Automatic recognition of non-native speech is problematic. A key challenge in developing spoken CALL systems is to design exercises that enable learning but which are still technically feasible. This especially applies to systems intended for practicing grammar. In the current paper we focus on the issue of matching design and speech technology. On the one hand we are developing and testing spe...

We describe a robust speech understanding system based on our newly developed approach to spoken language processing. We show that a robust NLU system can be rapidly developed using a relatively simple speech recognizer to provide sufficient information for database retrieval by spoken language. Our experimental system consists of three components: a speech recognizer based on HMM, a natural la...

It has been relatively difficult to develop natural language parsers for spoken dialog systems, not only because of the possible recognition errors, pauses, hesitations, out-ofvocabulary words, and the grammatically incorrect sentence structures, but because of the great efforts required to develop a general enough grammar with satisfactory coverage and flexibility to handle different applicati...

A stochastically based approach for the semantic analysis component of a natural spoken language system for the ATIS task has been developed. The semantic analyzer of the spoken language system already in use at LIMSI makes use of a rule-based case grammar. In this work, the system of rules for the semantic analysis is replaced with a relatively simple, first order Hidden Markov Model. The perf...

This paper reports on recent developments for the creation and analysis of very large databases of emotional and attitudinallymarked speech for the support of research into concatenative methods for producing synthesised speech which is capable of expressing the range of prosody and phonation styles to emulate human spoken interactions. It addresses the problems of ensuring high spontaneity in ...

We present in this paper a parser relying on a constraint-based formalism called Property Grammar. We show how constraints constitute an efficient solution in parsing non canonical material such as spoken language transcription or e-mails. This technique, provided that it is implemented with some control mechanisms, is very efficient. Some results are presented, from the French parsing evaluati...

The most common speech understanding architecture for spoken dialogue systems is a combination of speech recognition based on a class N-gram language model, and robust parsing. For many types of applications, however, grammar-based recognition can offer concrete advantages. Training a good class N-gram language model requires substantial quantities of corpus data, which is generally not availab...

The meaning of an idiomatic expression cannot be transparently worked out from the meanings of its constituent words due to its figurative and unpredictable nature. Consequently, the syntactic composition and the structural paradigm of an idiomatic expression are supposed to be the same in every context. However, this is not the case in the institutionalized second language varieties of English...

We use machine learners trained on a combination of acoustic confidence and pragmatic plausibility features computed from dialogue context to predict the accuracy of incoming n-best recognition hypotheses to a spoken dialogue system. Our best results show a 25% weighted f-score improvement over a baseline system that implements a “grammar-switching” approach to context-sensitive speech recognit...

This paper describes the Grammar Programming Language (GPL), a new formalism for feature-structure-based linguistic computation. GPL was designed to meet the needs of spoken language translation. GPL is easy to use, concise, and efficient, and it allows the direct expression of detailed linguistic algorithms. GPL was used successfully as the basis of Sony's machine translation project.