Deriving word association networks from text corpora

This article presents and evaluates a model to automatically derive word association networks from text corpora. Two aspects were evaluated: To what degree can corpus-based word association networks (CANs) approximate human word association networks with respect to (1) their ability to quantitatively predict word associations and (2) their structural network characteristics. Word association networks are the basis of the human mental lexicon. However, extracting such networks from human subjects is laborious, time consuming and thus necessarily limited in relation to the breadth of human vocabulary. Automatic derivation of word associations from text corpora would address these limitations. In both evaluations corpus-based processing provided vector representations for words. These representations were then employed to derive CANs using two measures: (1) the well known cosine metric, which is a symmetric measure, and (2) a new asymmetric measure computed from orthogonal vector projections. For both evaluations, the full set of 4068 free association networks (FANs) from the University of South Florida word association norms were used as baseline human data. Two corpus based models were benchmarked for comparison: a latent topic model and latent semantic analysis (LSA). We observed that CANs constructed using the asymmetric measure were slightly less effective than the topic model in quantitatively predicting free associates, and slightly better than LSA. The structural networks analysis revealed that CANs do approximate the FANs to an encouraging degree.

Scaffolding understanding of tables and graphs

The literacy demands of tables and graphs are different from those of prose texts such as narrative. This paper draws from part of a qualitative case study which sought to investigate strategies that scaffold and enhance the teaching and learning of varied representations in text. As indicated in the paper, the method focused on the teaching and learning of tables and graphs with use of Freebody and Luke's (1990) four resources model from literacy education.

Medical semantic similarity with a neural language model

Advances in neural network language models have demonstrated that these models can effectively learn representations of words meaning. In this paper, we explore a variation of neural language models that can learn on concepts taken from structured ontologies and extracted from free-text, rather than directly from terms in free-text. This model is employed for the task of measuring semantic similarity between medical concepts, a task that is central to a number of techniques in medical informatics and information retrieval. The model is built with two medical corpora (journal abstracts and patient records) and empirically validated on two ground-truth datasets of human-judged concept pairs assessed by medical professionals. Empirically, our approach correlates closely with expert human assessors ($\approx$ 0.9) and outperforms a number of state-of-the-art benchmarks for medical semantic similarity. The demonstrated superiority of this model for providing an effective semantic similarity measure is promising in that this may translate into effectiveness gains for techniques in medical information retrieval and medical informatics (e.g., query expansion and literature-based discovery).

Enhancing the teaching and learning of mathematical visual images

The literacy demands of mathematics are very different to those in other subjects (Gough, 2007; O'Halloran, 2005; Quinnell, 2011; Rubenstein, 2007) and much has been written on the challenges that literacy in mathematics poses to learners (Abedi and Lord, 2001; Lowrie and Diezmann, 2007, 2009; Rubenstein, 2007). In particular, a diverse selection of visuals typifies the field of mathematics (Carter, Hipwell and Quinnell, 2012), placing unique literacy demands on learners. Such visuals include varied tables, graphs, diagrams and other representations, all of which are used to communicate information.

Reflections and encounters: Exploring awareness in an academic environment

Earlier work within the CSCW community treated the notion of awareness as an important resource for supporting shared work and work-related activities. However, new trends have emerged in recent times that utilize the notion of awareness beyond work-related activities and explore social, emotional and interpersonal aspects of people’s everyday lives. To investigate this broader notion of awareness, we carried out a field study using ethnographic and cultural probe based methods in an academic setting. Our aim was to study staff members’ everyday activities in their natural surroundings; understand how awareness beyond work-related activities plays out and how it is dealt with. Our field study results shed light on two broad and sometimes overlapping themes of interaction between staff members: 1) self-representations and 2) casual encounters. We provide examples from the field illustrating these two themes. In general, our results show how awareness is closely associated with people’s everyday lives, where they creatively and artfully utilize ordinary resources from their environments to carry out their routine activities. Using the results of our field study, we describe the design of a situated display called Panorama that is meant to support non-critical, non-work-related awareness within work environments.

Pattern-based topic modelling and its application for information filtering and information retrieval

This thesis targets on a challenging issue that is to enhance users' experience over massive and overloaded web information. The novel pattern-based topic model proposed in this thesis can generate high-quality multi-topic user interest models technically by incorporating statistical topic modelling and pattern mining. We have successfully applied the pattern-based topic model to both fields of information filtering and information retrieval. The success of the proposed model in finding the most relevant information to users mainly comes from its precisely semantic representations to represent documents and also accurate classification of the topics at both document level and collection level.

Applying digital product memories in industrial production

Industrial production and supply chains face increased demands for mass customization and tightening regulations on the traceability of goods, leading to higher requirements concerning flexibility, adaptability, and transparency of processes. Technologies for the ’Internet of Things' such as smart products and semantic representations pave the way for future factories and supply chains to fulfill these challenging market demands. In this chapter a backend-independent approach for information exchange in open-loop production processes based on Digital Product Memories DPMs is presented. By storing order-related data directly on the item, relevant lifecycle information is attached to the product itself. In this way, information handover between several stages of the value chain with focus on the manufacturing phase of a product has been realized. In order to report best practices regarding the application of DPM in the domain of industrial production, system prototype implementations focusing on the use case of producing and handling a smart drug case are illustrated.

Towards robust automatic affective classification of images using facial expressions for practical applications

Affect is an important feature of multimedia content and conveys valuable information for multimedia indexing and retrieval. Most existing studies for affective content analysis are limited to low-level features or mid-level representations, and are generally criticized for their incapacity to address the gap between low-level features and high-level human affective perception. The facial expressions of subjects in images carry important semantic information that can substantially influence human affective perception, but have been seldom investigated for affective classification of facial images towards practical applications. This paper presents an automatic image emotion detector (IED) for affective classification of practical (or non-laboratory) data using facial expressions, where a lot of “real-world” challenges are present, including pose, illumination, and size variations etc. The proposed method is novel, with its framework designed specifically to overcome these challenges using multi-view versions of face and fiducial point detectors, and a combination of point-based texture and geometry. Performance comparisons of several key parameters of relevant algorithms are conducted to explore the optimum parameters for high accuracy and fast computation speed. A comprehensive set of experiments with existing and new datasets, shows that the method is effective despite pose variations, fast, and appropriate for large-scale data, and as accurate as the method with state-of-the-art performance on laboratory-based data. The proposed method was also applied to affective classification of images from the British Broadcast Corporation (BBC) in a task typical for a practical application providing some valuable insights.

Parallel streaming signature EM-tree: A clustering algorithm for web scale applications

The proliferation of the web presents an unsolved problem of automatically analyzing billions of pages of natural language. We introduce a scalable algorithm that clusters hundreds of millions of web pages into hundreds of thousands of clusters. It does this on a single mid-range machine using efficient algorithms and compressed document representations. It is applied to two web-scale crawls covering tens of terabytes. ClueWeb09 and ClueWeb12 contain 500 and 733 million web pages and were clustered into 500,000 to 700,000 clusters. To the best of our knowledge, such fine grained clustering has not been previously demonstrated. Previous approaches clustered a sample that limits the maximum number of discoverable clusters. The proposed EM-tree algorithm uses the entire collection in clustering and produces several orders of magnitude more clusters than the existing algorithms. Fine grained clustering is necessary for meaningful clustering in massive collections where the number of distinct topics grows linearly with collection size. These fine-grained clusters show an improved cluster quality when assessed with two novel evaluations using ad hoc search relevance judgments and spam classifications for external validation. These evaluations solve the problem of assessing the quality of clusters where categorical labeling is unavailable and unfeasible.

Projecting the new world city: The city as spectacle in an urban light festival

As a precursor to the 2014 G20 Leaders’ Summit held in Brisbane, Australia, the Queensland Government sponsored a program of G20 Cultural Celebrations, designed to showcase the Summit’s host city. The cultural program’s signature event was the Colour Me Brisbane festival, a two-week ‘citywide interactive light and projection installations’ festival that was originally slated to run from 24 October to 9 November, but which was extended due to popular demand to conclude with the G20 Summit itself on 16 November. The Colour Me Brisbane festival comprised a series projection displays that promoted visions of the city’s past, present, and future at landmark sites and iconic buildings throughout the city’s central business district and thus transformed key buildings into forms of media architecture. In some instances the media architecture installations were interactive, allowing the public to control aspects of the projections through a computer interface situated in front of the building; however, the majority of the installations were not interactive in this sense. The festival was supported by a website that included information regarding the different visual and interactive displays and links to social media to support public discussion regarding the festival (Queensland Government 2014). Festival-goers were also encouraged to follow a walking-tour map of the projection sites that would take them on a 2.5 kilometre walk from Brisbane’s cultural precinct, through the city centre, concluding at parliament house. In this paper, we investigate the Colour Me Brisbane festival and the broader G20 Cultural Celebrations as a form of strategic placemaking—designed, on the one hand, to promote Brisbane as a safe, open, and accessible city in line with the City Council’s plan to position Brisbane as a ‘New World City’ (Brisbane City Council 2014). On the other hand, it was deployed to counteract growing local concerns and tensions over the disruptive and politicised nature of the G20 Summit by engaging the public with the city prior to the heightened security and mobility restrictions of the Summit weekend. Harnessing perspectives from media architecture (Brynskov et al. 2013), urban imaginaries (Cinar & Bender 2007), and social media analysis, we take a critical approach to analysing the government-sponsored projections, which literally projected the city onto itself, and public responses to them via the official, and heavily promoted, social media hashtags (#colourmebrisbane and #g20cultural). Our critical framework extends the concepts of urban phantasmagoria and urban imaginaries into the emerging field of media architecture to scrutinise its potential for increased political and civic engagement. Walter Benjamin’s concept of phantasmagoria (Cohen 1989; Duarte, Firmino, & Crestani 2014) provides an understanding of urban space as spectacular projection, implicated in commodity and techno-culture. The concept of urban imaginaries (Cinar & Bender 2007; Kelley 2013)—that is, the ways in which citizens’ experiences of urban environments are transformed into symbolic representations through the use of imagination—similarly provides a useful framing device in thinking about the Colour Me Brisbane projections and their relation to the construction of place. Employing these critical frames enables us to examine the ways in which the installations open up the potential for multiple urban imaginaries—in the sense that they encourage civic engagement via a tangible and imaginative experience of urban space—while, at the same time, supporting a particular vision and way of experiencing the city, promoting a commodified, sanctioned form of urban imaginary. This paper aims to dissect the urban imaginaries intrinsic to the Colour Me Brisbane projections and to examine how those imaginaries were strategically deployed as place-making schemes that choreograph reflections about and engagement with the city.

Design of a tangible data visualization (extended)

In this paper we describe the design of DNA Jewelry, which is a wearable tangible data representation of personal DNA profile data. An iterative design process was followed to develop a 3D form-language that could be mapped to standard DNA profile data, with the aim of retaining readability of data while also producing an aesthetically pleasing and unique result in the area of personalised design. The work explores design issues with the production of data tangibles, contributes to a growing body of research exploring tangible representations of data and highlights the importance of approaches that move between technology, art and design.

It’s my site, and I’ll do what I want: Performing female identity through digital identity curation in online spaces

This chapter is based on a qualitative case study that researched the perceptions of nine male and female pre-service English teachers’ in regards to their preparedness to mentor positive digital conduct in Social network sites (SNS). These sites enable individuals to perform public representations of identity, consumed by virtual audiences, with various degrees of perceived privacy. The chapter frames what we call “identity curation” through three theoretical lenses; of performativity, customisation and critical literacy. This chapter discusses one of the themes that emerged from the research, which is the way in which “normalised” and naturalised representations of femineity on SNS were judged more harshly than masculine representations.

Dataset-invariant covariance normalization for out-domain PLDA speaker verification

In this paper we introduce a novel domain-invariant covariance normalization (DICN) technique to relocate both in-domain and out-domain i-vectors into a third dataset-invariant space, providing an improvement for out-domain PLDA speaker verification with a very small number of unlabelled in-domain adaptation i-vectors. By capturing the dataset variance from a global mean using both development out-domain i-vectors and limited unlabelled in-domain i-vectors, we could obtain domain- invariant representations of PLDA training data. The DICN- compensated out-domain PLDA system is shown to perform as well as in-domain PLDA training with as few as 500 unlabelled in-domain i-vectors for NIST-2010 SRE and 2000 unlabelled in-domain i-vectors for NIST-2008 SRE, and considerable relative improvement over both out-domain and in-domain PLDA development if more are available.

Is being #instagay different from an #lgbttakeover? A cross-platform investigation of sexual and gender identity performances

This paper introduces research in progress that examines how queer women perform sexual identity across social media platforms. Applying a lens of queer theory and Actor Network Theory, it discusses women’s embodied self-representations as taking on forms that both conform to and elaborate upon the selfie genre of digital representation. Acknowledging similarities and differences across platforms, specifically between Instagram and Vine, a novel walkthrough method is introduced to identify platform characteristics that shape identity performances. This method provides insights into the role of platforms in identity performances, which can be combined with analysis of user-generated content and interviews to better understand digital media’s constraints and affordances for queer representation.