This demo demonstrates a small and portable system embedding microphone array processing and robu... more This demo demonstrates a small and portable system embedding microphone array processing and robust speech recognition for distant-speech interaction to control a floor lamp. The system is entirely contained inside the lamp and operates in real-time, "always-listening" mode. It runs on a small, low-power, fanless board and acts as the light control interface. The prototype shows the feasibility, potential and limits of the integration of speech technology in devices of everyday use.
This work proposes the use of hierarchical LMs as an effective method both for efficiently dealin... more This work proposes the use of hierarchical LMs as an effective method both for efficiently dealing with contextdependent LMs in a dialogue system and for increasing the robustness of LM estimation and adaptation. Starting from basic LMs that express elementary semantic units, concepts, or data-types, sentence level LMs are recursively built. The resulting LMs may be a combination of grammars, word classes, and statistical LMs. Moreover, these LMs can be efficiently compiled into probabilistic recursive transition networks. A speech decoding algorithm directly exploits the recursive representation and produces the most probable parse tree matching the speech signal. The proposed approach has been implemented for a dataentry task which covers structured data, e.g. numbers, dates, and proper names, as well as free texts. In this task, the active LM must continuously change according to the current status, the active form, and the data entered so far. Finally, while the hierarchical approach results very convenient to cope with this task, it also looks very general and can give advantages in other applications, e.g. dictation.
Radiological reporting has already been identified as a field in which voice technologies can pro... more Radiological reporting has already been identified as a field in which voice technologies can prove to be very useful. Recent progress in automatic speech recognition and in hardware and software technology makes it possible to build large-vocabulary, continuous speech, speaker-independent, real-time systems.
In this paper we present a new application for speech technology. In the SpeeData project a user ... more In this paper we present a new application for speech technology. In the SpeeData project a user of the system will speak in any of the provided languages, which presently are the two languages German and Italian. The system will analyse the utterance and generate a data base entry. At the same time, relevant information will be translated into the other language. This paper presents an overview of multilingual speech recognition when using a single recognition module.
In the last few years the computational power of general-purpose machines has dramatically increa... more In the last few years the computational power of general-purpose machines has dramatically increased; furthermore high performance workstations quite often o er multimedial capabilities. Therefore it is now possible to implement automatic Speech Recognizers (SRs) without dedicated hardware devices such as DSP boards. In this paper performance and costs for di erent con gurations of some SRs based on workstations, DSP boards, and mixed architectures (workstations equipped with DSP boards), developed at IRST, are presented and discussed. Although the response time depends both on the speech modalities, ( isolated words or continuous speech ), and the hardware platform on which the SR runs, nevertheless, for isolated words, a 486 machine can give very interesting results, above all considering the overall system price. The conclusions are that the use of Digital Signal Processor boards (DSP) as computational engines for Automatic Speech Recognition (ASR) systems will probably be no longer convenient, whilst they can still be used to implement sophisticated recognition systems or single chip SRs.
This paper reports on experiments of porting the ITC-irst Italian broadcast news recognition syst... more This paper reports on experiments of porting the ITC-irst Italian broadcast news recognition system to two spontaneous dialogue domains. The trade-off between performance and the required amount of task specific data was investigated. Porting was experimented by applying supervised adaptation methods on acoustic and language models. By using two hours of manually transcribed speech, word error rates of 26.0% and 28.4% were achieved by the adapted systems. Two reference systems, developed on a larger training corpus, achieved word error rates of 22.6% and 21.2%, respectively.
In the last few years the computational power of general-purpose machines has dramatically increa... more In the last few years the computational power of general-purpose machines has dramatically increased; furthermore high performance workstations quite often offer multimedial capabilities. Therefore it is now possible to implement automatic Speech Recognizers (SRs) without dedicated hardware devices such as DSP boards. In this paper performance and costs for different configurations of some SRs based on workstations and mixed architectures (workstations equipped with DSP boards), developed at IRST, are presented and discussed. Although the response time depends both on the speech modalities, ( isolated words or continuous speech ), and the hardware platform on which the SR runs, nevertheless, for isolated words, considering the overall system price, a 486 machine can give very interesting results. 1 Introduction Integer and floating point performance of new machines based on the Hp PA-RISC and the DEC Alpha chips can be compared to mainframe computer capabilities. Furthermore, high p...
This paper describes an efficient way of representing a bigram language model with a finite state... more This paper describes an efficient way of representing a bigram language model with a finite state network used by a beam-search based and continuous speech HMM recognizer. In a previous paper [1], a compact tree-based organization of the search space was presented, that could be further reduced through an optimization algorithm. There, it was pointed out that for a 10,000-word newspaper dictation task the minimization step could have taken a lot of time and space on a standard workstation. In this paper, a new compilation technique that takes into account the particular tree-based topology is described. Results show that without additional time and space costs, the new technique produces networks equivalent to the tree-based ones but almost as small as the optimized one. 1 INTRODUCTION The most widely used Language Models (LMs) in speech recognition are n-gram models, due to both easy inference from the training corpus and easy integrability with the decoding algorithms commonly use...
In a large vocabulary isolated word automatic speech recognition system, using Hidden Markov Mode... more In a large vocabulary isolated word automatic speech recognition system, using Hidden Markov Models (HMM), it is time consuming to perform a detailed likelihood computation of an input utterance for all the words of the dictionary. A tree based A search algorithm is presented for finding the N best word hypotheses, using a new likelihood estimator to drive time asyncronously the search. The proposed likelihood estimator has been tested with two different tasks: a large vocabulary (5000 words) multi-speaker radiological reports recognizer, and a 94 word speaker-independent command language recognizer. In both cases, the speed performance was improved by a factor of about 6, without losing admissibility. Using discrete models, the mean word hypothesis evaluation time is 8 ms on Sun4/330 workstation, the recognition rates vary from 94% to 98% depending on the speaker in the 5000 word task with a bigram language model, and 91% in the speaker-independent task. Introduction A system base...
This paper presents the development of an Italian broadcast news transcription system, to be appl... more This paper presents the development of an Italian broadcast news transcription system, to be applied for the indexing of multimedia archives. Moreover, a broadcast news corpus under collection at ITC-irst is introduced. The system processes the input audio stream in four stages. The first one performs audio segmentation via the Bayesian Information Criterion (BIC) and classification by Gaussians mixtures modeling. The second stage groups spectrally homogeneous speech segments, again using the BIC method, in order to provide speaker clusters suitable for the following adaptation module. The third stage adapts the acoustic models to each selected cluster and, finally, the fourth stage transcribes the audio data employing cluster adapted models. The achieved word error rate, measured on a 1h:15m test set, corresponding to 6 news programs, was 21.5%.
This paper reports on several experiments performed during the development of the ITC-irst transc... more This paper reports on several experiments performed during the development of the ITC-irst transcription system for the TC-STAR '06 evaluation campaign. The aim is to nd methods of exploting a set of alternative hypotheses produced by different systems to derive a transcription that is more acccurate than any of these. The most used technique for combining alternative hypotheses relies on
International Conference on Acoustics, Speech, and Signal Processing, 1994
This paper compares different ways of estimating bigram language models and of representing them ... more This paper compares different ways of estimating bigram language models and of representing them in a finite state network used by a beam-search based, continuous speech, and speaker independent HMM recognizer. Attention is focused on the n-gram interpolation scheme for which seven models are considered. Among them, the Stacked estimated linear interpolated model favourably compares with the best known ones.
This paper considers the problems of estimating bigram language models and of efficiently represe... more This paper considers the problems of estimating bigram language models and of efficiently representing them by a finite state network, which can be employed by an hidden Markov model based, beam-search, continuous speech recognizer.
[Proceedings] ICASSP-92: 1992 IEEE International Conference on Acoustics, Speech, and Signal Processing, 1992
Stochastic signal models represent a powerful tool for automatic speech recognition. A particular... more Stochastic signal models represent a powerful tool for automatic speech recognition. A particular type of stochastic modeling based on first-order hidden Markov models (HMMs), has been increasingly popular, because it has a solid theoretical basis and offers practical advantages. The authors extend the standard HMM theory to parallel hidden Markov models (PHMMs). The parallel model consists of two statistically related
This demo demonstrates a small and portable system embedding microphone array processing and robu... more This demo demonstrates a small and portable system embedding microphone array processing and robust speech recognition for distant-speech interaction to control a floor lamp. The system is entirely contained inside the lamp and operates in real-time, "always-listening" mode. It runs on a small, low-power, fanless board and acts as the light control interface. The prototype shows the feasibility, potential and limits of the integration of speech technology in devices of everyday use.
2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100), 2000
This paper presents the first achievements in the development of a broadcast news transcription s... more This paper presents the first achievements in the development of a broadcast news transcription system to be applied for the processing of huge audio archives. In particular, the Italian broadcast news corpus under collection is introduced, and the first implemented baseline system is outlined. The baseline system consists of an audio segmentation module and a speech recognizer featuring a recursive Viterbi beam search, a 64K-word lexicon, a treebased trigram LM representation, and MLLR adaptation. The word error rate of the baseline was 20.9% on planned studio speech and 28.8% on the whole test set.
This demo demonstrates a small and portable system embedding microphone array processing and robu... more This demo demonstrates a small and portable system embedding microphone array processing and robust speech recognition for distant-speech interaction to control a floor lamp. The system is entirely contained inside the lamp and operates in real-time, "always-listening" mode. It runs on a small, low-power, fanless board and acts as the light control interface. The prototype shows the feasibility, potential and limits of the integration of speech technology in devices of everyday use.
This work proposes the use of hierarchical LMs as an effective method both for efficiently dealin... more This work proposes the use of hierarchical LMs as an effective method both for efficiently dealing with contextdependent LMs in a dialogue system and for increasing the robustness of LM estimation and adaptation. Starting from basic LMs that express elementary semantic units, concepts, or data-types, sentence level LMs are recursively built. The resulting LMs may be a combination of grammars, word classes, and statistical LMs. Moreover, these LMs can be efficiently compiled into probabilistic recursive transition networks. A speech decoding algorithm directly exploits the recursive representation and produces the most probable parse tree matching the speech signal. The proposed approach has been implemented for a dataentry task which covers structured data, e.g. numbers, dates, and proper names, as well as free texts. In this task, the active LM must continuously change according to the current status, the active form, and the data entered so far. Finally, while the hierarchical approach results very convenient to cope with this task, it also looks very general and can give advantages in other applications, e.g. dictation.
Radiological reporting has already been identified as a field in which voice technologies can pro... more Radiological reporting has already been identified as a field in which voice technologies can prove to be very useful. Recent progress in automatic speech recognition and in hardware and software technology makes it possible to build large-vocabulary, continuous speech, speaker-independent, real-time systems.
In this paper we present a new application for speech technology. In the SpeeData project a user ... more In this paper we present a new application for speech technology. In the SpeeData project a user of the system will speak in any of the provided languages, which presently are the two languages German and Italian. The system will analyse the utterance and generate a data base entry. At the same time, relevant information will be translated into the other language. This paper presents an overview of multilingual speech recognition when using a single recognition module.
In the last few years the computational power of general-purpose machines has dramatically increa... more In the last few years the computational power of general-purpose machines has dramatically increased; furthermore high performance workstations quite often o er multimedial capabilities. Therefore it is now possible to implement automatic Speech Recognizers (SRs) without dedicated hardware devices such as DSP boards. In this paper performance and costs for di erent con gurations of some SRs based on workstations, DSP boards, and mixed architectures (workstations equipped with DSP boards), developed at IRST, are presented and discussed. Although the response time depends both on the speech modalities, ( isolated words or continuous speech ), and the hardware platform on which the SR runs, nevertheless, for isolated words, a 486 machine can give very interesting results, above all considering the overall system price. The conclusions are that the use of Digital Signal Processor boards (DSP) as computational engines for Automatic Speech Recognition (ASR) systems will probably be no longer convenient, whilst they can still be used to implement sophisticated recognition systems or single chip SRs.
This paper reports on experiments of porting the ITC-irst Italian broadcast news recognition syst... more This paper reports on experiments of porting the ITC-irst Italian broadcast news recognition system to two spontaneous dialogue domains. The trade-off between performance and the required amount of task specific data was investigated. Porting was experimented by applying supervised adaptation methods on acoustic and language models. By using two hours of manually transcribed speech, word error rates of 26.0% and 28.4% were achieved by the adapted systems. Two reference systems, developed on a larger training corpus, achieved word error rates of 22.6% and 21.2%, respectively.
In the last few years the computational power of general-purpose machines has dramatically increa... more In the last few years the computational power of general-purpose machines has dramatically increased; furthermore high performance workstations quite often offer multimedial capabilities. Therefore it is now possible to implement automatic Speech Recognizers (SRs) without dedicated hardware devices such as DSP boards. In this paper performance and costs for different configurations of some SRs based on workstations and mixed architectures (workstations equipped with DSP boards), developed at IRST, are presented and discussed. Although the response time depends both on the speech modalities, ( isolated words or continuous speech ), and the hardware platform on which the SR runs, nevertheless, for isolated words, considering the overall system price, a 486 machine can give very interesting results. 1 Introduction Integer and floating point performance of new machines based on the Hp PA-RISC and the DEC Alpha chips can be compared to mainframe computer capabilities. Furthermore, high p...
This paper describes an efficient way of representing a bigram language model with a finite state... more This paper describes an efficient way of representing a bigram language model with a finite state network used by a beam-search based and continuous speech HMM recognizer. In a previous paper [1], a compact tree-based organization of the search space was presented, that could be further reduced through an optimization algorithm. There, it was pointed out that for a 10,000-word newspaper dictation task the minimization step could have taken a lot of time and space on a standard workstation. In this paper, a new compilation technique that takes into account the particular tree-based topology is described. Results show that without additional time and space costs, the new technique produces networks equivalent to the tree-based ones but almost as small as the optimized one. 1 INTRODUCTION The most widely used Language Models (LMs) in speech recognition are n-gram models, due to both easy inference from the training corpus and easy integrability with the decoding algorithms commonly use...
In a large vocabulary isolated word automatic speech recognition system, using Hidden Markov Mode... more In a large vocabulary isolated word automatic speech recognition system, using Hidden Markov Models (HMM), it is time consuming to perform a detailed likelihood computation of an input utterance for all the words of the dictionary. A tree based A search algorithm is presented for finding the N best word hypotheses, using a new likelihood estimator to drive time asyncronously the search. The proposed likelihood estimator has been tested with two different tasks: a large vocabulary (5000 words) multi-speaker radiological reports recognizer, and a 94 word speaker-independent command language recognizer. In both cases, the speed performance was improved by a factor of about 6, without losing admissibility. Using discrete models, the mean word hypothesis evaluation time is 8 ms on Sun4/330 workstation, the recognition rates vary from 94% to 98% depending on the speaker in the 5000 word task with a bigram language model, and 91% in the speaker-independent task. Introduction A system base...
This paper presents the development of an Italian broadcast news transcription system, to be appl... more This paper presents the development of an Italian broadcast news transcription system, to be applied for the indexing of multimedia archives. Moreover, a broadcast news corpus under collection at ITC-irst is introduced. The system processes the input audio stream in four stages. The first one performs audio segmentation via the Bayesian Information Criterion (BIC) and classification by Gaussians mixtures modeling. The second stage groups spectrally homogeneous speech segments, again using the BIC method, in order to provide speaker clusters suitable for the following adaptation module. The third stage adapts the acoustic models to each selected cluster and, finally, the fourth stage transcribes the audio data employing cluster adapted models. The achieved word error rate, measured on a 1h:15m test set, corresponding to 6 news programs, was 21.5%.
This paper reports on several experiments performed during the development of the ITC-irst transc... more This paper reports on several experiments performed during the development of the ITC-irst transcription system for the TC-STAR '06 evaluation campaign. The aim is to nd methods of exploting a set of alternative hypotheses produced by different systems to derive a transcription that is more acccurate than any of these. The most used technique for combining alternative hypotheses relies on
International Conference on Acoustics, Speech, and Signal Processing, 1994
This paper compares different ways of estimating bigram language models and of representing them ... more This paper compares different ways of estimating bigram language models and of representing them in a finite state network used by a beam-search based, continuous speech, and speaker independent HMM recognizer. Attention is focused on the n-gram interpolation scheme for which seven models are considered. Among them, the Stacked estimated linear interpolated model favourably compares with the best known ones.
This paper considers the problems of estimating bigram language models and of efficiently represe... more This paper considers the problems of estimating bigram language models and of efficiently representing them by a finite state network, which can be employed by an hidden Markov model based, beam-search, continuous speech recognizer.
[Proceedings] ICASSP-92: 1992 IEEE International Conference on Acoustics, Speech, and Signal Processing, 1992
Stochastic signal models represent a powerful tool for automatic speech recognition. A particular... more Stochastic signal models represent a powerful tool for automatic speech recognition. A particular type of stochastic modeling based on first-order hidden Markov models (HMMs), has been increasingly popular, because it has a solid theoretical basis and offers practical advantages. The authors extend the standard HMM theory to parallel hidden Markov models (PHMMs). The parallel model consists of two statistically related
This demo demonstrates a small and portable system embedding microphone array processing and robu... more This demo demonstrates a small and portable system embedding microphone array processing and robust speech recognition for distant-speech interaction to control a floor lamp. The system is entirely contained inside the lamp and operates in real-time, "always-listening" mode. It runs on a small, low-power, fanless board and acts as the light control interface. The prototype shows the feasibility, potential and limits of the integration of speech technology in devices of everyday use.
2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100), 2000
This paper presents the first achievements in the development of a broadcast news transcription s... more This paper presents the first achievements in the development of a broadcast news transcription system to be applied for the processing of huge audio archives. In particular, the Italian broadcast news corpus under collection is introduced, and the first implemented baseline system is outlined. The baseline system consists of an audio segmentation module and a speech recognizer featuring a recursive Viterbi beam search, a 64K-word lexicon, a treebased trigram LM representation, and MLLR adaptation. The word error rate of the baseline was 20.9% on planned studio speech and 28.8% on the whole test set.
Uploads
Papers by Fabio Brugnara