Exponentiated gradient algorithms for large-margin structured classification

We consider the problem of structured classification, where the task is to predict a label y from an input x, and y has meaningful internal structure. Our framework includes supervised training of Markov random fields and weighted context-free grammars as special cases. We describe an algorithm that solves the large-margin optimization problem defined in [12], using an exponential-family (Gibbs distribution) representation of structured objects. The algorithm is efficient—even in cases where the number of labels y is exponential in size—provided that certain expectations under Gibbs distributions can be calculated efficiently. The method for structured labels relies on a more general result, specifically the application of exponentiated gradient updates [7, 8] to quadratic programs.

Automatic domain ontology extraction for context-sensitive opinion mining

Automated analysis of the sentiments presented in online consumer feedbacks can facilitate both organizations’ business strategy development and individual consumers’ comparison shopping. Nevertheless, existing opinion mining methods either adopt a context-free sentiment classification approach or rely on a large number of manually annotated training examples to perform context sensitive sentiment classification. Guided by the design science research methodology, we illustrate the design, development, and evaluation of a novel fuzzy domain ontology based contextsensitive opinion mining system. Our novel ontology extraction mechanism underpinned by a variant of Kullback-Leibler divergence can automatically acquire contextual sentiment knowledge across various product domains to improve the sentiment analysis processes. Evaluated based on a benchmark dataset and real consumer reviews collected from Amazon.com, our system shows remarkable performance improvement over the context-free baseline.

Accelerated leap methods for simulating discrete stochastic chemical kinetics

Biologists are increasingly conscious of the critical role that noise plays in cellular functions such as genetic regulation, often in connection with fluctuations in small numbers of key regulatory molecules. This has inspired the development of models that capture this fundamentally discrete and stochastic nature of cellular biology - most notably the Gillespie stochastic simulation algorithm (SSA). The SSA simulates a temporally homogeneous, discrete-state, continuous-time Markov process, and of course the corresponding probabilities and numbers of each molecular species must all remain positive. While accurately serving this purpose, the SSA can be computationally inefficient due to very small time stepping so faster approximations such as the Poisson and Binomial τ-leap methods have been suggested. This work places these leap methods in the context of numerical methods for the solution of stochastic differential equations (SDEs) driven by Poisson noise. This allows analogues of Euler-Maruyuma, Milstein and even higher order methods to be developed through the Itô-Taylor expansions as well as similar derivative-free Runge-Kutta approaches. Numerical results demonstrate that these novel methods compare favourably with existing techniques for simulating biochemical reactions by more accurately capturing crucial properties such as the mean and variance than existing methods.

Geometry-free stochastic analysis of BDS triple frequency signals

The paper presents a geometry-free approach to assess the variation of covariance matrices of undifferenced triple frequency GNSS measurements and its impact on positioning solutions. Four independent geometryfree/ ionosphere-free (GFIF) models formed from original triple-frequency code and phase signals allow for effective computation of variance-covariance matrices using real data. Variance Component Estimation (VCE) algorithms are implemented to obtain the covariance matrices for three pseudorange and three carrier-phase signals epoch-by-epoch. Covariance results from the triple frequency Beidou System (BDS) and GPS data sets demonstrate that the estimated standard deviation varies in consistence with the amplitude of actual GFIF error time series. The single point positioning (SPP) results from BDS ionosphere-free measurements at four MGEX stations demonstrate an improvement of up to about 50% in Up direction relative to the results based on a mean square statistics. Additionally, a more extensive SPP analysis at 95 global MGEX stations based on GPS ionosphere-free measurements shows an average improvement of about 10% relative to the traditional results. This finding provides a preliminary confirmation that adequate consideration of the variation of covariance leads to the improvement of GNSS state solutions.

Continued use of process modeling grammars : the impact of individual difference factors

Process modeling grammars are used by analysts to describe information systems domains in terms of the business operations an organization is conducting. While prior research has examined the factors that lead to continued usage behavior, little knowledge has been established as to what extent characteristics of the users of process modeling grammars inform usage behavior. In this study, a theoretical model is advanced that incorporates determinants of continued usage behavior as well as key antecedent individual difference factors of the grammar users, such as modeling experience, modeling background and perceived grammar familiarity. Findings from a global survey of 529 grammar users support the hypothesized relationships of the model. The study offers three central contributions. First, it provides a validated theoretical model of post-adoptive modeling grammar usage intentions. Second, it discusses the effects of individual difference factors of grammar users in the context of modeling grammar usage. Third, it provides implications for research and practice.

Near-infrared spectroscopic studies of human scalp hair in a forensic context

Human hair is a relatively inert biopolymer and can survive through natural disasters. It is also found as trace evidence at crime scenes. Previous studies by FTIRMicrospectroscopy and – Attenuated Total Reflectance (ATR) successfully showed that hairs can be matched and discriminated on the basis of gender, race and hair treatment, when interpreted by chemometrics. However, these spectroscopic techniques are difficult to operate at- or on-field. On the other hand, some near infrared spectroscopic (NIRS) instruments equipped with an optical probe, are portable and thus, facilitate the on- or at –field measurements for potential application directly at a crime or disaster scene. This thesis is focused on bulk hair samples, which are free of their roots, and thus, independent of potential DNA contribution for identification. It explores the building of a profile of an individual with the use of the NIRS technique on the basis of information on gender, race and treated hair, i.e. variables which can match and discriminate individuals. The complex spectra collected may be compared and interpreted with the use of chemometrics. These methods can then be used as protocol for further investigations. Water is a common substance present at forensic scenes e.g. at home in a bath, in the swimming pool; it is also common outdoors in the sea, river, dam, puddles and especially during DVI incidents at the seashore after a tsunami. For this reason, the matching and discrimination of bulk hair samples after the water immersion treatment was also explored. Through this research, it was found that Near Infrared Spectroscopy, with the use of an optical probe, has successfully matched and discriminated bulk hair samples to build a profile for the possible application to a crime or disaster scene. Through the interpretation of Chemometrics, such characteristics included Gender and Race. A novel approach was to measure the spectra not only in the usual NIR range (4000 – 7500 cm-1) but also in the Visible NIR (7500 – 12800 cm-1). This proved to be particularly useful in exploring the discrimination of differently coloured hair, e.g. naturally coloured, bleached or dyed. The NIR region is sensitive to molecular vibrations of the hair fibre structure as well as that of the dyes and damage from bleaching. But the Visible NIR region preferentially responds to the natural colourants, the melanin, which involves electronic transitions. This approach was shown to provide improved discrimination between dyed and untreated hair. This thesis is an extensive study of the application of NIRS with the aid of chemometrics, for matching and discrimination of bulk human scalp hair. The work not only indicates the strong potential of this technique in this field but also breaks new ground with the exploration of the use of the NIR and Visible NIR ranges for spectral sampling. It also develops methods for measuring spectra from hair which has been immersed in different water media (sea, river and dam)

Slum free India program. Report on the findings and recommendations of QUT BEE workshop to investigate strategies to deliver better outcomes

As a result of a broad invitation extended by Professor Martin Betts, Executive Dean of the Faculty of Built Environment and Engineering, to the community of interest at QUT, a cross-disciplinary collaborative workshop was conducted to contribute ideas about responding to the Government of India’s urgent requirement to implement a program to re-house slum dwellers. This is a complex problem facing the Indian Ministry of Housing. Not only does the government aspire to eradicate existing slum conditions and to achieve tangible results within five years, but it must also ensure that slums do not form in the future. The workshop focused on technological innovation in construction to deliver transformation from the current unsanitary and overcrowded informal urban settlements to places that provide the economically weaker sections of Indian society with healthy, environmentally sustainable, economically viable mass housing that supports successful urban living. The workshop was conducted in two part process as follows: Initially, QUT academics from diverse fields shared current research and provided technical background to contextualise the challenge at a pre-workshop briefing session. This was followed by a one-day workshop during which participants worked intensively in multi-disciplinary groups through a series of exercises to develop innovative approaches to the complex problem of slum redevelopment. Dynamic, compressed work sessions, interspersed with cross-functional review and feedback by the whole group took place throughout the day. Reviews emphasised testing the concepts for their level of complexity, and likelihood of success. The two-stage workshop process achieved several objectives:  Inspired a sense of shared purpose amongst a diverse group of academics  Built participants’ knowledge of each other’s capacity  Engaged multi disciplinary team in an innovative design research process  Built participants’ confidence in the collaborative process  Demonstrated that collaborative problem solving can create solutions that represent transformative change.  Developed a framework of how workable solutions might be developed for the program through follow up workshops and charrettes of a similar nature involving stakeholders drawn from the context of the slum housing program management.

Rule-based approach for identifying assertions in clinical free-text data

A rule-based approach for classifying previously identified medical concepts in the clinical free text into an assertion category is presented. There are six different categories of assertions for the task: Present, Absent, Possible, Conditional, Hypothetical and Not associated with the patient. The assertion classification algorithms were largely based on extending the popular NegEx and Context algorithms. In addition, a health based clinical terminology called SNOMED CT and other publicly available dictionaries were used to classify assertions, which did not fit the NegEx/Context model. The data for this task includes discharge summaries from Partners HealthCare and from Beth Israel Deaconess Medical Centre, as well as discharge summaries and progress notes from University of Pittsburgh Medical Centre. The set consists of 349 discharge reports, each with pairs of ground truth concept and assertion files for system development, and 477 reports for evaluation. The system’s performance on the evaluation data set was 0.83, 0.83 and 0.83 for recall, precision and F1-measure, respectively. Although the rule-based system shows promise, further improvements can be made by incorporating machine learning approaches.

Representation of assertions in clinical free-­text using SNOMED CT

Information retrieval (IR) by clinicians in the healthcare setting is critical for informing clinical decision-making. However, a large part of this information is in the form of free-text and inhibits clinical decision support and effective healthcare services. This makes meaningful use of clinical free-­text in electronic health records (EHRs) for patient care a difficult task. Within the context of IR, given a repository of free-­text clinical reports, one might want to retrieve and analyse data for patients who have a known clinical finding.

Modeling the viability of the free-ranging cheetah population in Namibia : an object-oriented Bayesian network approach

Conservation of free-ranging cheetah (Acinonyx jubatus) populations is multi faceted and needs to be addressed from an ecological, biological and management perspective. There is a wealth of published research, each focusing on a particular aspect of cheetah conservation. Identifying the most important factors, making sense of various (and sometimes contrasting) findings, and taking decisions when little or no empirical data is available, are everyday challenges facing conservationists. Bayesian networks (BN) provide a statistical modeling framework that enables analysis and integration of information addressing different aspects of conservation. There has been an increased interest in the use of BNs to model conservation issues, however the development of more sophisticated BNs, utilizing object-oriented (OO) features, is still at the frontier of ecological research. We describe an integrated, parallel modeling process followed during a BN modeling workshop held in Namibia to combine expert knowledge and data about free-ranging cheetahs. The aim of the workshop was to obtain a more comprehensive view of the current viability of the free-ranging cheetah population in Namibia, and to predict the effect different scenarios may have on the future viability of this free-ranging cheetah population. Furthermore, a complementary aim was to identify influential parameters of the model to more effectively target those parameters having the greatest impact on population viability. The BN was developed by aggregating diverse perspectives from local and independent scientists, agents from the national ministry, conservation agency members and local fieldworkers. This integrated BN approach facilitates OO modeling in a multi-expert context which lends itself to a series of integrated, yet independent, subnetworks describing different scientific and management components. We created three subnetworks in parallel: a biological, ecological and human factors network, which were then combined to create a complete representation of free-ranging cheetah population viability. Such OOBNs have widespread relevance to the effective and targeted conservation management of vulnerable and endangered species.

Building sensitising terms to understand free-play in open-ended interactive art environments

In this paper we introduce and discuss the nature of free-play in the context of three open-ended interactive art installation works. We observe the interaction work of situated free-play of the participants in these environments and, building on precedent work, devise a set of sensitising terms derived both from the literature and from what we observe from participants interacting there. These sensitising terms act as guides and are designed to be used by those who experience, evaluate or report on open-ended interactive art. That is, we propose these terms as a common-ground language to be used by participants communicating while in the art work to describe their experience, by researchers in the various stages of research process (observation, coding activity, analysis, reporting, and publication), and by inter-disciplinary researchers working across the fields of HCI and art. This work builds a foundation for understanding the relationship between free-play, open-ended environments, and interactive installations and contributes sensitising terms useful for the HCI community for discussion and analysis of open-ended interactive art works.

From Kafka to KAFTA: Intellectual property, and the Korea-Australia free trade agreement

This paper provides a critical examination of the intellectual property sections of the Korea-Australia Free Trade Agreement 2014. Chapter 13 of the Korea-Australia Free Trade Agreement 2014 deals with the subject of intellectual property law. The Chapter covers such topics as the purposes and objectives of intellectual property law; copyright law; trade mark law; patent law; and intellectual property enforcement. The Joint Standing Committee on Treaties in the Australian Parliament highlighted the controversy surrounding this chapter of the agreement: The intellectual property rights chapter of KAFTA has drawn considerable attention from academics and stakeholders regarding the proposed need for changes to Australian intellectual property law and the inclusion of intellectual property in the definition of investment with regard to the investor-state dispute mechanism. Other concerns raised with the Committee include the prescriptive nature of the chapter, the lack of recognition of the broader public interests of intellectual property rights, and possible changes to fair use provisions. Article 13.1.1 of the Korea-Australia Free Trade Agreement 2014 provides that: ‘Each Party recognises the importance of adequate and effective protection of intellectual property rights, while ensuring that measures to enforce those rights do not themselves become barriers to legitimate trade.’ This is an unsatisfactory description of the objectives and purposes of intellectual property law in both Australia and Korea. There is a failure to properly consider the range of public purposes served by intellectual property law – such as providing for access to knowledge, promoting competition and innovation, protecting consumer rights, and allowing for the protection of public health, food security, and the environment. Such a statement of principles and objectives detracts from the declaration in the TRIPS Agreement 1994 of the public interest objectives to be served by intellectual property. Chapter 11 of the Korea-Australia Free Trade Agreement 2014 is an investment chapter, with an investor-state dispute settlement regime. This chapter is highly controversial – given the international debate over investor-state dispute settlement; the Australian context for the debate; and the text of the Korea-Australia Free Trade Agreement 2014. In April 2014, the United Nations Conference on Trade and Development (UNCTAD) released a report on Recent Developments in Investor-State Dispute Settlement. The overall figures are staggering. UNCTAD reports a significant growth in investment-state dispute settlement, across a wide array of different fields of public regulation. Given the broad definition of investment, intellectual property owners will be able to use the investor-state dispute settlement regime in the Korea-Australia Free Trade Agreement 2014. This will have significant implications for all the various disciplines of intellectual property – including copyright law, trade mark law, and patent law.