iLectures - Exploring the interactive lecture using web-based broadcasting technology

The continued use of traditional lecturing across Higher Education as the main teaching and learning approach in many disciplines must be challenged. An increasing number of studies suggest that this approach, compared to more active learning methods, is the least effective. In counterargument, the use of traditional lectures are often justified as necessary given a large student population. By analysing the implementation of a web based broadcasting approach which replaced the traditional lecture within a programming-based module, and thereby removed the student population rationale, it was hoped that the student learning experience would become more active and ultimately enhance learning on the module. The implemented model replaces the traditional approach of students attending an on-campus lecture theatre with a web-based live broadcast approach that focuses on students being active learners rather than passive recipients. Students ‘attend’ by viewing a live broadcast of the lecturer, presented as a talking head, and the lecturer’s desktop, via a web browser. Video and audio communication is primarily from tutor to students, with text-based comments used to provide communication from students to tutor. This approach promotes active learning by allowing student to perform activities on their own computer rather than the passive viewing and listening common encountered in large lecture classes. By analysing this approach over two years (n = 234 students) results indicate that 89.6% of students rated the approach as offering a highly positive learning experience. Comparing student performance across three academic years also indicates a positive change. A small data analytic analysis was conducted into student participation levels and suggests that the student cohort's willingness to engage with the broadcast lectures material is high.

Mitigating the Impact of Faults in Unreliable Memories for Error-Resilient Applications

Inherently error-resilient applications in areas such as signal processing, machine learning and data analytics provide opportunities for relaxing reliability requirements, and thereby reducing the overhead incurred by conventional error correction schemes. In this paper, we exploit the tolerable imprecision of such applications by designing an energy-efficient fault-mitigation scheme for unreliable data memories to meet target yield. The proposed approach uses a bit-shuffling mechanism to isolate faults into bit locations with lower significance. This skews the bit-error distribution towards the low order bits, substantially limiting the output error magnitude. By controlling the granularity of the shuffling, the proposed technique enables trading-off quality for power, area, and timing overhead. Compared to error-correction codes, this can reduce the overhead by as much as 83% in read power, 77% in read access time, and 89% in area, when applied to various data mining applications in 28nm process technology.

Protocol for a pilot randomised controlled trial of an intervention to increase the use of traffic light food labelling in UK shoppers (the FLICC trial)

Background: Traffic light labelling of foods—a system that incorporates a colour-coded assessment of the level of total fat, saturated fat, sugar and salt on the front of packaged foods—has been recommended by the UK Government and is currently in use or being phased in by many UK manufacturers and retailers. This paper describes a protocol for a pilot randomised controlled trial of an intervention designed to increase the use of traffic light labelling during real-life food purchase decisions.

Methods/design: The objectives of this two-arm randomised controlled pilot trial are to assess recruitment, retention and data completion rates, to generate potential effect size estimates to inform sample size calculations for the main trial and to assess the feasibility of conducting such a trial. Participants will be recruited by email from a loyalty card database of a UK supermarket chain. Eligible participants will be over 18 and regular shoppers who frequently purchase ready meals or pizzas. The intervention is informed by a review of previous interventions encouraging the use of nutrition labelling and the broader behaviour change literature. It is designed to impact on mechanisms affecting belief and behavioural intention formation as well as those associated with planning and goal setting and the adoption and maintenance of the behaviour of interest, namely traffic light label use during purchases of ready meals and pizzas. Data will be collected using electronic sales data via supermarket loyalty cards and web-based questionnaires and will be used to estimate the effect of the intervention on the nutrition profile of purchased ready meals and pizzas and the behavioural mechanisms associated with label use. Data collection will take place over 48 weeks. A process evaluation including semi-structured interviews and web analytics will be conducted to assess feasibility of a full trial.

Discussion: The design of the pilot trial allows for efficient recruitment and data collection. The intervention could be generalised to a wider population if shown to be feasible in the main trial.

DocLite: A Docker-Based Lightweight Cloud Benchmarking Tool

Existing benchmarking methods are time consuming processes as they typically benchmark the entire Virtual Machine (VM) in order to generate accurate performance data, making them less suitable for real-time analytics. The research in this paper is aimed to surmount the above challenge by presenting DocLite - Docker Container-based Lightweight benchmarking tool. DocLite explores lightweight cloud benchmarking methods for rapidly executing benchmarks in near real-time. DocLite is built on the Docker container technology, which allows a user-defined memory size and number of CPU cores of the VM to be benchmarked. The tool incorporates two benchmarking methods - the first referred to as the native method employs containers to benchmark a small portion of the VM and generate performance ranks, and the second uses historic benchmark data along with the native method as a hybrid to generate VM ranks. The proposed methods are evaluated on three use-cases and are observed to be up to 91 times faster than benchmarking the entire VM. In both methods, small containers provide the same quality of rankings as a large container. The native method generates ranks with over 90% and 86% accuracy for sequential and parallel execution of an application compared against benchmarking the whole VM. The hybrid method did not improve the quality of the rankings significantly.

A Scalable General Purpose System for Large-Scale Graph Processing

Graph analytics is an important and computationally demanding class of data analytics. It is essential to balance scalability, ease-of-use and high performance in large scale graph analytics. As such, it is necessary to hide the complexity of parallelism, data distribution and memory locality behind an abstract interface. The aim of this work is to build a scalable graph analytics framework that does not demand significant parallel programming experience based on NUMA-awareness.
The realization of such a system faces two key problems:
(i)~how to develop a scale-free parallel programming framework that scales efficiently across NUMA domains; (ii)~how to efficiently apply graph partitioning in order to create separate and largely independent work items that can be distributed among threads.