Scaffolding for integrated reflection

This paper presents some theoretical perspectives that might inform the design and development of information and communications technology (ICT) tools to support integrated (in-session) reflection and deep learning during e-learning. The role of why questioning provides the focus of discussion and is informed by the literature on critical thinking, sense-making, and reflective practice, as well as recent developments in knowledge management, computational linguistics and automated question generation. It is argued that there exists enormous scope for the development of ICT scaffolding targeted at supporting reflective practice during e-learning. The first generations of e-Portfolio tools provide some evidence for the significance of the benefits of integrating reflection into the design of ICT systems; however, following the review of a number of such systems, as well as a range of ICT applications and services designed to support e-learning, it is argued that the scope of implementation is limited.

Cognitive engagement and questioning online

This chapter discusses a range of issues associated with supporting inquiry and deep reasoning while utilising information and communications technology (ICT). The role of questioning in critical thinking and reflection is considered in the context of scaffolding and new opportunities for ICT-enabled scaffolding identified. In particular, why-questioning provides a key point of focus and is presented as an important consideration in the design of systems that not only require cognitive engagement but aim to nurture it. Advances in automated question generation within intelligent tutoring systems are shown to hold promise for both teaching and learning in a range of other applications. While shortening attention spans appear to be a hazard of engaging with digital media cognitive engagement is presented as something with broader scope than attention span and is best conceived of as a crucible within which a rich mix of cognitive activities take place and from which new knowledge is created.

Responding to the 5Rs : an alternate perspective of Slowmation

This paper is a response to Hoban and Neilsen's (2010) Five Rs model for understanding how learners engage with slowmation. An alternative model (the Learning MMAEPER Model) that builds on the 5Rs model is explained in terms of its use in secondary science preservice teacher education. To probe into the surface and deep learning that can occur during the creation of a slowmation, the learning and relearning model is explored in terms of learning elements. This model can assist teachers to monitor the learning of their students and direct them to a deeper understanding of science concepts.

Whole-of-Curriculum Design in Law

In this book teaching professionalism is characterised by the scholarly underpinning of each contribution; and every contribution provides a rich resource for enhancing teaching practice. The critical concerns for legal education have been identified and discussed: curriculum design that includes graduate attributes; embedding specific attributes across the curriculum; empowering students to learn; academic teamwork to manage large student cohorts; first year and final year transition strategies; tracking students' personal development through the use of ePortfolio; assessment strategies; improving student well-being and promoting resilience; teaching practice to achieve deep learning; flexibility in delivery; the use of Web 2.0 technology; and understanding the 21st century student.

Mathematics, programming, and STEM

Learning mathematics is a complex and dynamic process. In this paper, the authors adopt a semiotic framework (Yeh & Nason, 2004) and highlight programming as one of the main aspects of the semiosis or meaning-making for the learning of mathematics. During a 10-week teaching experiment, mathematical meaning-making was enriched when primary students wrote Logo programs to create 3D virtual worlds. The analysis of results found deep learning in mathematics, as well as in technology and engineering areas. This prompted a rethinking about the nature of learning mathematics and a need to employ and examine a more holistic learning approach for the learning in science, technology, engineering, and mathematics (STEM) areas.

Enhancing human action recognition with region proposals

Deep convolutional network models have dominated recent work in human action recognition as well as image classification. However, these methods are often unduly influenced by the image background, learning and exploiting the presence of cues in typical computer vision datasets. For unbiased robotics applications, the degree of variation and novelty in action backgrounds is far greater than in computer vision datasets. To address this challenge, we propose an “action region proposal” method that, informed by optical flow, extracts image regions likely to contain actions for input into the network both during training and testing. In a range of experiments, we demonstrate that manually segmenting the background is not enough; but through active action region proposals during training and testing, state-of-the-art or better performance can be achieved on individual spatial and temporal video components. Finally, we show by focusing attention through action region proposals, we can further improve upon the existing state-of-the-art in spatio-temporally fused action recognition performance.

Abstract concepts made real: A pilot study examining pedagogical approaches in thermodynamics tutorials

BACKGROUND OR CONTEXT Thermodynamics is a core concept for mechanical engineers yet notoriously difficult. Evidence suggests students struggle to understand and apply the core fundamental concepts of thermodynamics with analysis indicating a problem with student learning/engagement. A contributing factor is that thermodynamics is a ‘science involving concepts based on experiments’ (Mayhew 1990) with subject matter that cannot be completely defined a priori. To succeed, students must engage in a deep-holistic approach while taking ownership of their learning. The difficulty in achieving this often manifests itself in students ‘not getting’ the principles and declaring thermodynamics ‘hard’. PURPOSE OR GOAL Traditionally, students practice and “learn” the application of thermodynamics in their tutorials, however these do not consider prior conceptions (Holman & Pilling 2004). As ‘hands on’ learning is the desired outcome of tutorials it is pertinent to study methods of improving their efficacy. Within the Australian context, the format of thermodynamics tutorials has remained relatively unchanged over the decades, relying anecdotally on a primarily didactic pedagogical approach. Such approaches are not conducive to deep learning (Ramsden 2003) with students often disengaged from the learning process. Evidence suggests (Haglund & Jeppsson 2012), however, that a deeper level and ownership of learning can be achieved using a more constructivist approach for example through self generated analogies. This pilot study aimed to collect data to support the hypothesis that the ‘difficulty’ of thermodynamics is associated with the pedagogical approach of tutorials rather than actual difficulty in subject content or deficiency in students. APPROACH Successful application of thermodynamic principles requires solid knowledge of the core concepts. Typically, tutorial sessions guide students in this application. However, a lack of deep and comprehensive understanding can lead to student confusion in the applications resulting in the learning of the ‘process’ of application without understanding ‘why’. The aim of this study was to gain empirical data on student learning of both concepts and application, within thermodynamic tutorials. The approach taken for data collection and analysis was: - 1 Four concurrent tutorial streams were timetabled to examine student engagement/learning in traditional ‘didactic’ (3 weeks) and non-traditional (3 weeks). In each week, two of the selected four sessions were traditional and two non-traditional. This provided a control group for each week. - 2 The non-traditional tutorials involved activities designed to promote student-centered deep learning. Specific pedagogies employed were: self-generated analogies, constructivist, peer-to-peer learning, inquiry based learning, ownership of learning and active learning. - 3 After a three-week period, teaching styles of the selected groups was switched, to allow each group to experience both approaches with the same tutor. This also acted to mimimise any influence of tutor personality / style on the data. - 4 At the conclusion of the trial participants completed a ‘5 minute essay’ on how they liked the sessions, a small questionnaire, modelled on the modified (Christo & Hoang, 2013)SPQ designed by Biggs (1987) and a small formative quiz to gauge the level of learning achieved. DISCUSSION Preliminary results indicate that overall students respond positively to in class demonstrations (inquiry based learning), and active learning activities. Within the active learning exercises, the current data suggests students preferred individual rather than group or peer-to-peer activities. Preliminary results from the open-ended questions such as “What did you like most/least about this tutorial” and “do you have other comments on how this tutorial could better facilitate your learning”, however, indicated polarising views on the nontraditional tutorial. Some student’s responded that they really like the format and emphasis on understanding the concepts, while others were very vocal that that ‘hated’ the style and just wanted the solutions to be presented by the tutor. RECOMMENDATIONS/IMPLICATIONS/CONCLUSION Preliminary results indicated a mixed, but overall positive response by students with more collaborative tutorials employing tasks promoting inquiry based, peer-to-peer, active, and ownership of learning activities. Preliminary results from student feedback supports evidence that students learn differently, and running tutorials focusing on only one pedagogical approached (typically didactic) may not be beneficial to all students. Further, preliminary data suggests that the learning / teaching style of both students and tutor are important to promoting deep learning in students. Data collection is still ongoing and scheduled for completion at the end of First Semester (Australian academic calendar). The final paper will examine in more detail the results and analysis of this project.

R1STM: One-class support tensor machine with randomised kernel

Identifying unusual or anomalous patterns in an underlying dataset is an important but challenging task in many applications. The focus of the unsupervised anomaly detection literature has mostly been on vectorised data. However, many applications are more naturally described using higher-order tensor representations. Approaches that vectorise tensorial data can destroy the structural information encoded in the high-dimensional space, and lead to the problem of the curse of dimensionality. In this paper we present the first unsupervised tensorial anomaly detection method, along with a randomised version of our method. Our anomaly detection method, the One-class Support Tensor Machine (1STM), is a generalisation of conventional one-class Support Vector Machines to higher-order spaces. 1STM preserves the multiway structure of tensor data, while achieving significant improvement in accuracy and efficiency over conventional vectorised methods. We then leverage the theory of nonlinear random projections to propose the Randomised 1STM (R1STM). Our empirical analysis on several real and synthetic datasets shows that our R1STM algorithm delivers comparable or better accuracy to a state-of-the-art deep learning method and traditional kernelised approaches for anomaly detection, while being approximately 100 times faster in training and testing.

Deeper and wider fully convolutional network coupled with conditional random fields for scene labeling

Deep convolutional neural networks (DCNNs) have been employed in many computer vision tasks with great success due to their robustness in feature learning. One of the advantages of DCNNs is their representation robustness to object locations, which is useful for object recognition tasks. However, this also discards spatial information, which is useful when dealing with topological information of the image (e.g. scene labeling, face recognition). In this paper, we propose a deeper and wider network architecture to tackle the scene labeling task. The depth is achieved by incorporating predictions from multiple early layers of the DCNN. The width is achieved by combining multiple outputs of the network. We then further refine the parsing task by adopting graphical models (GMs) as a post-processing step to incorporate spatial and contextual information into the network. The new strategy for a deeper, wider convolutional network coupled with graphical models has shown promising results on the PASCAL-Context dataset.

Towards vision-based deep reinforcement learning for robotic motion control

This paper introduces a machine learning based system for controlling a robotic manipulator with visual perception only. The capability to autonomously learn robot controllers solely from raw-pixel images and without any prior knowledge of configuration is shown for the first time. We build upon the success of recent deep reinforcement learning and develop a system for learning target reaching with a three-joint robot manipulator using external visual observation. A Deep Q Network (DQN) was demonstrated to perform target reaching after training in simulation. Transferring the network to real hardware and real observation in a naive approach failed, but experiments show that the network works when replacing camera images with synthetic images.

The influences of conceptions of learning and learner self-concept on high school students' approaches to learning

This study investigated the mediating effect of learner selfconcept between conceptions of learning and students' approaches to learning using structural equation modelling. Data were collected using a modified version of Biggs' Learning Process Questionnaire, together with the recently developed 'What is Learning Survey' and 'Learner Self-Concept Scale'. A sample of 355 high school students participated in the study. Results indicate that learner self-concept does mediate between conceptions of meaning and approaches to learning. Students who adopted a deep approach liked learning new things and indirectly viewed learning as experiential, involving social interaction and directly viewed learning as personal development. Implications for teachers are discussed, with consideration given to appropriate classroom practice.

Elementary school students’ learner self-concept, academic self-concepts and approaches to learning

The present study investigated the relationships between academic selfconcepts, learner self-concept, and approaches to learning in elementary school students. A sample of 580 Australian Grade 6 and 7 school students with a mean age of 10.7 years participated in the study. Weak negative correlations between learner self-concepts and surface approaches to learning were identiŽ ed. In contrast, deep approaches for both boys and girls showed the highest positive correlations with school self-concept and learning self-concept. Only slight variations in these Ž gures were found between boys and girls.

Students’ perceptions of teaching and learning: the influence of students’ approaches to learning and teachers’ approaches to teaching

Nearly 500 secondary students in 24 classes were surveyed and four students in each class interviewed concerning their approaches to learning and perceptions of their classroom environment. While interviewed students with deep approaches to learning generally demonstrated a more sophisticated understanding of the learning opportunities offered to them than did students with surface approaches, teaching strategies also influenced students' perceptions. When teachers focused strongly on actively engaging students and creating a supportive environment, students with both deep and surface approaches focused on student-centred aspects of the class. In contrast, when traditional expository teaching methods were used exclusively, students with deep and surface approaches both focused on transmission and reproduction.

Creating deep and broad change through research and systems approaches in early childhood education for sustainability

The co-authors raise two matters they consider essential for the future development of ECEfS. The first is the need to create deep foundations based in research. At a time of increasing practitioner interest, research in ECEfS is meagre. A robust research community is crucial to support quality in curriculum and pedagogy, and to promote learning and innovation in thinking and practice. The second 'essential' for the expansion and uptake of ECEfS is broad systemic change. All level within the early childhood education system - individual teachers and classrooms, whole centres and schools, professional associations and networks, accreditation and employing authorities, and teacher educators - must work together to create and reinforce the cultural and educational changes required for sustainability. This chapter provides explanations of processes to engender systemic change. It illustrates a systems approach, with reference to a recent study focused on embedding EfS into teacher education. This study emphasises the apparent contradiction that the answer to large-scale reform lies with small-scale reforms that build capacity and make connections.