The effect of language models on phonetic decoding for spoken term detection

Spoken term detection (STD) popularly involves performing word or sub-word level speech recognition and indexing the result. This work challenges the assumption that improved speech recognition accuracy implies better indexing for STD. Using an index derived from phone lattices, this paper examines the effect of language model selection on the relationship between phone recognition accuracy and STD accuracy. Results suggest that language models usually improve phone recognition accuracy but their inclusion does not always translate to improved STD accuracy. The findings suggest that using phone recognition accuracy to measure the quality of an STD index can be problematic, and highlight the need for an alternative that is more closely aligned with the goals of the specific detection task.

Spoken term detection using fast phonetic decoding

While spoken term detection (STD) systems based on word indices provide good accuracy, there are several practical applications where it is infeasible or too costly to employ an LVCSR engine. An STD system is presented, which is designed to incorporate a fast phonetic decoding front-end and be robust to decoding errors whilst still allowing for rapid search speeds. This goal is achieved through mono-phone open-loop decoding coupled with fast hierarchical phone lattice search. Results demonstrate that an STD system that is designed with the constraint of a fast and simple phonetic decoding front-end requires a compromise to be made between search speed and search accuracy.

Look before you leap : the epistemic violence that sometimes hides behind the word "inclusion"

This paper demonstrates how Indigenous Studies is controlled in some Australian universities in ways that continue the marginalisation, denigration and exploitation of Indigenous peoples. Moreover, it shows how the engagement of white notions of “inclusion” can result in the maintenance of racism, systemic marginalisation, white race privilege and radicalised subjectivity. A case study will be utilised which draws from the experience of two Indigenous scholars who were invited to be part of a panel to review one Australian university’s plan and courses in Indigenous studies. The case study offers the opportunity to destabilise the relationships between oppression and privilege and the epistemology that maintains them. The paper argues for the need to examine exactly what is being offered when universities provide opportunities for “inclusion”.

Word Segmentation for Chinese Wikipedia Using N-Gram Mutual Information

In this paper, we propose an unsupervised segmentation approach, named "n-gram mutual information", or NGMI, which is used to segment Chinese documents into n-character words or phrases, using language statistics drawn from the Chinese Wikipedia corpus. The approach alleviates the tremendous effort that is required in preparing and maintaining the manually segmented Chinese text for training purposes, and manually maintaining ever expanding lexicons. Previously, mutual information was used to achieve automated segmentation into 2-character words. The NGMI approach extends the approach to handle longer n-character words. Experiments with heterogeneous documents from the Chinese Wikipedia collection show good results.

Word-perfect Gatsby overcomes its flaws

Review of 'Gatz', Elevator Repair Company / Brisbane Powerhouse, published in The Australian, 12 May 2009.

Thai word segmentation with hidden Markov Model and decision tree

The Thai written language is one of the languages that does not have word boundaries. In order to discover the meaning of the document, all texts must be separated into syllables, words, sentences, and paragraphs. This paper develops a novel method to segment the Thai text by combining a non-dictionary based technique with a dictionary-based technique. This method first applies the Thai language grammar rules to the text for identifying syllables. The hidden Markov model is then used for merging possible syllables into words. The identified words are verified with a lexical dictionary and a decision tree is employed to discover the words unidentified by the lexical dictionary. Documents used in the litigation process of Thai court proceedings have been used in experiments. The results which are segmented words, obtained by the proposed method outperform the results obtained by other existing methods.

Improvement to Chinese information retrieval by incorporating word segmentation and query expansion

The increasing diversity of the Internet has created a vast number of multilingual resources on the Web. A huge number of these documents are written in various languages other than English. Consequently, the demand for searching in non-English languages is growing exponentially. It is desirable that a search engine can search for information over collections of documents in other languages. This research investigates the techniques for developing high-quality Chinese information retrieval systems. A distinctive feature of Chinese text is that a Chinese document is a sequence of Chinese characters with no space or boundary between Chinese words. This feature makes Chinese information retrieval more difficult since a retrieved document which contains the query term as a sequence of Chinese characters may not be really relevant to the query since the query term (as a sequence Chinese characters) may not be a valid Chinese word in that documents. On the other hand, a document that is actually relevant may not be retrieved because it does not contain the query sequence but contains other relevant words. In this research, we propose two approaches to deal with the problems. In the first approach, we propose a hybrid Chinese information retrieval model by incorporating word-based techniques with the traditional character-based techniques. The aim of this approach is to investigate the influence of Chinese segmentation on the performance of Chinese information retrieval. Two ranking methods are proposed to rank retrieved documents based on the relevancy to the query calculated by combining character-based ranking and word-based ranking. Our experimental results show that Chinese segmentation can improve the performance of Chinese information retrieval, but the improvement is not significant if it incorporates only Chinese segmentation with the traditional character-based approach. In the second approach, we propose a novel query expansion method which applies text mining techniques in order to find the most relevant words to extend the query. Unlike most existing query expansion methods, which generally select the highly frequent indexing terms from the retrieved documents to expand the query. In our approach, we utilize text mining techniques to find patterns from the retrieved documents that highly correlate with the query term and then use the relevant words in the patterns to expand the original query. This research project develops and implements a Chinese information retrieval system for evaluating the proposed approaches. There are two stages in the experiments. The first stage is to investigate if high accuracy segmentation can make an improvement to Chinese information retrieval. In the second stage, a text mining based query expansion approach is implemented and a further experiment has been done to compare its performance with the standard Rocchio approach with the proposed text mining based query expansion method. The NTCIR5 Chinese collections are used in the experiments. The experiment results show that by incorporating the text mining based query expansion with the hybrid model, significant improvement has been achieved in both precision and recall assessments.

Changing lanes : an exploration of the journey from dance through choreography to directing and the spoken word

This paper reveals a journey of theatrical exploration. It is a journey of enquiry and investigation backed by a vigorous, direct and dense professional history of creative work.

The nonconceptual gateway to early word learning

My research investigates why nouns are learned disproportionately more frequently than other kinds of words during early language acquisition (Gentner, 1982; Gleitman, et al., 2004). This question must be considered in the context of cognitive development in general. Infants have two major streams of environmental information to make meaningful: perceptual and linguistic. Perceptual information flows in from the senses and is processed into symbolic representations by the primitive language of thought (Fodor, 1975). These symbolic representations are then linked to linguistic input to enable language comprehension and ultimately production. Yet, how exactly does perceptual information become conceptualized? Although this question is difficult, there has been progress. One way that children might have an easier job is if they have structures that simplify the data. Thus, if particular sorts of perceptual information could be separated from the mass of input, then it would be easier for children to refer to those specific things when learning words (Spelke, 1990; Pylyshyn, 2003). It would be easier still, if linguistic input was segmented in predictable ways (Gentner, 1982; Gleitman, et al., 2004) Unfortunately the frequency of patterns in lexical or grammatical input cannot explain the cross-cultural and cross-linguistic tendency to favor nouns over verbs and predicates. There are three examples of this failure: 1) a wide variety of nouns are uttered less frequently than a smaller number of verbs and yet are learnt far more easily (Gentner, 1982); 2) word order and morphological transparency offer no insight when you contrast the sentence structures and word inflections of different languages (Slobin, 1973) and 3) particular language teaching behaviors (e.g. pointing at objects and repeating names for them) have little impact on children's tendency to prefer concrete nouns in their first fifty words (Newport, et al., 1977). Although the linguistic solution appears problematic, there has been increasing evidence that the early visual system does indeed segment perceptual information in specific ways before the conscious mind begins to intervene (Pylyshyn, 2003). I argue that nouns are easier to learn because their referents directly connect with innate features of the perceptual faculty. This hypothesis stems from work done on visual indexes by Zenon Pylyshyn (2001, 2003). Pylyshyn argues that the early visual system (the architecture of the "vision module") segments perceptual data into pre-conceptual proto-objects called FINSTs. FINSTs typically correspond to physical things such as Spelke objects (Spelke, 1990). Hence, before conceptualization, visual objects are picked out by the perceptual system demonstratively, like a finger pointing indicating ‘this’ or ‘that’. I suggest that this primitive system of demonstration elaborates on Gareth Evan's (1982) theory of nonconceptual content. Nouns are learnt first because their referents attract demonstrative visual indexes. This theory also explains why infants less often name stationary objects such as plate or table, but do name things that attract the focal attention of the early visual system, i.e., small objects that move, such as ‘dog’ or ‘ball’. This view leaves open the question how blind children learn words for visible objects and why children learn category nouns (e.g. 'dog'), rather than proper nouns (e.g. 'Fido') or higher taxonomic distinctions (e.g. 'animal').

Fast and accurate phonetic spoken term detection

For the first time in human history, large volumes of spoken audio are being broadcast, made available on the internet, archived, and monitored for surveillance every day. New technologies are urgently required to unlock these vast and powerful stores of information. Spoken Term Detection (STD) systems provide access to speech collections by detecting individual occurrences of specified search terms. The aim of this work is to develop improved STD solutions based on phonetic indexing. In particular, this work aims to develop phonetic STD systems for applications that require open-vocabulary search, fast indexing and search speeds, and accurate term detection. Within this scope, novel contributions are made within two research themes, that is, accommodating phone recognition errors and, secondly, modelling uncertainty with probabilistic scores. A state-of-the-art Dynamic Match Lattice Spotting (DMLS) system is used to address the problem of accommodating phone recognition errors with approximate phone sequence matching. Extensive experimentation on the use of DMLS is carried out and a number of novel enhancements are developed that provide for faster indexing, faster search, and improved accuracy. Firstly, a novel comparison of methods for deriving a phone error cost model is presented to improve STD accuracy, resulting in up to a 33% improvement in the Figure of Merit. A method is also presented for drastically increasing the speed of DMLS search by at least an order of magnitude with no loss in search accuracy. An investigation is then presented of the effects of increasing indexing speed for DMLS, by using simpler modelling during phone decoding, with results highlighting the trade-off between indexing speed, search speed and search accuracy. The Figure of Merit is further improved by up to 25% using a novel proposal to utilise word-level language modelling during DMLS indexing. Analysis shows that this use of language modelling can, however, be unhelpful or even disadvantageous for terms with a very low language model probability. The DMLS approach to STD involves generating an index of phone sequences using phone recognition. An alternative approach to phonetic STD is also investigated that instead indexes probabilistic acoustic scores in the form of a posterior-feature matrix. A state-of-the-art system is described and its use for STD is explored through several experiments on spontaneous conversational telephone speech. A novel technique and framework is proposed for discriminatively training such a system to directly maximise the Figure of Merit. This results in a 13% improvement in the Figure of Merit on held-out data. The framework is also found to be particularly useful for index compression in conjunction with the proposed optimisation technique, providing for a substantial index compression factor in addition to an overall gain in the Figure of Merit. These contributions significantly advance the state-of-the-art in phonetic STD, by improving the utility of such systems in a wide range of applications.

Entertaining the E Word : Barrie Kosky's On Ecstasy [A Review]

A review of Barrie Kosky's essay, On Ecstasy : Most of us describe the E word as a pleasant, out of this world experience—a type of boundless, artificial joy, deliberately induced by some kind of technicoloured drug. For others, it is that “lovey dovey” feeling. A spinning ceiling. Anything Lindt. For sensualist and soup connoisseur Barrie Kosky, it is easier than this. Being On Ecstasy involves, quite simply, his grandmother's chicken specialty—something warm and golden, surrendered with vegetables and a side of transcendental bliss. “A soup that took you to the beginning and end of time itself. A dazzling, pure, clear rhapsody” (7).