William Mackaness

A multi-agent-based, semantic-driven system for decision support in epidemic management

Issues in epidemiology are truly multidisciplinary, requiring knowledge from diverse disciplines such as sociology, medicine, biology, geography and information science. Such inherent complexity has led to a challenge in developing decision support systems for epidemic information management, especially when data are from heterogeneous origins. In order to achieve a solution, an integrative framework is proposed. The Semantic Web is introduced in the context of enriching meaningful and machine-readable descriptions of epidemiological data. Software agents are utilised to achieve automation in semantic discovery, composition of data and process services. The objective is to enhance the performance in information retrieval in a dynamic decision-making environment while concealing technical complexity from inexperienced users. We illustrate how a prototype system can be developed by considering an epidemiology management scenario in which spatio-temporal analysis is undertaken of a specified epidemic.

Identifying related landmark tags in urban scenes using spatial and semantic clustering

Abstract There is considerable interest in developing landmark saliency models as a basis for describing urban landscapes, and in constructing wayfinding instructions, for text and spoken dialogue based systems. The challenge lies in knowing the truthfulness of such models; is what the model considers salient the same as what is perceived by the user? The method developed in this research identifies related annotated tags supplied from a web based experiment in which users were asked to tag the most salient features on urban images for the purposes of navigation and exploration. The tag collections may be used to rank landmark popularity in each scene, but the challenge is in determining which tags relate to the same object (e.g. tags relating to a particular cafe). Existing clustering techniques did not perform well for this task, and it was therefore necessary to develop a new spatial-semantic clustering method which considered the proximity of nearby tags and the similarity of their label content. The annotation similarity was initially calculated using trigrams in conjunction with a synonym list, generating a set of networks formed from the links between related tags. These networks were used to build related word lists encapsulating conceptual connections (e.g. church tower related to clock) so that during a secondary pass of the data, related network segments could be merged. This approach gives interesting insight into the partonomic relationships between the constituent parts of landmarks and the range and frequency of terms used to describe them.

Improving the sampling strategy for point-to-point line-of-sight modelling in urban environments

ABSTRACT Visibility modelling calculates what an observer could theoretically see in the surrounding region based on a digital model of the landscape. In some cases, it is not necessary, nor desirable, to compute the visibility of an entire region (i.e. a viewshed), but instead it is sufficient and more efficient to calculate the visibility from point to point, or from a point to a small set of points, such as computing the intervisibility of predators and prey in an agent-based simulation. This paper explores how different line-of-sight (LoS) sample ordering strategies increase the number of early target rejections, where the target is considered to be obscured from view, thereby improving the computational efficiency of the LoS algorithm. This is of particular importance in dynamic environments where the locations of the observers, targets and other surface objects are being frequently updated. Trials were conducted in three UK cities, demonstrating a robust fivefold increase in performance for two strategies (hop, divide and conquer). The paper concludes that sample ordering methods do impact overall efficiency, and that approaches which disperse samples along the LoS perform better in urban regions than incremental scan methods. The divide and conquer method minimises elevation interception queries, making it suitable when elevation models are held on disk rather than in memory, while the hopping strategy was equally fast, algorithmically simpler, with minimal overhead for visible target cases.

The bounds of distortion: truth, meaning and efficacy in digital geographic representation

ABSTRACT Even maps that strive for a precise representation of spatial relationships use techniques of distortion to embed a geographic extent within a two-dimensional plane – be it a page or a screen. The graphical design of geospatial information does however conform to a consensus around the conceptual limits of that distortion, with an overall design framework that constrains the design to ensure comprehension and the effective recognition of geographic entities and relations. Although constraints are necessary, we argue that the advent of digital technology, particularly in mobile mapping, warrants re-examination of the parameters of these distortions. Here we introduce the concept of ‘the bounds of distortion’ as a device for considering the conceptual boundaries of map design, and as a foundation for further work to investigate how these bounds may be redefined to better support map users with more effective graphical information. The focus here is navigational information, and in particular urban navigation and interaction with the graphical representation of urban geography and public transport networks.

Mapping the visual magnitude of popular tourist sites in Edinburgh city

There is value in being able to automatically measure and visualise the visual exposure of city sites (monuments and buildings, tourist sites) – for example, in urban planning, as an aid to automated way finding, or in augmented reality city guides. Here we present the outputs of an algorithm able to calculate visual exposure – both as an absolute measure of the façade area and in terms of a buildings perceived size (its lesser importance with distance). Both metrics influence the photogenic nature of a site. We therefore compared against maps showing the locations from where geo-located Flickr images were taken. The results accord with the metrics and therefore help disambiguate the meaning of Flickr tags.

From taxonomies to ontologies: formalizing generalization knowledge for on-demand mapping

ABSTRACT Automation of the cartographic design process is central to the delivery of bespoke maps via the web. In this paper, ontological modeling is used to explicitly represent and articulate the knowledge used in this decision-making process. A use case focuses on the visualization of road traffic accident data as a way of illustrating how ontologies provide a framework by which salient and contextual information can be integrated in a meaningful manner. Such systems are in anticipation of web-based services in which the user knows what they need, but do not have the cartographic ability to get what they want.

A dialogue based mobile virtual assistant for tourists: The SpaceBook Project

Abstract Ubiquitous mobile computing offers innovative approaches in the delivery of information that can facilitate free roaming of the city, informing and guiding the tourist as the city unfolds before them. However making frequent visual reference to mobile devices can be distracting, the user having to interact via a small screen thus disrupting the explorative experience. This research reports on an EU funded project, SpaceBook, that explored the utility of a hands-free, eyes-free virtual tour guide, that could answer questions through a spoken dialogue user interface and notify the user of interesting features in view while guiding the tourist to various destinations. Visibility modelling was carried out in real-time based on a LiDAR sourced digital surface model, fused with a variety of map and crowd sourced datasets (e.g. Ordnance Survey, OpenStreetMap, Flickr, Foursquare) to establish the most interesting landmarks visible from the users location at any given moment. A number of variations of the SpaceBook system were trialled in Edinburgh (Scotland). The research highlighted the pleasure derived from this novel form of interaction and revealed the complexity of prioritising route guidance instruction alongside identification, description and embellishment of landmark information – there being a delicate balance between the level of information ‘pushed’ to the user, and the users requests for further information. Among a number of challenges, were issues regarding the fidelity of spatial data and positioning information required for pedestrian based systems – the pedestrian having much greater freedom of movement than vehicles.

Artificial Cognitive Maps: Selecting Heterogeneous Sets of Geographic Objects and Relations to Drive Highly Contextual Task-Oriented Map Views

We present work from an on-going project to develop techniques of automated cartography. We introduce Artificial Cognitive Maps as an approach to integrating insights from spatial cognition with geographic data. The ultimate goal is to drive highly contextual map views that more effectively support navigation tasks such as travelling across large, complex cities. With a focus on our now ubiquitous small screen mobile devices, we propose that distortions on the traditional metric cartographic representation may support a reduction in cognitive load for the user, but that the logic and parameters of these distortions should be founded on the natural distortions present in our cognitive representations of geographic objects and their relation.