Efficient minimax strategies for square loss games

We consider online prediction problems where the loss between the prediction and the outcome is measured by the squared Euclidean distance and its generalization, the squared Mahalanobis distance. We derive the minimax solutions for the case where the prediction and action spaces are the simplex (this setup is sometimes called the Brier game) and the \ell_2 ball (this setup is related to Gaussian density estimation). We show that in both cases the value of each sub-game is a quadratic function of a simple statistic of the state, with coefficients that can be efficiently computed using an explicit recurrence relation. The resulting deterministic minimax strategy and randomized maximin strategy are linear functions of the statistic.

Curiosity driven reinforcement learning for motion planning on humanoids

Most previous work on artificial curiosity (AC) and intrinsic motivation focuses on basic concepts and theory. Experimental results are generally limited to toy scenarios, such as navigation in a simulated maze, or control of a simple mechanical system with one or two degrees of freedom. To study AC in a more realistic setting, we embody a curious agent in the complex iCub humanoid robot. Our novel reinforcement learning (RL) framework consists of a state-of-the-art, low-level, reactive control layer, which controls the iCub while respecting constraints, and a high-level curious agent, which explores the iCub's state-action space through information gain maximization, learning a world model from experience, controlling the actual iCub hardware in real-time. To the best of our knowledge, this is the first ever embodied, curious agent for real-time motion planning on a humanoid. We demonstrate that it can learn compact Markov models to represent large regions of the iCub's configuration space, and that the iCub explores intelligently, showing interest in its physical constraints as well as in objects it finds in its environment.

The ‘state of play’ in Australia: Early childhood educators and play-based learning

This article provides an overview of the Education Meets Play study that will investigate early childhood educators’ use of play-based learning, now mandatory under the National Quality Standard. By building on what can be gleaned about educators’ approaches to play-based learning prior to the implementation of the Early Years Learning Framework, the study will contribute to the evidence base concerning the implementation and effects of Australia’s early childhood education and care policy reform initiatives.

Secular and religious moral grounds resonating across state schools in Indonesia

This paper describes moral education in Indonesia, more particularly, how teachers have implemented the Character Education policy issued by the Ministry of Education and Culture (MOEC) in 2010. This policy required teachers to instil certain values in every lesson, including EFL lessons, to contribute towards building a shared national moral character. Drawing on Durkheim's distinction between secular and religious morality, this paper considers how state schools accommodated and promoted this ‘rational moral education' or secular morality (Durkheim, 1925) in government schools, and how it interacted with religious moral education. This paper uses Bernstein's concepts of pedagogic discourse, instructional and regulative discourses to analyse how teachers have recontextualised this policy in the micro pedagogic settings of their EFL classes. Three types of data were collected for this study: interviews, class observations and teachers' lesson plans. In this way, four EFL teachers working in state schools were interviewed on two occasions and three of their classes were observed. The first interview identified teachers' beliefs and perceptions regarding the Character Education policy. Their classroom and lesson plans were observed to augment this information. Then the final interview asked about the teacher's thinking behind their actions in the observed classes. Since character education was issued within the broader frame of school based curriculum that offered schools and teachers more choices to develop the local curriculum and its intent, the analysis will focus on what moral premises were evident in their school and classes, and how such morality was transmitted through the EFL lessons. The conclusion suggests that teachers' implementation of moral education in their classes was dominated by their school communities and the teachers' own preferred value of religiosity. Such value played out in the classes through both the regulative discourse and the instructional discourse.

The effect of feedback from pupil to teacher on assessment for leaning and visible learning: An ethnographic case study in a community school in England and the outcome in a state high school in Queensland, Australia

This study shows that there is positive regulatory effect of feedback from pupils to teachers on Assessment for Learning (AfL), classroom proactiveness, and on visible and progressive learning but not on behaviour. This research finding further articulates feedback from pupil to teacher as a paradigm shift from the classical paradigm of feedback from teacher to pupil. Here, the emphasis is geared towards pupils understanding of objectives built from previous knowledge. These are then feedback onto the teachers by the pupils in the form of discrete loops of cues and questions, where they are with their learning. This therefore enables them to move to the next level of understanding, and thus acquired independence, which in turn is reflected by their success in both formative and summative assessments. This study therefore shows that when feedback from pupil to teacher is used in combination with teacher to pupil feedback, AfL is ameliorated and hence, visible and accelerated learning occurs in a gender, nor subject non-dependent manner.

Keeping kids in school is not as simple as carrots and sticks

The OECD’s 2009 Jobs for Youth report, released on the heels of the global financial crisis, made a number of policy recommendations to the Australian government to prevent a rise in youth unemployment. Five years on, youth unemployment has risen from around 8% to 14.8%, and around one in five children do not complete year 12, according to data from the National Centre for Vocational Education Research...

Evaluation of ANPR trials for traffic policing in Queensland. Report to the Queensland Police Service, State Traffic Support Branch

In response to the Travelsafe Committee Report No. 51 – report on the inquiry into Automatic Plate Recognition Technology – it was recommended that the Queensland Police Service continue to trial the deployment of ANPR technology for traffic enforcement work and to evaluate the road safety impacts and operational effectiveness of the technology. As such, the purpose of this report is to provide an independent evaluation of a trial of ANPR that was conducted by a project team within the State Traffic Support Branch of the Queensland Police Service (QPS) and provide recommendations as to the applicability and usability of the technology for use throughout Queensland...

Automation of underground truck haulage

This paper reviews the state-of-the art in the automation of underground truck haulage. Past attemps at automating LHDs and haul trucks are described and their particular strengths and weaknesses are listed. We argue that the simple auto-tram...

Emergency pre-hospital disaster response training. What's the state of play?

Study/Objective This study examines the current state of disaster response education for Australian paramedics from a national and international perspective and identifies both potential gaps in content and challenges to the sustainability of knowledge acquired through occasional training. Background As demands for domestic and international disaster response increase, experience in the field has begun to challenge traditional assumptions that response to mass casualty events requires little specialist training. The need for a “streamlined process of safe medical team deployment into disaster regions”1 is generally accepted and, in Australia, the emergence of national humanitarian aid training has begun to respond to this gap. However, calls for a national framework for disaster health education2 haven’t received much traction. Methods A critical analysis of the peer reviewed and grey literature on the core components/competencies and training methods required to prepare Australian paramedics to contribute to effective health disaster response has been conducted. Research from the past 10 years has been examined along with federal and state policy with regard to paramedic disaster education. Results The literature shows that education and training for disaster response is variable and that an evidence based study specifically designed to outline sets of core competencies for Australian health care professionals has never been undertaken. While such competencies in disaster response have been developed for the American paradigm it is suggested that disaster response within the Australian context is somewhat different to that of the US, and therefore a gap in the current knowledge base exists. Conclusion Further research is needed to develop core competencies specific to Australian paramedics in order to standardise teaching in the area of health disaster management. Until this occurs the task of evaluating or creating disaster curricula that adequately prepares and maintains paramedics for an effective all hazards disaster response is seen as largely unattainable.

Evaluation of cooperative exploration strategies using full system simulation

This paper compares different state-of-the-art exploration strategies for teams of mobile robots exploring an unknown environment. The goal is to help in determining a best strategy for a given multi-robot scenario and optimization target. Experiments are done in a 2D-simulation environment with 5 robots that are equipped with a horizontal laser range finder. Required components like SLAM, path planning and obstacle avoidance of every robot are included in a full-system simulation. To evaluate different strategies the time to finish exploration, the number of measurements that have been integrated into the map and the development in size of the explored area over time are used. The results of extensive test runs on three environments with different characteristics show that simple strategies can perform fairly well in many situations but specialized strategies can improve performance with regards to their targeted evaluation measure.

The State of Queensland Charities: An examination of the first Annual Information Statements of charities operating in Queensland

The advent of the Australian Charities and Not-for-profits Commission (ACNC) in 2012 and submission of Annual Information Statements (AIS) in 2013 by those charities which registered with them, have allowed new measures to be taken of charities and their activities. This report examines the filed AIS data for Queensland charities and compares it with the overall Australian population of charities.

A normal-distribution based rating aggregation method for generating product reputations

With the extensive use of rating systems in the web, and their significance in decision making process by users, the need for more accurate aggregation methods has emerged. The Naïve aggregation method, using the simple mean, is not adequate anymore in providing accurate reputation scores for items [6 ], hence, several researches where conducted in order to provide more accurate alternative aggregation methods. Most of the current reputation models do not consider the distribution of ratings across the different possible ratings values. In this paper, we propose a novel reputation model, which generates more accurate reputation scores for items by deploying the normal distribution over ratings. Experiments show promising results for our proposed model over state-of-the-art ones on sparse and dense datasets.

An archival analysis of green information technology: The current state and future directions

This thesis provides a review of 199 papers published on Green IT/IS between 2007−2014, in order to present taxonomy of segments in Green IT/IS publications, where the segments are later used for multiple analyses to facilitate future research and to provide a retrospective analysis of existing knowledge and gaps thereof. This research also attempts to make a unique contribution to our understanding of Green IT/IS, by consolidating papers it observes current patterns of literature through approach analysis and segmentation, as well as allocating studies to the technology, process, or outcome (TPO) stage. Highlighting the necessity of a consolidated approach, these classification systems have been combined into a TPO matrix so that the studies could be arranged according to which stage of the Green IT/IS cycle they were focused on. We believe that these analyses will provide a solid platform from which future Green IT/IS research can be launched.

The role of regulatory frameworks and non-state actors in creating sustainable diets

This presentation discussed the growing recognition of sustainable diets at international governance levels and how this reflects the challenges and win-win opportunities of living within our ecological limits. I assert that sustainable diets provide an example of how living within our ecological limits would actually make us better off even apart from environmental benefits. After determining whether Australians’ generally have a sustainable diet, I outlined how Australian regulators are attempting to address sustainable diets. I argued that the personal responsibility approach coupled with the focus on preventing or reducing overweight and obesity levels are proving incapable of bringing about long-term sustainable diets that will contribute to the health and well-being of Australian people.

Automation of underground truck haulage

This paper reviews the state-of-the-art in the automation of underground truck haulage. Past attempts at automating LHDs and haul trucks are described and their particular strengths and weaknesses are listed. We argue that the simple auto-tramming systems currently being commercialised, that follow rail-type guides placed along the back, cannot match the performance, flexibility and reliability of systems based on modern mobile robotic principles. In addition, the lack of collision detection research in the underground environment is highlighted.