Volker Warnke

Integrated recognition of words and prosodic phrase boundaries

Abstract In this paper, we present an integrated approach for recognizing both the word sequence and the syntactic–prosodic structure of a spontaneous utterance. The approach aims at improving the performance of the understanding component of speech understanding systems by exploiting not only acoustic–phonetic and syntactic information, but also prosodic information directly within the speech recognition process. Whereas spoken utterances are typically modelled as unstructured word sequences in the speech recognizer, our approach includes phrase boundary information in the language model and provides HMMs to model the acoustic and prosodic characteristics of phrase boundaries. This methodology has two major advantages compared to purely word-based speech recognizers. First, additional syntactic–prosodic boundaries are determined by the speech recognizer which facilitates parsing and resolve syntactic and semantic ambiguities. Second – after having removed the boundary information from the result of the recognizer – the integrated model yields a 4% relative word error rate (WER) reduction compared to a traditional word recognizer. The boundary classification performance is equal to that of a separate prosodic classifier operating on the word recognizer output, thus making a separate classifier unnecessary for this task and saving the computation time involved. Compared to the baseline word recognizer, the integrated word-and-boundary recognizer does not involve any computational overhead.

Fast and Robust Features for Prosodic Classification

In our previous research, we have shown that prosody can be used to dramatically improve the performance of the automatic speech translation system Verbmobil [5,7,8]. In Verbmobil, prosodic information is made available to the different modules of the system by annotating the output of a word recognizer with prosodic markers. These markers are determined in a classification process. The computation of the prosodic features used for classification was previously based on a time alignment of the phoneme sequence of the recognized words. The phoneme segmentation was needed for the normalization of duration and energy features. This time alignment was very expensive in terms of computational effort and memory requirement. In our new approach the normalization is done on the word level with precomputed duration and energy statistics, thus the phoneme segmentation can be avoided. With the new set of prosodic features better classification results can be achieved, the features extraction can be sped up by 64 %, and the memory requirements are even reduced by 92%.

Discriminative estimation of interpolation parameters for language model classifiers

In this paper we present a new approach for estimating the interpolation parameters of language models (LM) which are used as classifiers. With the classical maximum likelihood (ML) estimation theoretically one needs to have a huge amount of data and the fundamental density assumption has to be correct. Usually one of these conditions is violated, so different optimization techniques like maximum mutual information (MMI) and minimum classification error (MCE) can be used instead, where the interpolation parameters are not optimized on their own but in consideration of all models together. In this paper we present how MCE and MMI techniques can be applied to two different kind of interpolation strategies: the linear interpolation, which is the standard interpolation method and the rational interpolation. We compare ML, MCE and MMI on the German part of the Verbmobil corpus, where we get a reduction of 3% of classification error when discriminating between 18 dialog act classes.

Topic spotting using subword units

In this paper we present a new approach for topic spotting based on subword units and feature vectors instead of words. In our first approach, we only use vector quantized feature vectors and polygram language models for topic representation. In the second approach, we use phonemes instead of the vector quantized feature vectors and model the topics again using polygram language models. We trained and tested the two methods on two different corpora. The first is a part of a media corpus which contains data from TV shows for three different topics. The second is the VERBMOBIL-corpus where we used 18 dialog acts as topics. Each corpus was splitted into disjunctive test and training sets. We achieved recognition rates up to 82% for the three topics of the media corpus and up to 64% using 18 dialog acts of the VERBMOBIL-corpus as topics.

Telephone-based speech dialog systems

In this contribution we look back on the last years in the history of telephone-based speech dialog systems. We will start in 1993 when the world wide first natural language understanding dialog system using a mixed-initiative approach was made accessible for the public, the well-known EVAR system from the Chair for Pattern Recognition of the University of Erlangen-Nuremberg. Then we discuss certain requirements we consider necessary for the successful application of dialog systems. Finally we present trends and developments in the area of telephone-based dialog systems.

Research Issues for the Next Generation Spoken Dialogue Systems Revisited

In this paper we take a second look at current research issues for conversational dialogue systems addressed in [17]. We look at two systems, a movie information and a stock information system which were built based on the experiences with the train information system Evar, described in [17].

Recognition and Labelling of Prosodic Events in Slovenian Speech

The paper describes prosodic annotation procedures of the GOPOLIS Slovenian speech data database and methods for automatic classification of different prosodic events. Several statistical parameters concerning duration and loudness of words, syllables and allophones were computed for the Slovenian language, for the first time on such a large amount of speech data. The evaluation of the annotated data showed a close match between automatically determined syntactic-prosodic boundary marker positions and those obtained by a rule-based approach.