In this paper, we present the concept of "Ben-ware" as a beneficial software system capable of identifying anomalous human behaviour within a 'closed' organisation's IT infrastructure. We note that this behaviour may be malicious (for example, an employee is seeking to act against the best interest of the organisation by stealing confidential information) or benign (for example, an employee is applying some workaround to complete their job). To help distinguish between users who are intentionally malicious and those who are benign, we use human behaviour modelling along with Artificial Intelligence. Ben-ware has been developed as a distributed system comprising of probes for data collection, intermediate nodes for data routing and higher nodes for data analysis. This allows for real-time analysis with low impact on the overall infrastructure, which may contain legacy and low-power resources. We present an analysis of the appropriateness of the Ben-ware system for deployment within a large closed organisation, comprising of both new and legacy hardware, to protect its essential information. This analysis is performed in terms of the memory footprint, disk footprint and processing requirements of the different parts of the system.

Accurately forecasting air quality could lead to the development of dynamic, data-driven policy-making and improved early warning detection systems. Deep learning has demonstrated the potential to produce highly accurate forecasting models, but it is noted that much literature focuses on narrow datasets and typically considers one geographic area. In this research, three diverse air quality datasets are utilised to evaluate four deep learning algorithms, which are feedforward neural networks, Long Short-Term Memory (LSTM) recurrent neural networks, DeepAR and Temporal Fusion Transformers (TFTs). The study uses these modules to forecast CO, NO₂, O₃, and particulate matter 2.5 and 10 (PM_2.5, PM₁₀) individually, producing a 24 h forecast for a given sensor and pollutant. Each model is optimised using a hyperparameter and a feature selection process, evaluating the utility of exogenous data such as meteorological data, including wind speed and temperature, along with the inclusion of other pollutants. The findings show that the TFT and DeepAR algorithms achieve superior performance over their simpler counterparts, though they may prove challenging in practical applications. It is noted that while some covariates such as CO are important covariates for predicting NO₂ across all three datasets, other parameters such as context length proved inconsistent across the three areas, suggesting that parameters such as context length are location and pollutant specific.

Workflows offer a great potential for enacting co-related jobs in an automated manner. This is especially desirable when workflows are large or there is a desire to run a workflow multiple times. Much research has been conducted in reducing the make span of running workflows and maximising the utilisation of the resources they run on, with some existing research investigates how to reduce the energy consumption of workflows on dedicated resources. We extend the HTC-Sim simulation framework to support workflows allowing us to evaluate different scheduling strategies on the overheads and energy consumption of workflows run on non-dedicated systems. We evaluate a number of scheduling strategies from the literature in an environment where (workflow) jobs can be evicted by higher priority users.

Interest in Artificial Intelligence (AI) and the popularity of such courses has increased over the past few years, as a consequence higher education (HE) institutions are now offering courses in such fields. However, there is a current lack of research relating to best practice for teaching this complex topic area which encompasses both computing and mathematics knowledge. This paper outlines an initial study that set out to determine the threshold concepts within this domain through use of the One Minute Paper technique with students currently studying AI. Our results identified a number of specific models which students found troublesome including the support vector machine, recurrent neural network and the multilayer perceptron. The results indicated that topics related to deep learning were more complex for the students to fully comprehend and students may require greater in-depth tuition and alternative support mechanisms in this area in particular. Further investigation is needed to determine the heterogeneity of content delivered on courses relating to AI as well as additional studies at other HE institutions to gather more data on potential threshold concepts and the best methods to teach them.