Probabilistic topographic information visualisation

The focus of this thesis is the extension of topographic visualisation mappings to allow for the incorporation of uncertainty. Few visualisation algorithms in the literature are capable of mapping uncertain data with fewer able to represent observation uncertainties in visualisations. As such, modifications are made to NeuroScale, Locally Linear Embedding, Isomap and Laplacian Eigenmaps to incorporate uncertainty in the observation and visualisation spaces. The proposed mappings are then called Normally-distributed NeuroScale (N-NS), T-distributed NeuroScale (T-NS), Probabilistic LLE (PLLE), Probabilistic Isomap (PIso) and Probabilistic Weighted Neighbourhood Mapping (PWNM). These algorithms generate a probabilistic visualisation space with each latent visualised point transformed to a multivariate Gaussian or T-distribution, using a feed-forward RBF network. Two types of uncertainty are then characterised dependent on the data and mapping procedure. Data dependent uncertainty is the inherent observation uncertainty. Whereas, mapping uncertainty is defined by the Fisher Information of a visualised distribution. This indicates how well the data has been interpolated, offering a level of ‘surprise’ for each observation. These new probabilistic mappings are tested on three datasets of vectorial observations and three datasets of real world time series observations for anomaly detection. In order to visualise the time series data, a method for analysing observed signals and noise distributions, Residual Modelling, is introduced. The performance of the new algorithms on the tested datasets is compared qualitatively with the latent space generated by the Gaussian Process Latent Variable Model (GPLVM). A quantitative comparison using existing evaluation measures from the literature allows performance of each mapping function to be compared. Finally, the mapping uncertainty measure is combined with NeuroScale to build a deep learning classifier, the Cascading RBF. This new structure is tested on the MNist dataset achieving world record performance whilst avoiding the flaws seen in other Deep Learning Machines.

A matter of professionalism? Reflection and reflexivity in continuing engineering education and practice

The argument that this paper sets out to critique is that in order to promote professionalism in Engineering Education and Practice, graduate level engineering programmes need to introduce the concepts of reflection and reflexivity into the curriculum right from the onset. By focusing upon the delivery of a newly developed „Work Based‟ Master’s level programme in Professional Engineering, this paper provides an overview of the first part of an empirical study which sets out to investigate the challenges associated with embedding reflection and reflexivity into Engineering Education. The paper concludes by noting that whilst student engineers may struggle with the concepts of reflection and reflexivity, with support and encouragement such difficulties can be overcome. Moreover, by encouraging students to reflect upon their Professional Practice, the programme not only enables students to consider how they may apply what they have learnt to their Professional Practice, but also encourages them to think about how they can link their experiences as Professional Engineers to what and how they learn both whilst on the programme but also as lifelong learners.

Improved data visualisation through multiple dissimilarity modelling

Popular dimension reduction and visualisation algorithms rely on the assumption that input dissimilarities are typically Euclidean, for instance Metric Multidimensional Scaling, t-distributed Stochastic Neighbour Embedding and the Gaussian Process Latent Variable Model. It is well known that this assumption does not hold for most datasets and often high-dimensional data sits upon a manifold of unknown global geometry. We present a method for improving the manifold charting process, coupled with Elastic MDS, such that we no longer assume that the manifold is Euclidean, or of any particular structure. We draw on the benefits of different dissimilarity measures allowing for the relative responsibilities, under a linear combination, to drive the visualisation process.

Learning user and product distributed representations using a sequence model for sentiment analysis

In product reviews, it is observed that the distribution of polarity ratings over reviews written by different users or evaluated based on different products are often skewed in the real world. As such, incorporating user and product information would be helpful for the task of sentiment classification of reviews. However, existing approaches ignored the temporal nature of reviews posted by the same user or evaluated on the same product. We argue that the temporal relations of reviews might be potentially useful for learning user and product embedding and thus propose employing a sequence model to embed these temporal relations into user and product representations so as to improve the performance of document-level sentiment analysis. Specifically, we first learn a distributed representation of each review by a one-dimensional convolutional neural network. Then, taking these representations as pretrained vectors, we use a recurrent neural network with gated recurrent units to learn distributed representations of users and products. Finally, we feed the user, product and review representations into a machine learning classifier for sentiment classification. Our approach has been evaluated on three large-scale review datasets from the IMDB and Yelp. Experimental results show that: (1) sequence modeling for the purposes of distributed user and product representation learning can improve the performance of document-level sentiment classification; (2) the proposed approach achieves state-of-The-Art results on these benchmark datasets.

Optical fiber sensors embedded in flexible polymer foils

In traditional electrical sensing applications, multiplexing and interconnecting the different sensing elements is a major challenge. Recently, many optical alternatives have been investigated including optical fiber sensors of which the sensing elements consist of fiber Bragg gratings. Different sensing points can be integrated in one optical fiber solving the interconnection problem and avoiding any electromagnetical interference (EMI). Many new sensing applications also require flexible or stretchable sensing foils which can be attached to or wrapped around irregularly shaped objects such as robot fingers and car bumpers or which can even be applied in biomedical applications where a sensor is fixed on a human body. The use of these optical sensors however always implies the use of a light-source, detectors and electronic circuitry to be coupled and integrated with these sensors. The coupling of these fibers with these light sources and detectors is a critical packaging problem and as it is well-known the costs for packaging, especially with optoelectronic components and fiber alignment issues are huge. The end goal of this embedded sensor is to create a flexible optical sensor integrated with (opto)electronic modules and control circuitry. To obtain this flexibility, one can embed the optical sensors and the driving optoelectronics in a stretchable polymer host material. In this article different embedding techniques for optical fiber sensors are described and characterized. Initial tests based on standard manufacturing processes such as molding and laser structuring are reported as well as a more advanced embedding technique based on soft lithography processing.

'Catching them young':inspiring future engineers, an exploratory study

The requirement that primary school children appreciate fully the pivotal role played by engineering in the sustainable development of future society is reflected in the literature with much attention being paid to the need to spark childrens engineering imagination early-on in their school careers. Moreover, UK policy documents highlight the value of embedding engineering into the school curriculum, arguing that programmes aimed at inspiring children through a process of real-life learning experiences are vital pedagogical tools in promoting engineering to future generations. Despite such attention, engineering education at school-level remains sporadic, often reliant on individual engineering-entrepreneurs such as teachers who, through personal interest, get children involved in what are usually extra-curriculum, time-limited engineering focused programmes and competitions. This paper briefly discusses an exploratory study aimed at investigating the issues surrounding embedding engineering into the primary school curriculum. It gives some insight into the perceptions of various stakeholders in respect of the viability and value of introducing engineering education into the primary school curriculum from the age of 6 or 7. A conceptual framework of primary level engineering education, bringing together the theoretical, pedagogical and policy related phenomena influencing the development of engineering education is proposed. The paper concludes by arguing that in order to avert future societal disaster, childrens engineering imagination needs to be ignited from an early age and that to do this primary engineering education needs to be given far more educational, social and political attention. © 2009 Authors.