An investigation into inductive parameter learning in complex hierarchical knowledge structures representing clinical expertise

This dissertation investigates the very important and current problem of modelling human expertise. This is an apparent issue in any computer system emulating human decision making. It is prominent in Clinical Decision Support Systems (CDSS) due to the complexity of the induction process and the vast number of parameters in most cases. Other issues such as human error and missing or incomplete data present further challenges. In this thesis, the Galatean Risk Screening Tool (GRiST) is used as an example of modelling clinical expertise and parameter elicitation. The tool is a mental health clinical record management system with a top layer of decision support capabilities. It is currently being deployed by several NHS mental health trusts across the UK. The aim of the research is to investigate the problem of parameter elicitation by inducing them from real clinical data rather than from the human experts who provided the decision model. The induced parameters provide an insight into both the data relationships and how experts make decisions themselves. The outcomes help further understand human decision making and, in particular, help GRiST provide more accurate emulations of risk judgements. Although the algorithms and methods presented in this dissertation are applied to GRiST, they can be adopted for other human knowledge engineering domains.

Semisupervised learning of hierarchical latent trait models for data visualization

Recently, we have developed the hierarchical Generative Topographic Mapping (HGTM), an interactive method for visualization of large high-dimensional real-valued data sets. In this paper, we propose a more general visualization system by extending HGTM in three ways, which allows the user to visualize a wider range of data sets and better support the model development process. 1) We integrate HGTM with noise models from the exponential family of distributions. The basic building block is the Latent Trait Model (LTM). This enables us to visualize data of inherently discrete nature, e.g., collections of documents, in a hierarchical manner. 2) We give the user a choice of initializing the child plots of the current plot in either interactive, or automatic mode. In the interactive mode, the user selects "regions of interest," whereas in the automatic mode, an unsupervised minimum message length (MML)-inspired construction of a mixture of LTMs is employed. The unsupervised construction is particularly useful when high-level plots are covered with dense clusters of highly overlapping data projections, making it difficult to use the interactive mode. Such a situation often arises when visualizing large data sets. 3) We derive general formulas for magnification factors in latent trait models. Magnification factors are a useful tool to improve our understanding of the visualization plots, since they can highlight the boundaries between data clusters. We illustrate our approach on a toy example and evaluate it on three more complex real data sets. © 2005 IEEE.

Learning multi-faceted activities from heterogeneous data with the product space hierarchical Dirichlet processes

Hierarchical Dirichlet processes (HDP) was originally designed and experimented for a single data channel. In this paper we enhanced its ability to model heterogeneous data using a richer structure for the base measure being a product-space. The enhanced model, called Product Space HDP (PS-HDP), can (1) simultaneously model heterogeneous data from multiple sources in a Bayesian nonparametric framework and (2) discover multilevel latent structures from data to result in different types of topics/latent structures that can be explained jointly. We experimented with the MDC dataset, a large and real-world data collected from mobile phones. Our goal was to discover identity–location– time (a.k.a who-where-when) patterns at different levels (globally for all groups and locally for each group). We provided analysis on the activities and patterns learned from our model, visualized, compared and contrasted with the ground-truth to demonstrate the merit of the proposed framework. We further quantitatively evaluated and reported its performance using standard metrics including F1-score, NMI, RI, and purity. We also compared the performance of the PS-HDP model with those of popular existing clustering methods (including K-Means, NNMF, GMM, DP-Means, and AP). Lastly, we demonstrate the ability of the model in learning activities with missing data, a common problem encountered in pervasive and ubiquitous computing applications.

Learning approach and perception of learning context in economics education : a causality test

In the study of student learning literature, the traditional view holds that when students are faced with heavy workload, poor teaching, and content that they cannot relate to – important aspects of the learning context, they will more likely utilise the surface approach to learning due to stresses, lack of understanding and lack of perceived relevance of the content (Kreber, 2003; Lizzio, Wilson, & Simons, 2002; Ramdsen, 1989; Ramsden, 1992; Trigwell & Prosser, 1991; Vermunt, 2005). For example, in studies involving health and medical sciences students, courses that utilised student-centred, problem-based approaches to teaching and learning were found to elicit a deeper approach to learning than the teacher-centred, transmissive approach (Patel, Groen, & Norman, 1991; Sadlo & Richardson, 2003). It is generally accepted that the line of causation runs from the learning context (or rather students’ self reported data on the learning context) to students’ learning approaches. That is, it is the learning context as revealed by students’ self-reported data that elicit the associated learning behaviour. However, other research studies also found that the same teaching and learning environment can be perceived differently by different students. In a study of students’ perceptions of assessment requirements, Sambell and McDowell (1998) found that students “are active in the reconstruction of the messages and meanings of assessment” (p. 391), and their interpretations are greatly influenced by their past experiences and motivations. In a qualitative study of Hong Kong tertiary students, Kember (2004) found that students using the surface learning approach reported heavier workload than students using the deep learning approach. According to Kember if students learn by extracting meanings from the content and making connections, they will more likely see the higher order intentions embodied in the content and the high cognitive abilities being assessed. On the other hand, if they rote-learn for the graded task, they fail to see the hierarchical relationship in the content and to connect the information. These rote-learners will tend to see the assessment as requiring memorising and regurgitation of a large amount of unconnected knowledge, which explains why they experience a high workload. Kember (2004) thus postulate that it is the learning approach that influences how students perceive workload. Campbell and her colleagues made a similar observation in their interview study of secondary students’ perceptions of teaching in the same classroom (Campbell et al., 2001). The above discussions suggest that students’ learning approaches can influence their perceptions of assessment demands and other aspects of the learning context such as relevance of content and teaching effectiveness. In other words, perceptions of elements in the teaching and learning context are endogenously determined. This study attempted to investigate the causal relationships at the individual level between learning approaches and perceptions of the learning context in economics education. In this study, students’ learning approaches and their perceptions of the learning context were measured. The elements of the learning context investigated include: teaching effectiveness, workload and content. The authors are aware of existence of other elements of the learning context, such as generic skills, goal clarity and career preparation. These aspects, however, were not within the scope of this present study and were therefore not investigated.

Hierarchical fuzzy logic traffic control at a road junction using genetic algorithms

Traffic control at a road junction by a complex fuzzy logic controller is investigated. The increase in the complexity of junction means more number of input variables must be taken into account, which will increase the number of fuzzy rules in the system. A hierarchical fuzzy logic controller is introduced to reduce the number of rules. Besides, the increase in the complexity of the controller makes formulation of the fuzzy rules difficult. A genetic algorithm based off-line leaning algorithm is employed to generate the fuzzy rules. The learning algorithm uses constant flow-rates as training sets. The system is tested by both constant and time-varying flow-rates. Simulation results show that the proposed controller produces lower average delay than a fixed-time controller does under various traffic conditions.

A stochastic model for natural feature representation

This paper presents a robust stochastic model for the incorporation of natural features within data fusion algorithms. The representation combines Isomap, a non-linear manifold learning algorithm, with Expectation Maximization, a statistical learning scheme. The representation is computed offline and results in a non-linear, non-Gaussian likelihood model relating visual observations such as color and texture to the underlying visual states. The likelihood model can be used online to instantiate likelihoods corresponding to observed visual features in real-time. The likelihoods are expressed as a Gaussian Mixture Model so as to permit convenient integration within existing nonlinear filtering algorithms. The resulting compactness of the representation is especially suitable to decentralized sensor networks. Real visual data consisting of natural imagery acquired from an Unmanned Aerial Vehicle is used to demonstrate the versatility of the feature representation.

A Stochastic View of Optimal Regret through Minimax Duality

We study the regret of optimal strategies for online convex optimization games. Using von Neumann's minimax theorem, we show that the optimal regret in this adversarial setting is closely related to the behavior of the empirical minimization algorithm in a stochastic process setting: it is equal to the maximum, over joint distributions of the adversary's action sequence, of the difference between a sum of minimal expected losses and the minimal empirical loss. We show that the optimal regret has a natural geometric interpretation, since it can be viewed as the gap in Jensen's inequality for a concave functional--the minimizer over the player's actions of expected loss--defined on a set of probability distributions. We use this expression to obtain upper and lower bounds on the regret of an optimal strategy for a variety of online learning problems. Our method provides upper bounds without the need to construct a learning algorithm; the lower bounds provide explicit optimal strategies for the adversary. Peter L. Bartlett, Alexander Rakhlin

Assessment policy change in relation to English language teaching and learning in China : a study of perspectives from two universities

In response to concerns about the quality of English Language Learning (ELL) education at tertiary level, the Chinese Ministry of Education (CMoE) launched the College English Reform Program (CERP) in 2004. By means of a press release (CMoE, 2005) and a guideline document titled College English Curriculum Requirements (CECR) (CMoE, 2007), the CERP proposed two major changes to the College English assessment policy, which were: (1) the shift to optional status for the compulsory external test, the College English Test Band 4 (CET4); and (2) the incorporation of formative assessment into the existing summative assessment framework. This study investigated the interactions between the College English assessment policy change, the theoretical underpinnings, and the assessment practices within two Chinese universities (one Key University and one Non-Key University). It adopted a sociocultural theoretical perspective to examine the implementation process as experienced by local actors of institutional and classroom levels. Systematic data analysis using a constant comparative method (Merriam, 1998) revealed that contextual factors and implementation issues did not lead to significant differences in the two cases. Lack of training in assessment and the sociocultural factors such as the traditional emphasis on the product of learning and hierarchical teacher/students relationship are decisive and responsible for the limited effect of the reform.

Hierarchical robust design optimisation of shock control bump devices for airfoil drag reduction

The paper investigates a detailed Active Shock Control Bump Design Optimisation on a Natural Laminar Flow (NLF) aerofoil; RAE 5243 to reduce cruise drag at transonic flow conditions using Evolutionary Algorithms (EAs) coupled to a robust design approach. For the uncertainty design parameters, the positions of boundary layer transition (xtr) and the coefficient of lift (Cl) are considered (250 stochastic samples in total). In this paper, two robust design methods are considered; the first approach uses a standard robust design method, which evaluates one design model at 250 stochastic conditions for uncertainty. The second approach is the combination of a standard robust design method and the concept of hierarchical (multi-population) sampling (250, 50, 15) for uncertainty. Numerical results show that the evolutionary optimization method coupled to uncertainty design techniques produces useful and reliable Pareto optimal SCB shapes which have low sensitivity and high aerodynamic performance while having significant total drag reduction. In addition,it also shows the benefit of using hierarchical robust method for detailed uncertainty design optimization.

Meta-evaluation to improve learning, evaluation capacity development and sustainability : Findings from a participatory evaluation project in Nepal

Recent studies suggest that meta-evaluation can be valuable in developing new approaches to evaluation, building evaluation capacities, and enhancing organizational learning. These new extensions of the concept of meta-evaluation are significant, given the growing emphasis on improving the quality and effectiveness of evaluation practices in the South Asian region. Following a review of the literature, this paper presents a case study of the use of concurrent meta-evaluation in the four-year project Assessing Communication for Social Change which developed and trialled a participatory impact assessment methodology in collaboration with a development communication Non-government organization (NGO) in Nepal. Key objectives of the meta-evaluation included to: continuously develop, adapt and improve the impact assessment methodology, Monitoring and Evaluation (M&E) systems and process and other project activities; identify impacts of the project; and build capacities in critical reflection and review. Our analysis indicates that this meta-evaluation was essential to understanding various constraints related to the organizational context that affected the success of the project and the development of improved M&E systems and capacities within the NGO. We identified several limitations of our meta-evaluation methods, which were balanced by the strengths of other methods. Our case study suggests that as well as assessing the quality, credibility and value of evaluation practices, meta-evaluations need to focus on important contextual issues that can have significant impacts on the outcomes of participatory evaluation projects. They include hierarchical organizational cultures, communication barriers, power/knowledge relations, and the time and resources available. Meta-evaluations also need to consider wider issues such as the sustainability of evaluation systems and approaches.

Personalized ontology learning for enhancing text mining effectiveness

Over the last decade, the majority of existing search techniques is either keyword- based or category-based, resulting in unsatisfactory effectiveness. Meanwhile, studies have illustrated that more than 80% of users preferred personalized search results. As a result, many studies paid a great deal of efforts (referred to as col- laborative filtering) investigating on personalized notions for enhancing retrieval performance. One of the fundamental yet most challenging steps is to capture precise user information needs. Most Web users are inexperienced or lack the capability to express their needs properly, whereas the existent retrieval systems are highly sensitive to vocabulary. Researchers have increasingly proposed the utilization of ontology-based tech- niques to improve current mining approaches. The related techniques are not only able to refine search intentions among specific generic domains, but also to access new knowledge by tracking semantic relations. In recent years, some researchers have attempted to build ontological user profiles according to discovered user background knowledge. The knowledge is considered to be both global and lo- cal analyses, which aim to produce tailored ontologies by a group of concepts. However, a key problem here that has not been addressed is: how to accurately match diverse local information to universal global knowledge. This research conducts a theoretical study on the use of personalized ontolo- gies to enhance text mining performance. The objective is to understand user information needs by a \bag-of-concepts" rather than \words". The concepts are gathered from a general world knowledge base named the Library of Congress Subject Headings. To return desirable search results, a novel ontology-based mining approach is introduced to discover accurate search intentions and learn personalized ontologies as user profiles. The approach can not only pinpoint users' individual intentions in a rough hierarchical structure, but can also in- terpret their needs by a set of acknowledged concepts. Along with global and local analyses, another solid concept matching approach is carried out to address about the mismatch between local information and world knowledge. Relevance features produced by the Relevance Feature Discovery model, are determined as representatives of local information. These features have been proven as the best alternative for user queries to avoid ambiguity and consistently outperform the features extracted by other filtering models. The two attempt-to-proposed ap- proaches are both evaluated by a scientific evaluation with the standard Reuters Corpus Volume 1 testing set. A comprehensive comparison is made with a num- ber of the state-of-the art baseline models, including TF-IDF, Rocchio, Okapi BM25, the deploying Pattern Taxonomy Model, and an ontology-based model. The gathered results indicate that the top precision can be improved remarkably with the proposed ontology mining approach, where the matching approach is successful and achieves significant improvements in most information filtering measurements. This research contributes to the fields of ontological filtering, user profiling, and knowledge representation. The related outputs are critical when systems are expected to return proper mining results and provide personalized services. The scientific findings have the potential to facilitate the design of advanced preference mining models, where impact on people's daily lives.

Motion planning and stochastic control with experimental validation on a planetary rover

Motion planning for planetary rovers must consider control uncertainty in order to maintain the safety of the platform during navigation. Modelling such control uncertainty is difficult due to the complex interaction between the platform and its environment. In this paper, we propose a motion planning approach whereby the outcome of control actions is learned from experience and represented statistically using a Gaussian process regression model. This model is used to construct a control policy for navigation to a goal region in a terrain map built using an on-board RGB-D camera. The terrain includes flat ground, small rocks, and non-traversable rocks. We report the results of 200 simulated and 35 experimental trials that validate the approach and demonstrate the value of considering control uncertainty in maintaining platform safety.

Learning information literacy

Introduction This paper reports on university students' experiences of learning information literacy. Method Phenomenography was selected as the research approach as it describes the experience from the perspective of the study participants, which in this case is a mixture of undergraduate and postgraduate students studying education at an Australian university. Semi-structured, one-on-one interviews were conducted with fifteen students. Analysis The interview transcripts were iteratively reviewed for similarities and differences in students' experiences of learning information literacy. Categories were constructed from an analysis of the distinct features of the experiences that students reported. The categories were grouped into a hierarchical structure that represents students' increasingly sophisticated experiences of learning information literacy. Results The study reveals that students experience learning information literacy in six ways: learning to find information; learning a process to use information; learning to use information to create a product; learning to use information to build a personal knowledge base; learning to use information to advance disciplinary knowledge; and learning to use information to grow as a person and to contribute to others. Conclusions Understanding the complexity of the concept of information literacy, and the collective and diverse range of ways students experience learning information literacy, enables academics and librarians to draw on the range of experiences reported by students to design academic curricula and information literacy education that targets more powerful ways of learning to find and use information.

Enhanced stochastic mobility prediction with multi-output Gaussian processes

Outdoor robots such as planetary rovers must be able to navigate safely and reliably in order to successfully perform missions in remote or hostile environments. Mobility prediction is critical to achieving this goal due to the inherent control uncertainty faced by robots traversing natural terrain. We propose a novel algorithm for stochastic mobility prediction based on multi-output Gaussian process regression. Our algorithm considers the correlation between heading and distance uncertainty and provides a predictive model that can easily be exploited by motion planning algorithms. We evaluate our method experimentally and report results from over 30 trials in a Mars-analogue environment that demonstrate the effectiveness of our method and illustrate the importance of mobility prediction in navigating challenging terrain.