David JonesAberystwyth University | AU · Department of Computer Science
David Jones
PhD MSc BSc
About
25
Publications
10,880
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
1,906
Citations
Introduction
My current research interests are in the application of advanced reasoning techniques within the context of the digital twin. These are state-of-the-art computational approaches such as computational causal inference, functional imagination, and artificial intuition that in theory should provide solutions to the higher cognitive challenges of Industry 4.0 such as adaptability and fault diagnosis.
Publications
Publications (25)
While there has been a recent growth of interest in the Digital Twin, a variety of definitions employed across industry and academia remain. There is a need to consolidate research such to maintain a common understanding of the topic and ensure future research efforts are to be based on solid foundations. Through a systematic literature review and...
Engineering and the manner in which engineers think is largely visual and functional, and yet engineers are typically provided with search engines that are text-based. While software based on a visual and functional ethos exist (CAD for example), when searching for information engineers are still required to enter a text query into a search box. Th...
During engineering design, designers employ three types of model: physical, virtual and cognitive. The role and contribution of each is documented in literature albeit fragmented in nature. Consequentially, a gap in understanding exists in terms of how these models and the transitions between them impact the designer and design process. This paper...
Prior work has shown Convolutional Neural Networks (CNNs) trained on surrogate Computer Aided Design (CAD) models are able to detect and classify real-world artefacts from photographs. The applications of which support twinning of digital and physical assets in design, including rapid extraction of part geometry from model repositories, information...
Modern version control strategies are highly capable at supporting the management of virtual artefacts. The process of developing a new product, however, is not limited to virtual artefacts. Today’s fast-paced industrial processes require a diverse range of both virtual and physical artefacts to explore, refine, and evaluate designs. These virtual...
The management of data related to prototypes created during new product development is seen as a beneficial yet challenging activity. While attempts have been made to understand prototypes and their context in a range of use-cases, there is a gap in the understanding of the data that captures a prototype's context and physical form. This paper high...
Transdisciplinary engineering is the exchange of knowledge about product, process,
organization, or social environment in the context of innovation. The ATDE book series aims to explore the evolution of engineering, and promote transdisciplinary practices, in which the exchange of different types of knowledge from a diverse range of disciplines is...
People prescribed physiotherapy exercises can struggle to engage with exercises due to a lack of mental stimulation in the repetitive tasks. The introduction of VR to motion-based physiotherapy can be beneficial, however, currently available physiotherapy applications are focused on gaming and the gamification of physiotherapy, something that will...
The digital twin is often presented as the solution to Industry 4.0 and, while there are many areas where this may be the case, there is a risk that a reliance on existing machine learning methods will not be able to deliver the high level cognitive capabilities such as adaptability, cause and effect, and planning that Industry 4.0 requires. As the...
The digital twin is often presented as the solution to Industry 4.0 and, while there are many areas where this may be the case, there is a risk that a reliance on existing machine learning methods will not be able to deliver the high level cognitive capabilities such as adaptability, cause and effect, and planning that Industry 4.0 requires. As the...
Physical prototyping during early stage design typically represents an iterative process. Commonly, a single prototype will be used throughout the process, with its form being modified as the design evolves. If the form of the prototype is not captured as each iteration occurs understanding how specific design changes impact upon the satisfaction o...
Classifying shape and form is a core feature of Engineering Design and one that we do this instinctively on a daily basis. Matching similar components to then reduce unique component counts, determining whether a competitors design infringes on copyright and receiving market feedback on product styling are all examples where shape and form comes in...
Prototyping is an indispensable activity in product development that facilitates the generation of knowledge in the design process. It is crucial that this knowledge is the right knowledge (e.g., type, fidelity, and accessibility) to ensure stakeholders can evaluate and decision-make effectively. While this is well-recognised, prior work has focuse...
Prior work has shown Convolutional Neural Networks (CNNs) trained on surrogate Computer Aided Design (CAD) models are able to detect and classify real-world artefacts from photographs. The applications of which support twinning of digital and physical assets in design, including rapid extraction of part geometry from model repositories, information...
While extensive modelling - both physical and virtual - is imperative to develop right-first-time products, the parallel use of virtual and physical models gives rise to two interrelated issues: the lack of revision control for physical prototypes; and the need for designers to manually inspect, measure, and interpret modifications to either virtua...
This paper contributes to a better understanding and design of dashboards for monitoring of engineering projects based on the projects’ digital footprint and user-centered design approach.
The use of Mixed Reality (MR) tools can improve information retrieval, collaboration and decision making, thus aiding the management of buildings within the operation and maintenance (O&M) lifecycle stages. In this paper, we focus on the use of MR in visualising BIM data to aid building lifecycle management. This paper compares current and emerging...
Modern engineering work, both project-based and operations, is replete with complexity and variety making the effective development of detailed understanding of work underway difficult, which in turn impacts on management and assurance of performance. Leveraging the digital nature of modern engineering work, recent research has demonstrated the cap...
The use of Mixed Reality (MR) tools can improve information retrieval, collaboration and decision making, thus aiding the management of buildings within the operation and maintenance (O&M) lifecycle stages. In this paper, we focus on the use of MR in visualising BIM data to aid building lifecycle management.
Modern engineering work, both project-based and operations, is replete with complexity and variety making the effective development of detailed understanding of work underway difficult, which in turn impacts on management and assurance of performance.
Engineering Information Management (EIM) and Information Retrieval (IR) systems are central to the day to day running of large engineering organisations. The capture, interrogation, retrieval and presentation of information from design to disposal is considered to be a key enabler for greater efficiency and decision making and in turn improved prod...
Understanding how users formulate search queries can allow the development of search engines that are tailored to the way users search and thus improve the knowledge discovery process, a key challenge for Product Lifecycle Management (PLM) systems.
This paper presents part-of-speech (POS) statistical analysis on two sets of ‘Top 500’ search query l...
Questions
Question (1)
If I have three data sets and want to see how well rules can be elicited from each data set, would it be best practice to take an off-the-shelf pre-trained model and train it on the three data sets, or to use something like Keras Tuner and develop a model from scratch for each of the data sets?
My current thinking is there could be some variance in a pre-trained model's ability to adapt to each data set, leading to bias, and so the way to control for that variance is to train a new model using Keras Tuner for each of the data sets (with all the tuner parameters being controlled).
Any one have any thoughts?