Intelligent agents for fault tolerance: from multi-agent simulation to cluster-based implementation

Recent research in multi-agent systems incorporate fault tolerance concepts, but does not explore the extension and implementation of such ideas for large scale parallel computing systems. The work reported in this paper investigates a swarm array computing approach, namely 'Intelligent Agents'. A task to be executed on a parallel computing system is decomposed to sub-tasks and mapped onto agents that traverse an abstracted hardware layer. The agents intercommunicate across processors to share information during the event of a predicted core/processor failure and for successfully completing the task. The feasibility of the approach is validated by simulations on an FPGA using a multi-agent simulator, and implementation of a parallel reduction algorithm on a computer cluster using the Message Passing Interface.

Pipelining considerations for an FPGA case

This paper presents a semi-synchronous pipeline scheme, here referred as single-pulse pipeline, to the problem of mapping pipelined circuits to a Field Programmable Gate Array (FPGA). Area and timing considerations are given for a general case and later applied to a systolic circuit as illustration. The single-pulse pipeline can manage asynchronous worst-case data completion and it is evaluated against two chosen asynchronous pipelining: a four-phase bundle-data pipeline and a doubly-latched asynchronous pipeline. The semi-synchronous pipeline proposal takes less FPGA area and operates faster than the two selected fully-asynchronous schemes for an FPGA case.

Cursor relocation techniques to help older adults find 'lost' cursors

Older adult computer users often lose track of the mouse cursor and so resort to methods such as mouse shaking or searching the screen to find the cursor again. Hence, this paper describes how a standard optical mouse was modified to include a touch sensor, activated by releasing and touching the mouse, which automatically centers the mouse cursor to the screen, potentially making it easier to find a 'lost' cursor. Six older adult computer users and six younger computer users were asked to compare the touch sensitive mouse with cursor centering with two alternative techniques for locating the mouse cursor: manually shaking the mouse and using the Windows sonar facility. The time taken to click on a target following a distractor task was recorded, and results show that centering the mouse was the fastest to use, with a 35% improvement over shaking the mouse. Five out of six older participants ranked the touch sensitive mouse with cursor centering as the easiest to use.

Multiple haptic targets for motion-impaired computer users

Although a number of studies have reported that force feedback gravity wells can improve performance in "point-and-click" tasks, there have been few studies addressing issues surrounding the use of gravity wells for multiple on-screen targets. This paper investigates the performance of users, both with and without motion-impairments, in a "point-and-click" task when an undesired haptic distractor is present. The importance of distractor location is studied explicitly. Results showed that gravity wells can still improve times and error rates, even on occasions when the cursor is pulled into a distractor. The greatest improvement is seen for the most impaired users. In addition to traditional measures such as time and errors, performance is studied in terms of measures of cursor movement along a path. Two cursor measures, angular distribution and temporal components, are proposed and their ability to explain performance differences is explored.

A multi-agent system for pervasive healthcare

Wireless technology based pervasive healthcare has been proposed in many applications such as disease management and accident prevention for cost saving and promoting citizen’s wellbeing. However, the emphasis so far is on the artefacts with limited attentions to guiding the development of an effective and efficient solution for pervasive healthcare. Therefore, this paper aims to propose a framework of multi-agent systems design for pervasive healthcare by adopting the concept of pervasive informatics and using the methods of organisational semiotics. The proposed multi-agent system for pervasive healthcare utilises sensory information to support healthcare professionals for providing appropriate care. The key contributions contain theoretical aspect and practical aspect. In theory, this paper articulates the information interactions between the pervasive healthcare environment and stakeholders by using the methods of organisational semiotics; in practice, the proposed framework improves the healthcare quality by providing appropriate medical attentions when and as needed. In this paper, both systems and functional architecture of the multi-agent system are elaborated with the use of wireless technologies such as RFID and wireless sensor networks. The future study will focus on the implementation of the proposed framework.

Non-uniform data distribution for communication-efficient parallel clustering

Global communicationrequirements andloadimbalanceof someparalleldataminingalgorithms arethe major obstacles to exploitthe computational power of large-scale systems. This work investigates how non-uniform data distributions can be exploited to remove the global communication requirement and to reduce the communication costin parallel data mining algorithms and, in particular, in the k-means algorithm for cluster analysis. In the straightforward parallel formulation of the k-means algorithm, data and computation loads are uniformly distributed over the processing nodes. This approach has excellent load balancing characteristics that may suggest it could scale up to large and extreme-scale parallel computing systems. However, at each iteration step the algorithm requires a global reduction operationwhichhinders thescalabilityoftheapproach.Thisworkstudiesadifferentparallelformulation of the algorithm where the requirement of global communication is removed, while maintaining the same deterministic nature ofthe centralised algorithm. The proposed approach exploits a non-uniform data distribution which can be either found in real-world distributed applications or can be induced by means ofmulti-dimensional binary searchtrees. The approachcanalso be extended to accommodate an approximation error which allows a further reduction ofthe communication costs. The effectiveness of the exact and approximate methods has been tested in a parallel computing system with 64 processors and in simulations with 1024 processing element

Interactive sensory objects for improving access to heritage

In this project we explore how to enhance the experience and understanding of cultural heritage in museums and heritage sites by creating interactive multisensory objects collaboratively with artists, technologists and people with learning disabilities. We focus here on workshops conducted during the first year of a three year project in which people with learning disabilities each constructed a 'sensory box' to represent their experiences of Speke Hall, a heritage site in the UK. The box is developed further in later workshops which explore aspects of physicality and how to appeal to the entire range of senses, making use of Arduino technology and basic sensors to enable an interactive user experience.

Cursor relocation techniques to help older adults find ‘lost’ cursors

Older adult computer users often lose track of the mouse cursor and so resort to methods such as shaking the mouse or searching the entire screen to find the cursor again. Hence, this paper describes how a standard optical mouse was modified to include a touch sensor, activated by releasing and touching the mouse, which automatically centers the mouse cursor to the screen, potentially making it easier to find a ‘lost’ cursor. Six older adult computer users and six younger computer users were asked to compare the touch sensitive mouse with cursor centering with two alternative techniques for locating the mouse cursor: manually shaking the mouse and using the Windows sonar facility. The time taken to click on a target after a distractor task was recorded, and results show that centering the mouse was the fastest to use, with a 35% improvement over shaking the mouse. Five out of six older participants ranked the touch sensitive mouse with cursor centering as the easiest to use.

Making electronics more accessible to people with learning disabilities

We extended 'littleBits' electronic components by attaching them to a larger base that was designed to help make them easier to pick up and handle, and easier to assemble into circuits for people with learning disabilities. A pilot study with a group of students with learning disabilities was very positive. There were fewer difficulties in assembling the components into circuits, and problems such as attempting to connect them the wrong way round or the wrong way up were eliminated completely.

Online persuasion for e-commerce websites

The persuasive design of e-commerce websites has been shown to support people with online purchases. Therefore, it is important to understand how persuasive applications are used and assimilated into e-commerce website designs. This paper demonstrates how the PSD model’s persuasive features could be used to build a bridge supporting the extraction and evaluation of persuasive features in such e-commerce websites; thus practically explaining how feature implementation can enhance website persuasiveness. To support a deeper understanding of persuasive e-commerce website design, this research, using the Persuasive Systems Design (PSD) model, identifies the distinct persuasive features currently assimilated in ten successful e-commerce websites. The results revealed extensive use of persuasive features; particularly features related to dialogue support, credibility support, and primary task support; thus highlighting weaknesses in the implementation of social support features. In conclusion we suggest possible ways for enhancing persuasive feature implementation via appropriate contextual examples and explanation.

Fast adaptive real-time classification for data streams with concept drift

An important application of Big Data Analytics is the real-time analysis of streaming data. Streaming data imposes unique challenges to data mining algorithms, such as concept drifts, the need to analyse the data on the fly due to unbounded data streams and scalable algorithms due to potentially high throughput of data. Real-time classification algorithms that are adaptive to concept drifts and fast exist, however, most approaches are not naturally parallel and are thus limited in their scalability. This paper presents work on the Micro-Cluster Nearest Neighbour (MC-NN) classifier. MC-NN is based on an adaptive statistical data summary based on Micro-Clusters. MC-NN is very fast and adaptive to concept drift whilst maintaining the parallel properties of the base KNN classifier. Also MC-NN is competitive compared with existing data stream classifiers in terms of accuracy and speed.

Scalable and fault tolerant failure detection and consensus

Future extreme-scale high-performance computing systems will be required to work under frequent component failures. The MPI Forum's User Level Failure Mitigation proposal has introduced an operation, MPI_Comm_shrink, to synchronize the alive processes on the list of failed processes, so that applications can continue to execute even in the presence of failures by adopting algorithm-based fault tolerance techniques. This MPI_Comm_shrink operation requires a fault tolerant failure detection and consensus algorithm. This paper presents and compares two novel failure detection and consensus algorithms. The proposed algorithms are based on Gossip protocols and are inherently fault-tolerant and scalable. The proposed algorithms were implemented and tested using the Extreme-scale Simulator. The results show that in both algorithms the number of Gossip cycles to achieve global consensus scales logarithmically with system size. The second algorithm also shows better scalability in terms of memory and network bandwidth usage and a perfect synchronization in achieving global consensus.

Connectivity recovery in epidemic membership protocols

Epidemic protocols are a bio-inspired communication and computation paradigm for extreme-scale network system based on randomized communication. The protocols rely on a membership service to build decentralized and random overlay topologies. In a weakly connected overlay topology, a naive mechanism of membership protocols can break the connectivity, thus impairing the accuracy of the application. This work investigates the factors in membership protocols that cause the loss of global connectivity and introduces the first topology connectivity recovery mechanism. The mechanism is integrated into the Expander Membership Protocol, which is then evaluated against other membership protocols. The analysis shows that the proposed connectivity recovery mechanism is effective in preserving topology connectivity and also helps to improve the application performance in terms of convergence speed.

En beskrivning av informationsmiljö därdatumstyrda förhållanden råder : En fallstudie av en myndighet och dess informationsmiljö med rättsakter

In this study a case study was conducted at the Transport Agency. The government agency is seeking better alternatives to its current ways of keeping track of important dates stated in legal documents. The agency wants to explore the possibility for an IT-solution for keeping track on those dates. In the case study change analysis was conducted on the vehicle type approval process. Qualitative interviews were conducted with a section manager, vehicle type and component approvers, and with a system administrator and a system manager at the Transport Agency in Borlänge. The study describes an information environment in an organization in which date-controlled conditions apply and challenges with such an environment.The study also provides a proposal for designing an IT-solution for such an environment. The result that emerged showed that the information environment in terms of the legal documents and the institutions’ (EU and UNECE) websites where they are distributed is challenging for keeping track on dates. The challenges of the websites are that they are limited and complicated in terms of information search. The challenges with the legal documents are that date-controlled conditions apply. In other words, the dates themselves determine which dates and what applies and not. The legal documents are also many, retroactive, refer to each other and contain regulations for different kinds of areas, which makes difficult to find the relevant dates. The design proposal that emerged was made based on the existing IT environment at the Transport Agency. It was proposed that important key concepts are identified, categorized and realized in a database with searchability on the most important common concepts as denominators, and that it the information conforms with E-message. The developed database was based on date, date type, vehicle category, legal document, minor version and other info. A web interface was created in ASP.NET and C # for access to the database, with the message that a webbbased IT-solution should be considered.