efficiently during runtime. To tackle the problem, we propose a novel approach based on Congestion Game in this paper. Firstly, we use a weighted directed acyclic graph to model the inter-relationship of the microservices that compose an application. Then we use M/G/1 Queue in Queue Theory to describe the arrival process of access requests, and combine it with the above graph to calculate the arrival rate of access requests to each microservice, which in turn is used to estimate response time in a newly-designed microservice revenue function. Finally, we define resources competing problem as a congestion game where each microservice is a player aiming to maximize its revenue, and propose an algorithm to find Nash equilibrium in polynomial time. Experiment results show that our approach can effectively improve the overall performance of the system with limited resources, and outperform Binpack and Spread, two scheduling strategies used in Docker Swarm.

One important feature required by autonomous vehicles is the ability to perform a localization task in order to navigate in both known (urban, suburban, and highways) and unknown environments. Instead of relying on LIDAR technology, we propose to leverage a bio-inspired algorithm relying on more conventional cameras and a large neural network (NN) [2]. Yet, this approach must be able to scale. We propose to investigate the development of an FPGA-based solution. Due to the size of NN, dynamic partial reconfiguration will be required, and an efficient (software-based) scheduler must place the hardware tasks on multiple FPGA chip. We intend to implement this algorithm using a unique custom board, Wizarde, which embeds a 3 × 3 matrix of Zynq SoCs with high-end FPGAs to prototype a possible solution.

We describe a system for paper-cutting with a knife with a blade attached to the tip of the stylus on a drawing display. The purpose of this research is to support controlling cutting pressure for novices. Novices tend to cut paper with an unstable pressure stronger than necessary. Therefore, some instructors teach practicing to control the pressure to novices We have developed a device to support cutting practice by a stylus and drawing display. We measured the difference between pressures of novices and experts. In addition, we developed a system that encourages appropriate pressure based on expert pressure. This system shows the pressure difference between the user and the experts in color and sound. We experimented to compare the effectiveness of the system. As a result, the novices practiced with the system, the range of pressure and variation improved than the existing practice method.

This paper describes different methods used for inshore ship detection because of complex background. In this paper near sea shore ship detection methods are split into model based sea shore ship detection method and contour based sea shore ship detection method. In model based sea shore ship detection method is eliminate the interference of complex background and color. Contour based sea shore ship detection method is matching sea shore ship convex contour segments of the coastline which was extracted from image segmentation and edge detection. In this survey paper ten types of inshore ship detection methods are discussed.

Peer rewiring aims to update peer's contacts to other peers in a network. It is one of the classic approaches to semantic overlay networks(SON), the network organization where peers with semantically similar content are clustered together. To perform peer rewiring, each peer periodically collects the information of other peers in the network and uses them to update its existing contacts. While the information are collected either by random or by similarity-based gradient strategy in the literature, we propose a centrality-based approach to improve the efficiency of peer rewiring. The centrality is defined as peer's capability to access other peers in its neighborhood in SON. By integrating centrality in the similarity-based gradient strategy, more similar peers are expected to be collected in a shorter time. We test our proposal by adding new peers in an unstructured P2P network with more than 20000 peers, where similar peers are already clustered up. The proposed approach greatly improves the rewiring efficiency of the new peers, comparing to similarity-based gradient strategy(the state-of-the-art approach).

In a Vehicular Ad hoc NETwork (VANET), vehicles broadcast safety messages disclosing their trajectory information in order to warn drivers of impending accidents. Precise location information needed for these safety applications, combined with the need to exclude attackers through the use of authentication, creates a significant privacy risk. One method proposed to improve privacy is the use of many pseudonyms, and changing pseudonyms while in a mix zone where all other vehicles also change pseudonyms. Previous work has evaluated the effectiveness of mix zones using traces generated based on traffic theory. In this paper, we analyze the privacy obtainable from using mix zones in VANETs based on actual recordings of vehicle movements. We choose rank instead of entropy as our privacy metric because, as we will show, entropy is difficult to measure in our scenarios.

Stochastic Petri nets (SPNs) are widely used for the performance evaluation of computer and telecommunication systems. They inherit from their untimed version the capability of modeling parallel computations in a simple, graphical way. Simulation of SPNs is an important way to assess the performance of a system measured as throughput, response time or expected number of customers/resources in some places. Perfect sampling allows for the selection of the initial state of a simulation with a probability which corresponds to its stationary probability and hence the warm-up period is not required any more. In this paper we present performance tests of spnps, a tool based on some previous works that implements perfect sampling algorithm for SPNs by using decision diagrams. We test performance of the tool on a class of stochastic models of great importance in the quantitative evaluation of distributed systems and communication networks: the fork-join queueing networks.

As the Internet of Things (IoT) grows, the attacks to these devices continue to increase, both in number and impact. We identify the requirements for a real-time data processing framework that can be used to detect attacks while they occur, and propose an architecture that would implement these requirements. The proposed framework can be used by vendors, thirdparty organizations, and end-users to provide on-line monitoring and real-time attack mitigation. Finally, we identify the research challenges that we will face when implementing this architecture.

Domain-Specific Languages (DSLs) provide a high-level and domain-specific abstraction to describe algorithms within a certain domain concisely. Since a DSL separates the algorithm description from the actual target implementation, it offers a high flexibility among heterogeneous hardware targets, such as CPUs and GPUs. With the recent uprise of promising High-Level Synthesis (HLS) tools, like Vivado HLS and Altera OpenCL, FPGAs are becoming another attractive target architecture. Particularly in the domain of image processing, applications often come with stringent requirements regarding performance, energy efficiency, and power, for which FPGA have been proven to be among the most suitable architectures. In this work, we present the Hipacc framework, a DSL and source-to-source compiler for image processing. We show that domain knowledge can be captured to generate tailored implementations for C-based HLS from a common high-level DSL description targeting FPGAs. Our approach includes FPGA-specific memory architectures for handling point and local operators, as well as several high-level transformations. We evaluate our approach by comparing the resulting hardware accelerators to GPU implementations, generated from exactly the same DSL source code.

This article summarizes our current studies aiming at a better understanding of the energy consumption of a microprocessor during the execution of an application through a combination of theoretical results and experimental validations, The analysis of the transient thermal behavior and energy gains (ranging from 20 to 40% in some cases) via the adaptation of the clock frequency are of obvious practical interest. A general Passive Cooling Rule (PCR) for an isothermal object subjected to radiation, convection and internal heat generation is proposed. This power-temperature model is observed on an Exynos 5410 processor. Several approximations to this cooling rule are formulated for practical use, particularly online. They are accompanied by general rules for assessing when passive cooling becomes non-negligible compared to active cooling in embedded systems. On another hand, a theoretical framework for the existence of an Energy/Frequency Convexity Rule (EFCR) of program consumption is established. It is validated both by the state of the art and by experimental measurements. Power consumption models are then explained. The novelty of these models is that they take into account certain characteristics of the program execution and that they can be directly reused in any simulation for other processors of similar architecture.

Recently there is rapidly increase in bulk of data that formed global by networked and multi-user storage systems. It is very quick and therefore, there is concern about data security, which will prevent many users from transferring data. The ultimate re is to encrypt the data before they leave the owner. Deduplication is mostly used by number of cloud backup providers and also by different cloud services such as Drop box. But Encrypted data is pseudorandom and so it cannot be deduplicated. Data De-duplication is a technique for erasing the repeating data in a particular data set. In the de-duplication process, extra copies of the same data are removed by leaving only one copy which is stored. In order to protect the acquaintance of personal data which is sensitive enough that supports de-duplication. The convergent encryption technique is used to encipher the data before purchase or deploying from an outside agency.

Cloud platforms often take advantage of virtualization technology and make their actual hosts virtualized. As network attack events occur frequently, providing system security in a virtualized environment is the focus of this study. We have designed and implemented a lightweight network-based intrusion prevention system (IPS) named VMM-IPS for the virtual machine (VM) execution environment. To ensure the system safety of VMs and the host system at the same time, VMM-IPS is operated in the Linux kernel of the host system and co-located with the Kernel-based Virtual Machine that turns Linux kernel into a hypervisor. As packets enter the system, no matter destined to VMs or passing through the host, they are detected by VMM-IPS. Unlike user-level IPS that needs switching protection domain and copying packets to user buffer for inspection, VMM-IPS is more efficient because of the capability to perform in-place packet inspection. It adopts signature-based detection and is implemented with the multiple-pattern search algorithm AC-BM for efficient string matching. Besides, VMM-IPS can protect the system against attacks using packet splitting and reassembly to evade introduction detection system (IDS). The experimental results demonstrate VMM-IPS can achieve system safety effectively and efficiently.

We present a unique approach for integrating research in High Performance Computing (HPC) as well as photovoltaic (PV) solar farming and battery technologies into a container-based compute center designed for a maximum of energy efficiency, performance and extensibility/scalability. We use NVIDIA Jetson TK1 boards to build a considerably dimensioned cluster of 60 low-power GPUs, attach a 7.5 kWp solar farm and a 8 kWh Lithium-Ion battery power supply and integrate everything into a single-container, standalone housing. We demonstrate the success of our system by evaluating the performance and energy efficiency for common versatile dense and sparse linear algebra kernels as well as a full CFD code. By this work we can show, that with current technology, energy consumption-induced follow-up cost of HPC can be reduced to zero.

Most prominent cryptocurrencies utilize proof of work (PoW) to secure their operation, yet PoW suffers from two key undesirable properties. First, the work done is generally wasted, not useful for anything but the gleaned security of the cryptocurrency. Second, PoW is naturally outsourceable, leading to inegalitarian concentration of power in the hands of few so-called pools that command large portions of the system’s computation power.

SpaceWire is integrated networking technology where the data packets and control traffic in the form of hard real-time signals are transmitted throw the same links of communication network. Last decades for hard real-time signals transmission in onboard networks the specialized buses like CAN and MIL-STD 155b and sideband signals were used. All buses have the property that the signal from the source node to all other nodes propagates simultaneously, so there is no jitter for signal delivery. When the industry moves to a common communication medium for control and data traffic transmission through network with routing switches and point-to-point connections, there is a problem of synchronization hard real-time signals delivery from the source to all other nodes because of different distance (number of switches and link speeds in a path) between source and receiver. There are many different algorithms and methods for the system time synchronization. In this paper we consider the questions of delivery synchronization of distributed interrupts which intended to control and inform the devices about critical system events in hard real-time.

Trust is the cornerstone of success in any relationship between two or more parties. Current technologies to include cloud computing, social networking, and mobile applications, coupled with the explosion in storage and processing power, are evolving large-scale marketplaces for a wide variety of resources and services. In such marketplaces, users (consumer, provider and broker) are largely autonomous with vastly diverse requirements, capabilities, and trust profiles, and thus trustworthy interactions and transactions become a challenging endeavor. We hypothesize the need for a dynamically customizable personalized trust management system for the robustness and wide-scale adoption of such large-scale heterogeneous marketplaces. In this paper, we explore and enable trust personalization and propose PETRA, PErsonalized TRust Architecture, and present a PETRA-based trust management system. Our simulation results demonstrate the effectiveness and efficiency of the PETRA-based approach and illustrate the impact of trust personalization.

We propose a heuristic method that generates a graph for order/degree problem. Target graphs of our heuristics have large order (> 4000) and diameter 3. We describe the observation of smaller graphs and basic structure of our heuristics. We also explain an evaluation function of each edge for efficient 2-opt local search. Using them, we found the best solutions for several graphs.

Mobile ad-hoc network is a collection of mobile nodes which forms an instant network without fixed topology and can be arranged dynamically. Energy saving and security are important issue in MANET. Network coding technique is used to reduce energy consumption by less transmissions in MANET. To achieve a security, there are many encryption scheme are available. Out of which p-coding technique is lightweight encryption scheme which provides confidentiality. P-coding is to let the source randomly permute the symbol of each packet. So eavesdropper can't obtain the meaningful information without knowing the permutation encryption function and coding vector.

Cloud computing is the growing technology. Each organization wants to connect to the cloud computing environment. But some of them resist connecting due to security issues. Various security factors had been raised, some also lead to include third party auditor for solving security issues. Another major factor in concern is the Trust on the cloud service provider's i.e., the level to which an organization can trust the cloud service provider for handling their company's data.

Studies in the optimization of sequence alignment have been carried out in bioinformatics. In this paper, we have focused on two aspects: memory usage and execution time. Our study suggests that cache memory does not have a significant effect on system performance. Our attention then turns to optimize Smith—Waterman's algorithm. Two instruction level methods have been proposed and 2—8 fold speed improvements have been observed after the optimization has been implemented. Further improvements on system performance have been achieved by overlapping computation with system I/O usage.

A promising approach for treating tissue or organ failure involves the use of bioartificial tissue substitutes grown in scaffolds with appropriate structure and shape. Currently, however, the engineering of tissue substitutes is a long and costly process based exclusively on experimentation. Predictive computer models can greatly reduce the development costs of tissue-engineered therapies by enabling scientists to rapidly evaluate the effect of system parameters on the growth rates and quality of regenerated tissues. We report here the parallel implementation of a three-dimensional model that employs cellular automata to describe the dynamic behavior of a population of mammalian cells that migrate, interact and proliferate to generate new tissues. The simulator uses MPI for interprocessor communication and is suitable for distributed memory multi-computers. Three parallel algorithms are developed to approximate the sequential algorithm describing this dynamic process of tissue growth. The parallel algorithms progressively relax the correctness requirements using different approaches to handle the cells that either move/ divide in the boundary layers of processors or cross sub-domain boundaries. Finally, a systematic study is carried out to evaluate the accuracy and performance of these algorithms.

Today's mobile devices are packed with sensors that are capable of gathering rich contextual information, such as location, wireless device signatures, ambient noise, and photographs. This paper exhorts the security community to re-design authentication mechanisms for users on mobile devices. Instead of relying on one simplistic, worst-case threat model, we should use contextual information to develop more nuanced models that assess the risk level of the user's current environment. This would allow us to decrease or eliminate the level of user interaction required to authenticate in some situations, improving usability without any effective impact on security. Ideally, authentication mechanisms will scale up or down to match users' own mental threat models of their environments. We sketch out several scenarios demonstrating how contextual information can be used to assess risks and adapt authentication mechanisms. This is a research-rich area, and we outline future research directions for developing and evaluating dynamic security mechanisms using contextual information.

Edge computing promises to bring computation close to the end-users to support emergent applications such as virtual reality. However, the computational capacity at the edge of the network is currently limited. To become a pervasive paradigm, edge computing needs highly dispersed decentralized deployments, that, contrary to cloud, cannot benefit from economies of scale. In this situation, crowdsourcing appears attractive - there are plenty of computing devices at the disposal of the general public, and these devices are located exactly where computing power is needed the most - at the edge of the network. Crowdsourcing has been a success maker for scientific computing projects, e.g., SETI@home, or distributed ledger systems empowering decentralized finance. However, as of now, there is no crowdsourced system that addresses the needs of edge computing. In this position paper, we aim to identify the causes of this shortcoming, analyze the potential ways to overcome it, and outline future directions.

This position paper argues that earphones hold the potential for major disruptions in mobile, wearable computing. The early signs are positive and the industrial wheels are in motion. However, whether earphones truly become a disruptive new platform, or stop at being a useful accessory, could depend on whether we -- the mobile computing researchers -- deliver. If we do, tomorrow's earphones will run augmented reality via 3D sound, will have Alexa and Siri whisper just-in-time information, will track our motion and health, will make authentications seamless, and much more. The leap from today's earphones to "earables" could mimic the transformation we saw from basic-phones to smart phones. On the other hand, if we are unable to provide some of the disruptive building blocks, tomorrow's earphones may saturate in its capabilities. We believe this is an important juncture in time where the mobile computing research community has an opportunity to shape the future. This paper aims to discuss this landscape, including some challenges, opportunities, and applications.

Background Rabies is an acute infectious disease of the central nervous system caused by the rabies virus. The mortality rate of rabies is almost 100%. For some countries with poor sanitation, the spread of rabies among dogs is very serious. Objective The objective of this paper was to study the ecological transmission mode of rabies to make theoretical contributions to the suppression of rabies in China. Methods A mathematical model of the transmission mode of rabies was constructed using relevant data from the literature and officially published figures in China. Using this model, we fitted the data of the number of patients with rabies and predicted the future number of patients with rabies. In addition, we studied the effectiveness of different rabies suppression measures. Results The results of the study indicated that the number of people infected with rabies will rise in the first stage, and then decrease. The model forecasted that in about 10 years, the number of rabies cases will be controlled within a relatively stable range. According to the prediction results of the model reported in this paper, the number of rabies cases will eventually plateau at approximately 500 people every year. Relatively effective rabies suppression measures include controlling the birth rate of domestic and wild dogs as well as increasing the level of rabies immunity in domestic dogs. Conclusions The basic reproductive number of rabies in China is still greater than 1. That is, China currently has insufficient measures to control rabies. The research on the transmission mode of rabies and control measures in this paper can provide theoretical support for rabies control in China.

Effective task scheduling in heterogeneous computing systems is a very challenging and crucial task. Inter process communication and the heterogeneity of resources plays an important role in task scheduling. To achieve the efficiency tasks are assigned to best suited processor while minimizing communication cost. This directly increases the performance and is referred as completion time. Such problems in a distributed system are considered as NP hard problems. Many solutions are proposed in literature for solving this issue. Directed Acyclic Graph (DAG) is also used to solve the issue of performance in distributed networks. A new heuristic is in this paper based on DAG is proposed for task scheduling on tasks order, average communication cost and best available processor/resource. The experimental study shows the proposed approach give promising results. The performance of proposed heuristic is illustrated by comparing the schedule time, efficiency and schedule length with other well-known algorithms for task scheduling.

Filariasis is one of the major public health concerns in India. Approximately 600 million people spread across 250 districts of India are at risk of filariasis. To predict this disease, a pilot scale study was carried out in 30 villages of Karimnagar district of Telangana from 2004 to 2007 to collect epidemiological and socio-economic data. The collected data are analysed by employing various machine learning techniques such as Naïve Bayes (NB), logistic model tree, probabilistic neural network, J48 (C4.5), classification and regression tree, JRip and gradient boosting machine. The performances of these algorithms are reported using sensitivity, specificity, accuracy and area under ROC curve (AUC). Among all employed classification methods, NB yielded the best AUC of 64% and was equally statistically significant with the rest of the classifiers. Similarly, the J48 algorithm generated 23 decision rules that help in developing an early warning system to implement better prevention and control efforts in the management of filariasis.

This is a demo for a full paper in CHI'17 proceedings http://dx.doi.org/10.1145/3025453.3025879 [6]. We designed and implemented a secure, robust and usable multi-expert recognizer, Garda, for gesture passwords. This demo demonstrates two real scenarios of using Garda as a gesture recognizer in mobile devices: under normal use and under a shoulder-surfing attack. As a supplement to the demo, we describe technical details of Garda in this report. Our demo shows the usability of our Garda gesture authentication method on mobile devices in daily life.

Awareness about one's own cognitive state is a critical first step towards better physical, physiological, psychological, behavioral, and social health. When users are aware of their mental states - particularly those which might be detrimental to their health (e.g. high stress, low excitement) - they can take the necessary steps to either alter their behavior, or find ways to effectively cope with it. From a technology viewpoint, it is thus important to find ways to detect a user's cognitive state as unobtrusively as possible, and thereafter either inform or assist the users in coping with it. Through this position paper, we wish to discuss the potential of using mobile context data gathered from a user's smartphone to infer the user's cognitive state, and thereafter using mobile notifications to deliver timely intervention messages.

In Hei Longjiang Province of China, ten million tons of soybean straws are produced every year, most of which are burned or discarded. The comprehensive utilization rate is less than 5 %. However, as a kind of high-quality raw material resources, the soybean straw is the main material when producing fuel ethanol and biodiesel. The traditional chemical detection technology was always applied to test the quality of straws’ internal components, which needs long time and high cost. Therefore, this research adopted near-infrared spectroscopy technology to establish models to achieve rapid detection of straw’s internal components. One hundred and fifty-two samples of straw were collected within Heilongjiang Province, and the main ingredients, lignin and hemicellulose, were chosen as the research objects. K-stone method is used to divide into calibration and validation sets. The correlation coefficient is adopted to select and validate characteristic bands. After removal of the abnormal sample by concentration of residual method, the emphasis is put on the effect of removal of spectrum noise by applying the wavelet transform. The experimental results are that the model has good stability under the full spectrum band 4000–12,000 cm−1. After the removal of 3 lignin and 3 hemicellulose samples, respectively, the R 2 of calibration set of lignin and hemicellulose models increased significantly, from 0.5836436 and 0.4994598 to 0.6994097 and 0.6943559, respectively. However, after the 5-layer decomposition of DB2 wavelet, the R 2 increased to 0.8075574 and 0.8214309 separately. The validation set’s root mean square error prediction (RMSEP) fell from 0.7738772 and 0.3069899 to 0.5979685 and 0.2761462. After the conversion, the relative standard deviations (RSD) are 1.92 and 1.86 %, respectively. Unknown samples can be real-time predicted in short times after building the model, and the prediction process needs only 1960s. The experiments show that it is feasible to employ near-infrared spectroscopy method for rapid detection of straw components, which provides a new method to detect the straw ingredients rapidly in the future.

Background Emotional state in everyday life is an essential indicator of health and well-being. However, daily assessment of emotional states largely depends on active self-reports, which are often inconvenient and prone to incomplete information. Automated detection of emotional states and transitions on a daily basis could be an effective solution to this problem. However, the relationship between emotional transitions and everyday context remains to be unexplored. Objective This study aims to explore the relationship between contextual information and emotional transitions and states to evaluate the feasibility of detecting emotional transitions and states from daily contextual information using machine learning (ML) techniques. Methods This study was conducted on the data of 18 individuals from a publicly available data set called ExtraSensory. Contextual and sensor data were collected using smartphone and smartwatch sensors in a free-living condition, where the number of days for each person varied from 3 to 9. Sensors included an accelerometer, a gyroscope, a compass, location services, a microphone, a phone state indicator, light, temperature, and a barometer. The users self-reported approximately 49 discrete emotions at different intervals via a smartphone app throughout the data collection period. We mapped the 49 reported discrete emotions to the 3 dimensions of the pleasure, arousal, and dominance model and considered 6 emotional states: discordant, pleased, dissuaded, aroused, submissive, and dominant. We built general and personalized models for detecting emotional transitions and states every 5 min. The transition detection problem is a binary classification problem that detects whether a person’s emotional state has changed over time, whereas state detection is a multiclass classification problem. In both cases, a wide range of supervised ML algorithms were leveraged, in addition to data preprocessing, feature selection, and data imbalance handling techniques. Finally, an assessment was conducted to shed light on the association between everyday context and emotional states. Results This study obtained promising results for emotional state and transition detection. The best area under the receiver operating characteristic (AUROC) curve for emotional state detection reached 60.55% in the general models and an average of 96.33% across personalized models. Despite the highly imbalanced data, the best AUROC curve for emotional transition detection reached 90.5% in the general models and an average of 88.73% across personalized models. In general, feature analyses show that spatiotemporal context, phone state, and motion-related information are the most informative factors for emotional state and transition detection. Our assessment showed that lifestyle has an impact on the predictability of emotion. Conclusions Our results demonstrate a strong association of daily context with emotional states and transitions as well as the feasibility of detecting emotional states and transitions using data from smartphone and smartwatch sensors.

Future urban scenarios will be characterized by the close integration of 'CT devices and humans. Citizens using their own capabilities integrated with 'CT technologies could collaboratively constitute a large-scale socio-technical superorganism to support collective urban awareness and activities. This position paper, with the help of a representative case study in the area of intelligent transportation systems, identifies the key challenges for future urban superorganisms and proposes a two-tier architecture to support their development.

International audience; As a result of the recent explosion of sensor-equipped mobile phone market, the phenomenal growth of Internet and social network users, and the large deployment of sensor network in public facilities, private buildings and outdoor environments, the "digital footprints" left by people while interacting with cyber-physical spaces are accumulating with an unprecedented breadth, depth and scale. The technology trend towards pervasive sensing and large-scale social and community computing is making "ocial and community intelligence (SCI)", a new research area take shape, that aims at mining the "digital footprints" to reveal the patterns of individual, group and societal behaviours. It is believed that the SCI technology has the potential to revolutionize the field of context-aware computing. The aim of this position paper is to identify this emerging research area, present the research background and some references to the relevant research fields, define the general system framework, predict some potential application areas, and propose some initial thoughts about the future research issues and challenges in social and community intelligence

Current and anticipated storage environments confront domain scientist and data center operators with usability, performance and cost challenges. The amount of data upcoming system will be required to handle is expected to grow exponentially, mainly due to increasing resolution and affordable compute power. Unfortunately, the relationship between cost and performance is not always well understood requiring considerable effort for educated procurement. Within the Centre of Excellence in Simulation of Weather and Climate in Europe (ESiWACE) models to better understand cost and performance of current and future systems are being explored. This paper presents models and methodology focusing on, but not limited to, data centers used in the context of climate and numerical weather prediction. The paper concludes with a case study of alternative deployment strategies and outlines the challenges anticipating their impact on cost and performance. By publishing these early results, we would like to make the case to work towards standard models and methodologies collaboratively as a community to create sufficient incentives for vendors to provide specifications in formats which are compatible to these modeling tools. In addition to that, we see application for such formalized models and information in I/O re lated middleware, which are expected to make automated but reasonable decisions in increasingly heterogeneous data centers.

This paper introduces hybrid LU-QR algorithms for solving dense linear systems of the form A x = b . Throughout a matrix factorization, these algorithms dynamically alternate LU with local pivoting and QR elimination steps based upon some robustness criterion. LU elimination steps can be very efficiently parallelized, and are twice as cheap in terms of floating-point operations, as QR steps. However, LU steps are not necessarily stable, while QR steps are always stable. The hybrid algorithms execute a QR step when a robustness criterion detects some risk for instability, and they execute an LU step otherwise. The choice between LU and QR steps must have a small computational overhead and must provide a satisfactory level of stability with as few QR steps as possible. In this paper, we introduce several robustness criteria and we establish upper bounds on the growth factor of the norm of the updated matrix incurred by each of these criteria. In addition, we describe the implementation of the hybrid algorithms through an extension of the PaRSEC software to allow for dynamic choices during execution. Finally, we analyze both stability and performance results compared to state-of-the-art linear solvers on parallel distributed multicore platforms. A comprehensive set of experiments shows that hybrid LU-QR algorithms provide a continuous range of trade-offs between stability and performances. New hybrid algorithm combining stability of QR and efficiency of LU factorizations.Flexible threshold criteria to select LU and QR steps.Comprehensive experimental bi-criteria study of stability and performance.

International audience; Distributed heterogeneous computing systems escalate the problem of choosing the appropriate programming model. Programming models such as message passing are efficient but require low-level management of communications. Higher level of programming such as shared memory are convenient for the application design but they usually have performance issues. With the recent development of distributed heterogeneous systems and new protocols to access remote memories, there is an opportunity for distributed shared memory systems to offer a satisfying level of abstraction while not giving up on performance. In this paper a video processing application is written using MPI, 0MQ and an in-house software-distributed shared memory (S-DSM) backend and deployed over a set of heterogeneous computing boards. Results show that 0MQ implementation is the most efficient but at the price of writing the application with the targeted platform in mind. The S-DSM implementation runs up to 2 times faster than the pure OpenMPI implementation and competes with 0MQ when the data granularity is small.

Live migration is an essential feature in virtual infrastructure and cloud computing datacenters. Using live migration, virtual machines can be online migrated from a physical machine to another with negligible service interruption. Load balance, power saving, dynamic resource allocation, and high availability algorithms in virtual data-centers and cloud computing environments are dependent on live migration. Live migration process has six phases that result in live migration overhead. Currently, virtual datacenters admins run live migrations without an idea about the migration cost prediction and without recommendations about the optimal timing for initiating a VM live migration especially for large memory VMs or for concurrently multiple VMs migration. Without cost prediction and timing optimization, live migration might face longer duration, network bottlenecks and migration failure in some cases. The previously proposed timing optimization approach is based on using machine learning for live migration cost prediction and the network utilization predict ion of the cluster. In this paper, we show how to integrate our machine learning based timing optimization algorithm with VMware vSphere. This integration deployment proves the practicality of the proposed algorithm by presenting the building blocks of the tools and backend scripts that should run to implement this timing optimization feature. The paper shows also how the IT admins can make use of this novel cost prediction and timing optimization option as an integrated plug-in within VMware vSphere UI to be notified with the optimal timing recommendation in case of a having live migration request.

Every day billions of people access web sites, blogs, and social media. Often they use their mobile devices and produce huge amount of data that can be effectively exploited for extracting valuable information concerning human dynamics and behaviors. Such data, commonly referred as Big Data, contains rich information about user activities, interests, and behaviors, which makes it intrinsically suited to a very large set of applications. For getting valuable information and knowledge from such data in a reasonable time, novel scalable frameworks and data analysis techniques on Cloud systems have been developed. This paper aims at describing some recent Cloud-based frameworks and methodologies for Big Data processing that can be used for developing and executing several data analysis applications, including trajectory mining and sentiment analysis. The paper is organized in two main parts. The first part focuses on tools for developing and executing scalable data analysis applications on Clouds. The second part presents data analysis methodologies for extracting knowledge from large datasets.