Learning vector representation of medical objects via EMR-driven nonnegative restricted Boltzmann machines (eNRBM)

Electronic medical record (EMR) offers promises for novel analytics. However, manual feature engineering from EMR is labor intensive because EMR is complex - it contains temporal, mixed-type and multimodal data packed in irregular episodes. We present a computational framework to harness EMR with minimal human supervision via restricted Boltzmann machine (RBM). The framework derives a new representation of medical objects by embedding them in a low-dimensional vector space. This new representation facilitates algebraic and statistical manipulations such as projection onto 2D plane (thereby offering intuitive visualization), object grouping (hence enabling automated phenotyping), and risk stratification. To enhance model interpretability, we introduced two constraints into model parameters: (a) nonnegative coefficients, and (b) structural smoothness. These result in a novel model called eNRBM (EMR-driven nonnegative RBM). We demonstrate the capability of the eNRBM on a cohort of 7578 mental health patients under suicide risk assessment. The derived representation not only shows clinically meaningful feature grouping but also facilitates short-term risk stratification. The F-scores, 0.21 for moderate-risk and 0.36 for high-risk, are significantly higher than those obtained by clinicians and competitive with the results obtained by support vector machines.

Learning how to engage students online in hard times

In a context of financial restraint and enterprising university managers, teacher-researchers have reason to be sceptical about the trend towards online teaching and away from learning for its own sake. This article departs from both economic and technological determinism and turns instead to ideas about technology embedded in social and political institutions. Activity theory offers a useful means of analysing such embeddedness. Its Marxian assumptions about human nature specify a non-deterministic approach to technology. Its dynamic model of the subjects, tools, and objects of activity within a context of rules, a community, and a division of labour helps to specify aspects of the authors process of learning how to use electronic conferencing effectively. A full deployment of activity theory would also analyse the activity of students. Here the evidence comes mainly from the activity of researcher-teachers engaging greater activity among students. The numbers of students involved precludes reliable quantitative analysis but qualitative evidence from students does support conclusions about researcher-teachers learning how to make best use of electronic conferencing.

Developing wikis to foster web-based learning communities : an iterative approach

This paper discusses ongoing research at Deakin University, which focuses on developing wikis to foster web-based learning communities. Research to date has used wikis to facilitate collaborative icebreaker exercises, discussions, and to create knowledge repositories. Student feedback has contributed to the iterative revision of the wiki interface, the icebreaker exercise and the development of new tasks for students to complete using the wiki. The analysis and discussion of the experiments presented in this paper focuses on usage trends such as the signature, viewing and editing patterns exhibited by the student cohort. The community building potential of wikis is discussed, highlighting the specific wiki features that can be used to foster a sense of community in a web-based learning environment. Finally, issues surrounding the development of web-based learning communities, which have emerged through the wiki study, are discussed and future directions are outlined.

Time-to-Collision (TTC) judgements with offsize objects show that Tau needs to be intergrated with familiar size to explain TTC performance

Observers judged TTC with computer-generated displays simulating an approaching object in three familiar-size conditions:

(i) Real-size (smaller, larger objects depicted as tennis, soccer balls respectively).
(ii) Off-size (smaller, larger objects depicted as soccer, tennis balls respectively).
(iii) Ambiguous-size (smaller, larger objects depicted as texture-less black balls of different size).

Displays simulated objects approaching observersí viewpoint from 24.96 m, and disappearing at 5.76 m. Manipulation of approach velocities (4.8-19.2 msec-1) produced viewing times from 1.0 to 4.0 sec, and delays between object disappearance and tau-based TTC ranging from 0.3 to 1.2 sec. Motion characteristics of smaller and larger objects in the three familiar-size conditions simulated those of approaching real-sized tennis and soccer balls respectively; that is, for each approach velocity, tau‚-based TTC was the same across the three conditions for smaller and larger objects.

Results showed that, consistent with the proposition of tau-determined TTC, TTC estimates in the real-size condition were uninfluenced by object size. This is contrary to previous reports that TTC for larger objects is underestimated relative to TTC for smaller objects. However, such size-dependent TTC differences were found in the ambiguous-size condition, with even larger differences in the off-size condition; TTCs for the ëlargerí tennis ball were much less than TTCs to the ësmallerí soccer ball compared to corresponding TTCs in the ambiguous-size condition. These results are problematic for the proposition that tau solely determines TTC. We discuss the role of perceptual learning in resolving this problem.

A constructivist approach to the teaching & learning of astronomy

Astronomy is one of the recurrent contemporary issues in the mass media where news related to comets, new stars, satellites, space tests, etc., frequently appear. Through this media presence students develop an interest in learning about aspects of astronomy. However, the information students gain from various sources, both inside and outside the classroom, doesn't often increase their knowledge about the most basic and common aspects of astronomy (Martinez Pena & Gil Quilez, 2001). Studies that have explored students' and teachers' understandings of astronomical concepts (Kalkan et al, 2007; Trumper, 2001) have found many alternative conceptions relating to basic astronomical processes such as day and night, the seasons, gravity and the relative distances between celestial objects.

Introducing a learning repository using a blended professional development approach

This paper outlines the professional development program used to introduce a learning repository at Deakin University. Providing appropriate, timely and effective professional development programs to support academic and other staff is one of the objectives of the Deakin University Teaching and Learning Functional Plan 2008. Our blended program combines web-based and face-to-face training with a wide variety of resources to support staff. Issues noted in the literature relating to the introduction and use of learning repositories informed the planning and development of our program. Challenges and issues we experienced at Deakin are also outlined.

A comparative study of supervised learning techniques for data-driven haptic simulation

This paper focuses on the choice of a supervised learning algorithm and possible data preprocessing in the domain of data-driven haptic simulation. This is done through a comparison of the performance of different supervised learning techniques with and without data preprocessing. The simulation of haptic interactions with deformable objects using data-driven methods has emerged as an alternative to parametric methods. The accuracy of the simulation depends on the empirical data and the learning method. Several methods were suggested in the literature and here we provide a comparison between their performance and applicability to this domain. We selected four examples to be compared: singular learning mechanism which is artificial neural networks (ANN), attribute selection followed by ANN learning process, ensemble of multiple learning techniques, and attribute selection followed by the learning ensemble. These methods performance was compared in the domain of simulating multiple interactions with a deformable object with nonlinear material behavior.

Students' intercultural development through language learning in Vietnamese tertiary education: a case study on the use of film as an innovative approach

Vietnam's open-door policy, its socialist-oriented market economy, recent growth in cross-border education and skills mobility, regionalisation and globalisation have created an increasing demand for Vietnamese graduates to develop not only their English language but also their intercultural competence. This paper discusses the issue of student intercultural learning and development in the Vietnamese English as a foreign language (EFL) class. Especially, it addresses the use of film as an innovative approach to engage Vietnamese students in intercultural learning and development in the EFL classroom. The study reported in this paper draws on rich sources of data which include in-depth interviews with students, student reflective journals and video-recorded class observations at a university in central Vietnam. Overall, five key themes relating to student intercultural learning through film have been identified in this study. These include enhancing knowledge about cultural differences, engaging in cross-cultural comparison, breaking cultural stereotypes, immersing students in authentic learning and living in the world of ‘other’ culture and the integrated mode of intercultural language learning. The study is a significant contribution to scholarly research on the use of media objects to enhance student intercultural learning in language classrooms in developing countries.

An ontology-based learning resources management framework for exploratory e-learning

This paper presents an ontology-based conceptual framework for effectively managing exploratory e-learning resources. The proposed framework has five significant novel features including authentication of retrieved resources, automatic ontology-based query refinement, reuse-oriented management of retrieved resources, adaptive retrieval of learning resources based on the style and preference of individual learners, and synthesisation of retrieval and management activities for creating reusable learning repositories. The applicability of the framework is demonstrated using a sample fragment of an ontology developed in the database domain.

Synchronous Objects for One Flat Thing, reproduced

Synchronous Objects is a website where a visitor/ reader can investigate the interlocking systems of organisation in the choreography of William Forsythe's dance One Flat Thing, reproduced (2000). These systems were quantified through the collection of data and transformed into a series of objects - synchronous objects - that work in harmony to explore these choreographic structures, reveal their patterns and re-imagine what they might look like. The goal of the project was to create a learning platform to engage a broad public in this exploration of choreography, to explore cross-disciplinary research and to stimulate creative discovery for specialists and non-specialists alike.

Learning in-between, across and beyond workplace boundaries

In this paper, we challenge conventional theories underpinning the practice of human resource development (HRD) that typically focus on the objects of learning - individuals, skills, jobs - or separate the contexts of work (performance) from learning (training). Using case study research from a public utility and a winery, we found similarities in workers' experiences of various organisational practices despite different objects of learning. These practices were, firstly, temporarily occupying (acting up) in a more senior position and, secondly, collaborating across functions. Unlike notions of boundary-crossing or job rotation, we argue that learning 'in-between' develops interactional competence that is grounded within the work practices of interacting individuals. This relational view opens up possibilities for redirecting the focus of HRD practice towards integrating work and learning in ways that develop the potential for learning in-between, across and beyond the boundaries of work. © 2012 Inderscience Enterprises Ltd.

Learning complementary saliency priors for foreground object segmentation in complex scenes

Object segmentation is widely recognized as one of the most challenging problems in computer vision. One major problem of existing methods is that most of them are vulnerable to the cluttered background. Moreover, human intervention is often required to specify foreground/background priors, which restricts the usage of object segmentation in real-world scenario. To address these problems, we propose a novel approach to learn complementary saliency priors for foreground object segmentation in complex scenes. Different from existing saliency-based segmentation approaches, we propose to learn two complementary saliency maps that reveal the most reliable foreground and background regions. Given such priors, foreground object segmentation is formulated as a binary pixel labelling problem that can be efficiently solved using graph cuts. As such, the confident saliency priors can be utilized to extract the most salient objects and reduce the distraction of cluttered background. Extensive experiments show that our approach outperforms 16 state-of-the-art methods remarkably on three public image benchmarks.

Supervised learning for labelling human body with attached props

 In this research, a novel method for generating training data of human postures with attached objects is proposed. The results has shown a significant increase in body-part classification accuracy for subjects with props from 60% to 94% using the generated image set

Young children's transfer of learning from a touchscreen device

Because young children are devoting increasing time to playing on handheld touchscreen devices, understanding children's ability to learn from this activity is important. Through two experiments we examined the ability of 4- to 6-year-old children to learn how to solve a problem (Tower of Hanoi) on a touchscreen device and subsequently apply this learning in their interactions with physical objects. The results were that participants demonstrated significant improvement at solving the task irrespective of the modality (touchscreen vs. physical version) with which they practiced. Moreover, children's learning on the touchscreen smoothly transferred to a subsequent attempt on the physical version. We conclude that, at least with respect to certain activities, children are quite capable of transferring learning from touchscreen devices. This result highlights the limitations of generalizing across screen-based activities (e.g., "screen time") in discussing the effects of media on young children's development.