A sociocultural analysis of language learning: New forms of literacy practices in a language and culture awareness programme

This paper reports on a sociocultural study conducted in a Catholic primary school in the Australian outback and provides insights into how policy related to Languages Other Than English (LOTE) programmes is implemented in a specific location and interwoven within the literacy practices of children, parents and teachers. A case study that tracked a Year Four student's learning and development during a Language and Culture Awareness Programme is discussed within a discourse of cultural and linguistic practices. Significant aspects of the student's learning related to a phenomenon called multi-tiered scaffolding temporarily disrupted the established literacy practices in the school community. Implications of the research for second-language teaching and learning in Australian primary schools are elaborated.

A sociocultural analysis of learning to teach

This paper addresses the question of how teachers learn from experience during their pre-service course and early years of teaching. It outlines a theoretical framework that may help us better understand how teachers' professional identities emerge in practice. The framework adapts Vygotsky's Zone of Proximal Development, and Valsiner's Zone of Free Movement and Zone of Promoted Action, to the field of teacher education. The framework is used to analyse the pre-service and initial professional experiences of a novice secondary mathematics teacher in integrating computer and graphics calculator technologies into his classroom practice. (Contains 1 figure.) [For complete proceedings, see ED496848.]

Playing in continuous spaces: Some analysis and extension of population-based incremental learning

As an alternative to traditional evolutionary algorithms (EAs), population-based incremental learning (PBIL) maintains a probabilistic model of the best individual(s). Originally, PBIL was applied in binary search spaces. Recently, some work has been done to extend it to continuous spaces. In this paper, we review two such extensions of PBIL. An improved version of the PBIL based on Gaussian model is proposed that combines two main features: a new updating rule that takes into account all the individuals and their fitness values and a self-adaptive learning rate parameter. Furthermore, a new continuous PBIL employing a histogram probabilistic model is proposed. Some experiments results are presented that highlight the features of the new algorithms.

Port community learning needs:analysis and design

The port industry is facing a dramatic wave of changes that have transformed the structure of the industry. Modern seaports are increasingly shifting from a “hardware-based” approach towards “knowhow intensive” configuration. In this context knowledge resources, learning processes and training initiatives increasingly represent key elements to guarantee the quality of service supplied and hence the competitiveness of modern seaport communities. This paper describes the learning needs analysis conducted amongst key port community actors in three ports in the south east of Ireland during 2005 in the context of the I-Sea.Net project. It goes on to describe the learning requirements report and the training design carried out based on this analysis.

Learning higher-level features with convolutional restricted Boltzmann machines for sentiment analysis

In recent years, learning word vector representations has attracted much interest in Natural Language Processing. Word representations or embeddings learned using unsupervised methods help addressing the problem of traditional bag-of-word approaches which fail to capture contextual semantics. In this paper we go beyond the vector representations at the word level and propose a novel framework that learns higher-level feature representations of n-grams, phrases and sentences using a deep neural network built from stacked Convolutional Restricted Boltzmann Machines (CRBMs). These representations have been shown to map syntactically and semantically related n-grams to closeby locations in the hidden feature space. We have experimented to additionally incorporate these higher-level features into supervised classifier training for two sentiment analysis tasks: subjectivity classification and sentiment classification. Our results have demonstrated the success of our proposed framework with 4% improvement in accuracy observed for subjectivity classification and improved the results achieved for sentiment classification over models trained without our higher level features.

The effectiveness of workplace coaching:a meta-analysis of learning and performance outcomes from coaching

This study presents a meta-analysis synthesizing the existing research on the effectiveness of workplace coaching. We exclusively explore workplace coaching provided by internal or external coaches and therefore exclude cases of manager-subordinate and peer coaching. We propose a framework of potential outcomes from coaching in organizations, which we examine meta-analytically (k = 17). Our analyses indicated that coaching had positive effects on organizational outcomes overall (δ = 0.36), and on specific forms of outcome criteria (skill-based δ = 0.28; affective δ = 0.51; individual-level results δ = 1.24). We also examined moderation by a number of coaching practice factors (use of multisource feedback; type of coach; coaching format; longevity of coaching). Our analyses of practice moderators indicated a significant moderation of effect size for type of coach (with effects being stronger for internal coaches compared to external coaches) and use of multisource feedback (with the use of multisource feedback resulting in smaller positive effects). We found no moderation of effect size by coaching format (comparing face-to-face, with blended face-to-face and e-coaching) or duration of coaching (number of sessions or longevity of intervention). The effect sizes give support to the potential utility of coaching in organizations. Implications for coaching research and practice are discussed.

A Comparative Analysis of Predictive Learning Algorithms on High-Dimensional Microarray Cancer Data

This research evaluates pattern recognition techniques on a subclass of big data where the dimensionality of the input space (p) is much larger than the number of observations (n). Specifically, we evaluate massive gene expression microarray cancer data where the ratio κ is less than one. We explore the statistical and computational challenges inherent in these high dimensional low sample size (HDLSS) problems and present statistical machine learning methods used to tackle and circumvent these difficulties. Regularization and kernel algorithms were explored in this research using seven datasets where κ < 1. These techniques require special attention to tuning necessitating several extensions of cross-validation to be investigated to support better predictive performance. While no single algorithm was universally the best predictor, the regularization technique produced lower test errors in five of the seven datasets studied.