Cross-Lingual spoken language understanding from unaligned data using discriminative classification models and machine translation

This paper investigates several approaches to bootstrapping a new spoken language understanding (SLU) component in a target language given a large dataset of semantically-annotated utterances in some other source language. The aim is to reduce the cost associated with porting a spoken dialogue system from one language to another by minimising the amount of data required in the target language. Since word-level semantic annotations are costly, Semantic Tuple Classifiers (STCs) are used in conjunction with statistical machine translation models both of which are trained from unaligned data to further reduce development time. The paper presents experiments in which a French SLU component in the tourist information domain is bootstrapped from English data. Results show that training STCs on automatically translated data produced the best performance for predicting the utterance's dialogue act type, however individual slot/value pairs are best predicted by training STCs on the source language and using them to decode translated utterances. © 2010 ISCA.

Exploring the impacts of the interactions between lifecycles and other dynamics that influence the development of technology-based industries

To address future uncertainty within strategy and innovation, managers extrapolate past patterns and trends into the future. Several disciplines make use of lifecycles, often with a linear sequence of identified phases, to make predictions and address likely uncertainties. Often the aggregation of several cycles is then interpreted as a new cycle - such as product lifecycles into an industry lifecycle. However, frequently different lifecycle terms - technology, product, industry - are used interchangeably and without clear definition. Within the interdisciplinary context of technology management, this juxtaposition of dynamics can create confusion, rather than clarification. This paper explores some typical dynamics associated with technology-based industries, using illustrative examples from the automotive industry. A wide range of dimensions are seen to influence the path of a technology-based industry, and stakeholders need to consider the likely causality and synchronicity of these. Some curves can simply present the aggregation of components; other dynamics incur time lags, rather than being superimposed, but still have a significant impact. To optimise alignment of the important dimensions within any development, and for future strategy decisions, understanding these interactions will be critical. © 2011 IEEE.

Customizing the technology roadmapping technique for software companies

Software importance keeps growing fast and consistently for many organizations. The growth of software functionality in manufactured products and the emergence of digital media, convergent spaces including digital content, software, and multi-channels to the market, are recent examples of organizational changes where software assumed a central position for the corporate strategy. This paper analyzes the alignment between strategic objectives and software development processes at software companies and proposes a methodology to ensure that development processes are aligned with the corporate capabilities required to exploit future market opportunities. The methodology includes the categorization of different software companies according to their core capabilities and the customization of the technology roadmapping technique for software companies. The research process included the realization of case studies and a survey. (c) 2006 PICMET.

An ngram-based statistical machine translation decoder

In this paper we describe MARIE, an Ngram-based statistical machine translation decoder. It is implemented using a beam search strategy, with distortion (or reordering) capabilities. The underlying translation model is based on an Ngram approach, extended to introduce reordering at the phrase level. The search graph structure is designed to perform very accurate comparisons, what allows for a high level of pruning, improving the decoder efficiency. We report several techniques for efficiently prune out the search space. The combinatory explosion of the search space derived from the search graph structure is reduced by limiting the number of reorderings a given translation is allowed to perform, and also the maximum distance a word (or a phrase) is allowed to be reordered. We finally report translation accuracy results on three different translation tasks.