Language model combination and adaptation using weighted finite state transducers

In speech recognition systems language model (LMs) are often constructed by training and combining multiple n-gram models. They can be either used to represent different genres or tasks found in diverse text sources, or capture stochastic properties of different linguistic symbol sequences, for example, syllables and words. Unsupervised LM adaptation may also be used to further improve robustness to varying styles or tasks. When using these techniques, extensive software changes are often required. In this paper an alternative and more general approach based on weighted finite state transducers (WFSTs) is investigated for LM combination and adaptation. As it is entirely based on well-defined WFST operations, minimum change to decoding tools is needed. A wide range of LM combination configurations can be flexibly supported. An efficient on-the-fly WFST decoding algorithm is also proposed. Significant error rate gains of 7.3% relative were obtained on a state-of-the-art broadcast audio recognition task using a history dependently adapted multi-level LM modelling both syllable and word sequences. ©2010 IEEE.

Dasher - A data entry interface using continuous gestures and language models

Existing devices for communicating information to computers are bulky, slow to use, or unreliable. Dasher is a new interface incorporating language modelling and driven by continuous two-dimensional gestures, e.g. a mouse, touchscreen, or eye-tracker. Tests have shown that this device can be used to enter text at a rate of up to 34 words per minute, compared with typical ten-finger keyboard typing of 40-60 words per minute. Although the interface is slower than a conventional keyboard, it is small and simple, and could be used on personal data assistants and by motion-impaired computer users.

Language model cross adaptation for LVCSR system combination

State-of-the-art large vocabulary continuous speech recognition (LVCSR) systems often combine outputs from multiple subsystems developed at different sites. Cross system adaptation can be used as an alternative to direct hypothesis level combination schemes such as ROVER. In normal cross adaptation it is assumed that useful diversity among systems exists only at acoustic level. However, complimentary features among complex LVCSR systems also manifest themselves in other layers of modelling hierarchy, e.g., subword and word level. It is thus interesting to also cross adapt language models (LM) to capture them. In this paper cross adaptation of multi-level LMs modelling both syllable and word sequences was investigated to improve LVCSR system combination. Significant error rate gains up to 6.7% rel. were obtained over ROVER and acoustic model only cross adaptation when combining 13 Chinese LVCSR subsystems used in the 2010 DARPA GALE evaluation. © 2010 ISCA.

Unified circuit model for the polysilicon TFT

This paper proposes two methods to improve the modelling of thin film transistors (TFTs). The first involves integrating Poissons equation numerically, given a density of trap states and other relevant material parameters including a constant mobility. Theresult is conductance as a numerical function of gate voltage. The second method recognizes that the data for areal conductance found by numerical integration, may easily be found by measurement without making assumptions about the density of trap states.

A unified circuit model for the polysilicon thin film transistor

This paper describes a unified approach to modelling the polysilicon thin film transistor (TFT) for the purposes of circuit design. The approach uses accurate methods of predicting the channel conductance and then fitting the resulting data with a polynomial. Two methods are proposed to find the channel conductance: a device model and measurement. The approach is suitable because the TFT does not have a well defined threshold voltage. The polynomial conductance is then integrated generally to find the drain current and channel charge, necessary for a complete circuit model. © 1991 The Japan Society of Applied Physics.

Contact and channel modelling using part and function libraries in a function-based design approach

To support the development and analysis of engineering designs at the embodiment stage, designers work iteratively with representations of those designs as they consider the function and form of their constituent parts. Detailed descriptions of "what a machine does" usually include flows of forces and active principles within the technical system, and their localization within parts and across the interfaces between them. This means that a representation should assist a designer in considering form and function at the same time and at different levels of abstraction. This paper describes a design modelling approach that enables designers to break down a system architecture into its subsystems and parts, while assigning functions and flows to parts and the interfaces between them. In turn, this may reveal further requirements to fulfil functions in order to complete the design. The approach is implemented in a software tool which provides a uniform, computable language allowing the user to describe functions and flows as they are iteratively discovered, created and embodied. A database of parts allows the user to search for existing design solutions. The approach is illustrated through an example: modelling the complex mechanisms within a humanoid robot. Copyright © 2010 by ASME.

Language model cross adaptation for LVCSR system combination

State-of-the-art large vocabulary continuous speech recognition (LVCSR) systems often combine outputs from multiple sub-systems that may even be developed at different sites. Cross system adaptation, in which model adaptation is performed using the outputs from another sub-system, can be used as an alternative to hypothesis level combination schemes such as ROVER. Normally cross adaptation is only performed on the acoustic models. However, there are many other levels in LVCSR systems' modelling hierarchy where complimentary features may be exploited, for example, the sub-word and the word level, to further improve cross adaptation based system combination. It is thus interesting to also cross adapt language models (LMs) to capture these additional useful features. In this paper cross adaptation is applied to three forms of language models, a multi-level LM that models both syllable and word sequences, a word level neural network LM, and the linear combination of the two. Significant error rate reductions of 4.0-7.1% relative were obtained over ROVER and acoustic model only cross adaptation when combining a range of Chinese LVCSR sub-systems used in the 2010 and 2011 DARPA GALE evaluations. © 2012 Elsevier Ltd. All rights reserved.

Paraphrastic language models

In natural languages multiple word sequences can represent the same underlying meaning. Only modelling the observed surface word sequence can result in poor context coverage, for example, when using n-gram language models (LM). To handle this issue, this paper presents a novel form of language model, the paraphrastic LM. A phrase level transduction model that is statistically learned from standard text data is used to generate paraphrase variants. LM probabilities are then estimated by maximizing their marginal probability. Significant error rate reductions of 0.5%-0.6% absolute were obtained on a state-ofthe-art conversational telephone speech recognition task using a paraphrastic multi-level LM modelling both word and phrase sequences.