Computing and Information Systems - Theses
Permanent URI for this collection
Search Results
1 - 3 of 3
-
ItemScaling conditional random fields for natural language processing( 2007-01)This thesis deals with the use of Conditional Random Fields (CRFs; Lafferty et al. (2001)) for Natural Language Processing (NLP). CRFs are probabilistic models for sequence labelling which are particularly well suited to NLP. They have many compelling advantages over other popular models such as Hidden Markov Models and Maximum Entropy Markov Models (Rabiner, 1990; McCallum et al., 2001), and have been applied to a number of NLP tasks with considerable success (e.g., Sha and Pereira (2003) and Smith et al. (2005)). Despite their apparent success, CRFs suffer from two main failings. Firstly, they often over-fit the training sample. This is a consequence of their considerable expressive power, and can be limited by a prior over the model parameters (Sha and Pereira, 2003; Peng and McCallum, 2004). Their second failing is that the standard methods for CRF training are often very slow, sometimes requiring weeks of processing time. This efficiency problem is largely ignored in current literature, although in practise the cost of training prevents the application of CRFs to many new more complex tasks, and also prevents the use of densely connected graphs, which would allow for much richer feature sets. (For complete abstract open document)
-
ItemComputing relationships and relatedness between contextually diverse entities( 2011)When presented with a pair of entities such as a ball and a bat, a person may make the connection that both of these entities are involved in sport (e.g., the sports baseball or cricket, based on the individual's background), that the composition of the two entities is similar (e.g., a wooden ball and a wooden stick), or if the person is especially creative, a fancy dress ball where someone has come dressed as a bat. All of these connections are equally valid, but depending on the context the person is familiar with (e.g., sport, wooden objects, fancy dress), a particular connection may be more apparent to that person. From a computational perspective, identifying these relationships and calculating the level of relatedness of entity pairs requires consideration of all ways in which the entities are able to interact with one another. Existing approaches to identifying the relatedness of entities and the semantic relationships that exist between them fail to take into account the multiple diverse ways in which these entities may interact, and hence do not explore all potential ways in which entities may be related. In this thesis, I use the collaborative encyclopedia Wikipedia as the basis for the formulation of a measure of semantic relatedness that takes into account the contextual diversity of entities (called the Related Article Conceptual Overlap, or RACO, method), and describe several methods of relationship extraction that utilise the taxonomic structure of Wikipedia to identify pieces of text that describe relations between contextually diverse entities. I also describe the construction of a dataset of museum exhibit relatedness judgements used to evaluate the performance of RACO. I demonstrate that RACO outperforms state-of-the-art measures of semantic relatedness over a collection of contextually diverse entities (museum exhibits), and that the taxonomic structure of Wikipedia provides a basis for identifying valid relationships between contextually diverse entities. As this work is presented in regard to the domain of Cultural Heritage and using Wikipedia as a basis for representation, I additionally describe the process for adapting the principle of conceptual overlap for calculating semantic relatedness and the relationship extraction methods based on taxonomic links to alternate contextually diverse domains, and for use with other representational resources.
-
ItemOrthographic support for passing the reading hurdle in Japanese( 2010)Learning a second language is, for the most part, a day-in day-out struggle against the mountain of new vocabulary a learner must acquire. Furthermore, since the number of new words to learn is so great, learners must acquire them autonomously. Evidence suggests that for languages with writing systems, native-like vocabulary sizes are only developed through reading widely, and that reading is only fruitful once learners have acquired the core vocabulary required for it to become smooth. Learners of Japanese have an especially high barrier in the form of the Japanese writing system, in particular its use of kanji characters. Recent work on dictionary accessibility has focused on compensating for learner errors in pronouncing unknown words, however much difficulty remains. This thesis uses the rich visual nature of the Japanese orthography to support the study of vocabulary in several ways. Firstly, it proposes a range of kanji similarity measures and evaluates them over several new data sets, finding that the stroke edit distance and tree edit distance metrics best approximate human judgements. Secondly, it uses stroke edit distance construct a model of kanji misrecognition, which we use as the basis for a new form of kanji search by similarity. Analysing query logs, we find that this new form of search was rapidly adopted by users, indicating its utility. We finally combine kanji confusion and pronunciation models into a new adaptive testing platform, Kanji Tester, modelled after aspects of the Japanese Language Proficiency Test. As the user tests themselves, the system adapts to their error patterns and uses this information to make future tests more difficult. Investigating logs of use, we find a weak positive correlation between ability estimates and time the system has been used. Furthermore, our adaptive models generated questions which were significantly more difficult than their control counterparts. Overall, these contributions make a concerted effort to improve tools for learner self-study, so that learners can successfully overcome the reading hurdle and propel themselves towards greater proficiency. The data collected from these tools also forms a useful basis for further study of learner error and vocabulary development.