The research in this dissertation consists of two parts: An active learning algorithm for hierarchical labels and an embedding-based retrieval algorithm. In the first part, we present a new approach for learning hierarchically decomposable concepts. The approach learns a high-level classifier (e.g., location vs. non-location) by separately…

The diversity of engineering education research provides an opportunity for cross-fertilisation of ideas and creativity, but it also can result in fragmentation of the field and duplication of effort. One solution is to establish a standardised taxonomy of engineering education terms to map the field and communicate and connect research…

This paper provides an overview of specifications and standards for metadata relating to learning materials. It is structured to present first the currently established metadata schemas in use today (specifically the IEEE LOM and Dublin Core metadata), then to examine current developments and activities before looking at what might be the future…

Introduction: We define the results list problem in information search and suggest the "associative index model", an ad-hoc, user-derived indexing solution based on Vannevar Bush's description of an associative indexing approach for his memex machine. We further define what selection means in indexing terms with reference to Charles…

Since the experiments of Saffran et al. [Saffran, J., Aslin, R., & Newport, E. (1996). Statistical learning in 8-month-old infants. "Science," 274, 1926-1928], there has been a great deal of interest in the question of how statistical regularities in the speech stream might be used by infants to begin to identify individual words. In this work, we…

Due to the popularity of 3D graphics in animation and games, usage of 3D geometry deformable models increases dramatically. Despite their growing importance, these models are difficult and time consuming to build. A distance learning system for the construction of these models could greatly facilitate students to learn and practice at different…

Present day shortcomings in information retrieval are the results of a failure to properly contend with the problem of data representation. The index provides the necessary linkage between a multiplicity of sources and a single receiver. Whether considering the source/document-space interface or the query/index interface, the elements of the…

The retrieval decision problem is considered from the viewpoint of a decision theory approach. A threshold rule based on earlier rules for indexing decisions is considered and analyzed for retrieval decisions as a measure of retrieval performance. The threshold rule is seen as a good descriptive design measure of what a reasonable retrieval system…

The retrieval effectiveness of Keywork-In-Context (KWIC) indexing is sought using a simple search simulation which constrains the user to approach a KWIC index with subject keywords and references dictated by a conventional index. A comparison of several KWIC index studies, using various methodologies, shows retrievability ratios of 35 percent -…

Explores the use of linear models and a combination of neural networks and linear classifiers to create a hybrid hierarchical mixture of experts (HME) model. Results confirm that using the hierarchical structure of the classification vocabulary improves categorization performance. (AEF)

Reports on a theoretical model of structured document indexing and retrieval based on the Dempster-Shafer Theory of Evidence. Highlights include passage retrieval, hypertext, relevance, information needs, implementation of the model, an evaluation scheme designed to test the model's effectiveness, and results of experiments that investigated…

Reports on two experiments investigating the relationship between topical subject expertise and the assessment of relevance. Results indicate that topical subject expertise enhances effects of the semantic distance model (SDM), and that relevance assessment is a contingent, psychological phenomenon that is part of the human-perception task of…

Describes the development of an information retrieval system designed for nonspecialist users that is based on the binary vector model. The syntactic structure of phrases used for indexing is examined, queries using an experimental collection of documents are described, and precision values are examined. (19 references) (LRW)

Presents a set-oriented view of information retrieval and defines a precise model to capture this approach. A number of consequences of the model, such as the limitations implied by a finite index vocabulary, are examined, and the various ways in which the set orientation can influence thinking about information retrieval is discussed. (14…

Discussion of the use of Latent Semantic Indexing to determine relevancy in information retrieval focuses on statistical regression and Bayesian methods. Topics include keyword searching; a multiple regression model; how the regression model can aid search methods; and limitations of this approach, including complexity, linearity, and…