The paper is concerning an approach for understanding speech using a new form of probabilistic models to represent syntactic and semantic knowledge of a restricted domain. One important feature of our grammar is that the parse tree directly represents the semantic content of the utterance. Since we determine that semantic content by an integrated search, we avoid consistency problems at the interface between the recognizer and the language understanding part of the speech understanding system. We succeeded in designing such an incremental algorithm, which integrates semantic, syntactic, and acoustic-phonetic knowledge in a seamless, consistent way. High efficiency is achieved by using a chart-parsing technique with structuresharing and a strict top-down strategy for opening new word hypotheses in the pronunciation layer.

We present an active chart parser which parses left connected wordgraphs in a strictly time synchronous way. The parser performs a beam search on the possible paths through the word graph and on the possible derivations of the unification grammar simultaneously. A metric is given to assign scores to edges, taking into account the whole left context thereby combining acoustic probabilities, n-gram probabilities and unification grammar probabilities. A specialized model for the derivation of typed unification grammars is introduced. Different ways of coupling the parser with an LR beam decoder in an online time synchronous fashion are defined and several experimental results are presented. Two top down and one bottom up method are investigated. In bottom up mode, the decoder sends word hypotheses as they are found from left to right, while the parser keeps step. In verify mode, the decoder is always a frame ahead, while the parser verifies received hypotheses, providing language informat...

. This article describes a probabilistic context free grammar approximation method for unification grammars. In order to produce good results, the method is combined with an N best parsing extension to chart parsing. The first part of the paper introduces the grammar approximation method, while the second part describes details of an efficient N-best packing and unpacking scheme for chart parsing. 1 Introduction Recently much attention has been payed to the integration of speech and language technology 1 . The concentration on spontaneous speech understanding led to the definition of a robust interface known as the word graph or word lattice between recognition and understanding. Depending on the application, systems are built to provide a shallow stochastic analysis or a deep linguistic analysis of the word lattice. Using a shallow stochastic approach, a rough template-based analysis can be achieved which makes sense in those cases where a fine grained reconstruction of meanings is...

This paper describes a domain-limited system for speech understanding as well as for speech translation. An integrated semantic decoder directly converts the preprocessed speech signal into its semantic representation by a maximum a-posteriori classification. With the combination of probabilistic knowledge on acoustic, phonetic, syntactic, and semantic levels, the semantic decoder extracts the most probable meaning of the utterance. Any separate speech recognition stage is not needed because of the integration of the Viterbi-algorithm (calculating acoustic probabilities by the use of Hidden-MarkovModels) and a probabilistic chart parser (calculating semantic and syntactic probabilities by especial models). The semantic structure is introduced as representation of an utterance's meaning. It can be used as intermediate level for a succeeding intention decoder (within a speech understanding system for the control of a running application by spoken inputs) as well as interlingua-level for...

Introduction L'utilisation de grammaires alg'ebriques pour la description de langages naturels a pris de l'int'eret du point de vue de la linguistique th'eorique pendant ces derni`eres ann'ees. Ces grammaires sont adapt'ees `a l'impl'ementation, et montrent qu'avec leur utilisation les syst`emes d'analyse de langages naturels peuvent incorporer des composantes `a la fois efficaces du point de vue du traitement et 'el'egantes du point de vue linguistique, `a l'oppos'e des approches heuristiques des syst`emes conventionnels. Nous ne sugg'erons pas qu'il existe une grammaire alg'ebrique pour d'ecrire chaque langage naturel; il est probablement plus indiqu'e de voir la grammaire comme une structure de controle guidant l'analyse du texte en entr'ee, l'analyse s'emantique post'erieure 'etant motiv'ee par un mod`ele linguistique plus sophistiqu'e. La section 2 d'ecrit le fonctionnement du module d"etiquetage du syst`eme. Le g'en'erateur d'analyseurs syntaxiques est pr'esent'e dans l

This work explores the problem of incremental analysis in the context of chart parsing, probably the most commonly used framework for the analysis of natural language. Incrementality here means that syntax correctness of a text is checked dynamically as the text is edited, changing the internal representation of the analysis rather than generating an entirely new one. This implies that the system then may interact with the user in order to resolve problems that occur. As a consequence, these kinds of techniques can be used to develop highly interactive and reactive natural language processors. The kernel of the work is an incremental parsing algorithm that analyses arbitrary changes of a text, allowing competing analyses to be developed in parallel.