Baden Hughes

Metadata quality evaluation: experience from the open language archives community

We describe the motivation, design and implementation of an infrastructure to support metadata quality assessment within a specialised Open Archives Initiative (OAI) sub-domain, the Open Language Archives Community (OLAC). While services for structural validation of metadata are widely used, there is little corresponding work regarding services which evaluate the semantic and syntactic content of metadata from a qualitative perspective. We posit that any measure of metadata quality benefits from both contextual and referential assessment – metadata on a per record and per collection basis is legitimately assessed against the baseline of broader community practice, as well as for compliance to any external standard. In this paper we describe the implementation of a metadata quality assessment scheme, and the corresponding interfaces to the evaluation tool.

A situated learning perspective on learning object design

Recent work suggests that learning object design can be improved by greater integration of instructional design, learning theory and software development methodologies. Despite this, there is a lack of research in the field that seeks to establish an association between the contextualised nature of learning object design and empirical properties of learner-computer interaction. In addressing this issue, we argue for a situated learning perspective on learning object design. Using the CASE framework as an exemplar of situated learning, we describe an holistic approach to eliciting socio-cultural properties of learning objects.

Frontiers in Linguistic Annotation for Lower-Density Languages

The languages that are most commonly subject to linguistic annotation on a large scale tend to be those with the largest populations or with recent histories of linguistic scholarship. In this paper we discuss the problems associated with lower-density languages in the context of the development of linguistically annotated resources. We frame our work with three key questions regarding the definition of lower-density languages; increasing available resources and reducing data requirements. A number of steps forward are identified for increasing the number lower-density language corpora with linguistic annotations.

Grid-enabling natural language engineering by stealth

We describe a proposal for an extensible, componentbased software architecture for natural language engineering applications. Our model leverages existing linguistic resource description and discovery mechanisms based on extended Dublin Core metadata. In addition, application design is flexible, allowing disparate components to be combined to suit the overall application functionality. An application specification language provides abstraction from the programming environment and allows ease of interface with computational grids via a broker.

Grid-based indexing of a newswire corpus

In this paper we report experience in the use of computational grids in the domain of natural language processing, particularly in the area of information extraction, to create query indices for information retrieval tasks. Given the prevalence of large corpora in the natural language processing domain, computational grids offer significant utility to researchers in the domain who are reaching the bounds of computational efficiency. We leverage the affinities between the segmented data sources prevalent in natural language processing and the parallelisation model from the grid domain. The experiment reported here is a large-scale newswire corpus indexing task, with the goal to efficiently create a queryable index of the entire corpus. By parallelising the indexing task and executing it on an Australian computational grid, we observe overall performance improvement of a 2.26x speedup over the same experiment on a single computational node. In addition to reporting the raw performance impact, we reflect on a number of interesting points discovered during the execution of the experiments and propose a number of new requirements for grid middleware.

A classification-based framework for learning object assembly

Relations between learning outcomes and the learning objects which are assembled to facilitate their achievement are the subject of increasingly prevalent investigation, particularly with approaches which advocate the aggregation of learning objects as complex constituencies for achieving learning outcomes. From the perspective of situated learning, we show how the CASE framework imbues learning objects with a closed set of properties which can be classified and aggregated into learning object assemblies in a principled fashion. We argue that the computational and pedagogical tractability of this model provides a new insight into learning object evaluation, and hence learning outcomes.

Pattern-Based End-User Development with Learning Objects

Previous research in Learning Object design has focused on automatic component selection and aggregation based on derived metadata relationships. However, research has yet to examine how end-user developers subsequently visualise and integrate such relationships during system design and use. In this paper we describe an approach and architecture for pattern-based design in End-User Development contexts using LO metadata. We discuss how existing approaches and techniques from machine learning can be integrated within our architecture to provide localised and distributed adaptation to EUD practices.

Semantic Decomposition of Character Encodings for Linguistic Knowledge Discovery

Analysis and knowledge representation of linguistic objects tends to focus on larger units (e.g. words) than print medium characters. We analyse characters as linguistic objects in their own right, with meaning, structure and form. Characters have meaning (the symbols of the International Phonetic Alphabet denote phonetic categories, the character represented by the glyph ‘∪’ denotes set union), structure (they are composed of stems and parts such as descenders or diacritics or are ligatures), and form (they have a mapping to visual glyphs). Character encoding initatives such as Unicode tend to concentrate on the structure and form of characters and ignore their meaning in the sense discussed here. We suggest that our approach of including semantic decomposition and defining font-based namespaces for semantic character domains provides a long-term perspective of interoperability and tractability with regard to data-mining over characters by integrating information about characters into a coherent semiotically-based ontology. We demonstrate these principles in a case study of the International Phonetic Alphabet.

CASE: a framework for evaluating learner-computer interaction in Computer-Assisted Language learning

Socio-cultural theories of language learning examine the impact of social interaction within cultural environments upon cognitive development and learner performance. Such theories emphasise the emergence of learner strategies and subsequent proficiency through involvement in continuously unravelling situated activities. However, this view is rarely expressed within Computer-Assisted Language Learning where practitioners often fail to consider implications of system design. This is in part due to the lack of a holistic and integrative framework for capturing, modelling and evaluating cognitive and social requirements of learner-computer interaction. In response, we propose the CASE (Cognition, Activity, Social Organisation and Environment) framework, and explore its application to a historical study on Computer-Assisted Language Learning software development. In sum, we argue that the CASE approach will greatly assist Computer-Assisted Language Learning initiatives in quality-driven system design.

Metadata Challenges for Situational Properties of Learning Objects

The abstraction of free-standing metadata describing learning objects is typified by an analytical model which primarily focuses on the encoding of discrete properties pertaining to the learning object itself. This process often neglects the contextual properties derived from the situations in which the learning object can be used optimally. We report the derivation of situated properties for learning objects through the use of the CASE framework and explore how these properties are best encoded.