A5078307880- Distributed and Parallel Computing Systems
- Cloud Computing and Resource Management
- Scientific Computing and Data Management
- Parallel Computing and Optimization Techniques
- Advanced Graph Neural Networks
- Complex Network Analysis Techniques
- Advanced Data Storage Technologies
- Smart Grid Energy Management
- Electric Power System Optimization
- Energy Load and Power Forecasting
- Peer-to-Peer Network Technologies
- Building Energy and Comfort Optimization
- Simulation Techniques and Applications
- Topic Modeling
- Anomaly Detection Techniques and Applications
- Cybercrime and Law Enforcement Studies
- Neural Networks and Applications
- Integrated Energy Systems Optimization
- Machine Learning in Healthcare
- Advanced Malware Detection Techniques
- Network Security and Intrusion Detection
- Machine Learning and Data Classification
- IoT and Edge/Fog Computing
- Advanced Neural Network Applications
- Species Distribution and Climate Change
Newcastle University
2013-2024
Durham University
2014-2017
Imperial College London
2001-2010
La Jolla Alcohol Research
2004-2006
Heidelberg (Poland)
2005-2006
University of Newcastle Australia
2000
Deep learning is rapidly becoming a go-to tool for many artificial intelligence problems due to its ability outperform other approaches and even humans at problems. Despite popularity we are still unable accurately predict the time it will take train deep network solve given problem. This training can be seen as product of per epoch number epochs which need performed reach desired level accuracy. Some work has been carried out an - most have based around assumption that linearly related...
Accurate predictive modelling of the growth microbial communities requires credible representation interactions biological, chemical and mechanical processes. However, although biological processes are represented in a number Individual-based Models (IbMs) interaction mechanics is limited. Conversely, there mechanically sophisticated IbMs with only elementary biology chemistry. This study focuses on addressing these limitations by developing flexible IbM that can robustly combine physical...
Abstract Urban flood risk modelling is a highly topical example of intensive computational processing. Such processing increasingly required by range organisations including local government, engineering consultancies and the insurance industry to fulfil statutory requirements provide professional services. As demands for this type work become more common, then ownership high-end resources warranted but if use sporadic with tight deadlines Cloud computing could cost-effective alternative....
We present NUFEB (Newcastle University Frontiers in Engineering Biology), a flexible, efficient, and open source software for simulating the 3D dynamics of microbial communities. The tool is based on Individual-based Modelling (IbM) approach, where microbes are represented as discrete units their behaviour changes over time due to variety processes. This approach allows us study population behaviours that emerge from interaction between individuals environment. built top classical molecular...
Toxic air pollutants such as PM2.5 and NO\textsubscript{2} have been linked to various health implications, including respiratory diseases, cardiovascular cancers. With recent updates from the World Health Organisation introducing even stricter daily annual exposure limits, expectations placed upon governing bodies better manage urban quality increased. It is suggested in this research that may be managed through use of smart city infrastructure data-driven techniques introduce dynamic...
In order to reliably generate electricity meet the demands of customer base, it is essential match supply with demand. Short-term load forecasting utilised in both real-time scheduling electricity, and load-frequency control. This paper aims improve accuracy load-forecasting by using machine learning techniques predict 30 minutes ahead smart meter data. We k-means clustering algorithm cluster similar individual consumers fit distinct models per cluster. Public holidays were taken into...
Graph embeddings have become a key and widely used technique within the field of graph mining, proving to be successful across broad range domains including social, citation, transportation biological. Unsupervised embedding techniques aim automatically create low-dimensional representation given graph, which captures structural elements in resulting space. However, date, there has been little work exploring exactly topological structures are being learned embeddings, could possible way...
Artificial Intelligence and its sub-disciplines are becoming increasingly relevant in numerous areas of academia as well industry can now be considered a core area Computer Science [ 84 ]. The Higher Education sector offering more courses Machine Learning than ever before. However, there is lack research pertaining to best practices for teaching this complex domain that heavily relies on both computing mathematical knowledge. We conducted literature review qualitative study with students...
Accurate occupancy prediction in smart buildings is a key element to reduce building energy consumption and control HVAC systems (Heating – Ventilation and– Air Conditioning) efficiently, resulting an increment of human comfort. This work focuses on the problem modelling (occupied / unoccupied) using environmental sensor data. A novel transfer learning approach was used enhance accuracy when amounts historical training data are limited. The proposed models applied case study three office...
Unpredictable job execution environments pose a significant barrier to the widespread adoption of Grid paradigm, because innate risk jobs failing execute at time specified by user. We demonstrate that predictability can be enhanced with supporting infrastructure consisting three parts: Performance modelling and monitoring, scheduling which exploits application structure an advanced reservation resource management service. prove theoretically times using reservations display less variance...
Graphs have become a crucial way to represent large, complex and often temporal datasets across wide range of scientific disciplines. However, when graphs are used as input machine learning models, this rich information is frequently disregarded during the process, resulting in suboptimal performance on certain inference tasks. To combat this, we introduce Temporal Neighbourhood Aggregation (TNA), novel vertex representation model architecture designed capture both topological directly...
Individual based models (IbM) must transition from research tools to engineering tools. To make the we aspire develop large, three dimensional and physically biologically credible models. Biological credibility can be promoted by grounding, as far possible, biology in thermodynamics. Thermodynamic principles are known have predictive power microbial ecology. However, this turn requires a model that incorporates pH chemical speciation. Physical implies plausible mechanics connection with...
Graphs are a commonly used construct for representing relationships between elements in complex high dimensional datasets. Many real-world phenomenon dynamic nature, meaning that any graph to represent them is inherently temporal. However, many of the machine learning models designed capture knowledge about structure these graphs ignore this rich temporal information when creating representations graph. This results which do not perform well make predictions future state - especially delta...
Effective exploitation of computational grids can only be achieved when applications are fully integrated with the grid middleware and underlying resources. Fundamental to this is information. Information about structure behavior application, capability networking resources, availability access these resources by an individual, a group or organization. This paper describes environment that open, extensible platform independent. We match high-level application specification, defined as...
The classification of graphs is a key challenge within many scientific fields using to represent data and an active area research. Graph can be critical in identifying labelling unknown dataset has seen application across fields. poses two distinct problems: the elements graph entire graph. Whilst there considerable work on first problem, efficient accurate massive into one or more classes has, thus far, received less attention. In this paper we propose Deep Topology Classification (DTC)...
Clinical data, such as evaluations, treatments, vital sign and lab test results, are usually observed recorded in hospital systems. Making use of data to help physicians evaluate the mortality risk in-hospital patients provides an invaluable source information that can ultimately with improving healthcare services. In particular, quick accurate predictions be valuable for who making decisions about interventions. this work we introduce a predictive Deep Learning model patients. Stacked...