In this paper we report our development work in Spanish spontaneous speech conversational systems... more In this paper we report our development work in Spanish spontaneous speech conversational systems. We describe the automatic telephone operator service (ATOS) and present the improvements introduced into it to deal with spontaneous speech, which are: (a) a task independent dialogue manager, that can be adapted to a new semantic domain by changing a configuration file. It also generates a prediction about the user's expected utterance to constrain the language model used by the speech recognizer; (b) a language modeling strategy, which allows to adapt the statistical language model to a new task with just few hundreds of sentences. This strategy reduces a 27% the word error rate. We also report the results, conclusions and the speech database collected in the evaluation of the ATOS system, which has been tested by 30 real users
We present a new speech rate classifier (SRC) which is directly based on the dynamic coefficients... more We present a new speech rate classifier (SRC) which is directly based on the dynamic coefficients of the feature vectors and it is suitable to be used in real time. We also report the study that has been carried out to determine what parameters of speech are the best regarding the speech rate classification problem. In this study we analyse
... Although it may be thought of as an `answering machine in the network' it offers much mo... more ... Although it may be thought of as an `answering machine in the network' it offers much more. ... The technologies involved in this project are automatic speech recognition, dialog management and text to speech conversion. 2.5.2. Technologies and platforms. ...
Statistical language models provide a powerful tool for modelling natural spoken language. Nevert... more Statistical language models provide a powerful tool for modelling natural spoken language. Nevertheless a large set of training sentences is required to estimate reliably the model parameters. The authors present a method for estimating n-gram probabilities from sparse data. The proposed language modeling strategy allows one to adapt a generic language model (LM) to a new semantic domain with just
... (b) The second compensation technique, Transition Probabilities Adaptation (TPA), modifies th... more ... (b) The second compensation technique, Transition Probabilities Adaptation (TPA), modifies the HMM state-transition probabilities to adapt them to fast and slow speech. This idea has been previously tried by other authors [3][4] and our experiments confirm its usefulness. ...
In this paper we report our development work in Spanish spontaneous speech conversational systems... more In this paper we report our development work in Spanish spontaneous speech conversational systems. We describe the automatic telephone operator service (ATOS) and present the improvements introduced into it to deal with spontaneous speech, which are: (a) a task independent dialogue manager, that can be adapted to a new semantic domain by changing a configuration file. It also generates a prediction about the user's expected utterance to constrain the language model used by the speech recognizer; (b) a language modeling strategy, which allows to adapt the statistical language model to a new task with just few hundreds of sentences. This strategy reduces a 27% the word error rate. We also report the results, conclusions and the speech database collected in the evaluation of the ATOS system, which has been tested by 30 real users
We present a new speech rate classifier (SRC) which is directly based on the dynamic coefficients... more We present a new speech rate classifier (SRC) which is directly based on the dynamic coefficients of the feature vectors and it is suitable to be used in real time. We also report the study that has been carried out to determine what parameters of speech are the best regarding the speech rate classification problem. In this study we analyse
... Although it may be thought of as an `answering machine in the network' it offers much mo... more ... Although it may be thought of as an `answering machine in the network' it offers much more. ... The technologies involved in this project are automatic speech recognition, dialog management and text to speech conversion. 2.5.2. Technologies and platforms. ...
Statistical language models provide a powerful tool for modelling natural spoken language. Nevert... more Statistical language models provide a powerful tool for modelling natural spoken language. Nevertheless a large set of training sentences is required to estimate reliably the model parameters. The authors present a method for estimating n-gram probabilities from sparse data. The proposed language modeling strategy allows one to adapt a generic language model (LM) to a new semantic domain with just
... (b) The second compensation technique, Transition Probabilities Adaptation (TPA), modifies th... more ... (b) The second compensation technique, Transition Probabilities Adaptation (TPA), modifies the HMM state-transition probabilities to adapt them to fast and slow speech. This idea has been previously tried by other authors [3][4] and our experiments confirm its usefulness. ...
Uploads
Papers by Daniel Tapias