Real-Time Robust Automatic Speech Recognition Using Compact Support Vector Machines
| Autor: | Ana I. García-Moral, Carmen Peláez-Moreno, Fernando Diaz-de-Maria, Manel Martínez-Ramón, Rubén Solera-Ureña |
|---|---|
| Rok vydání: | 2012 |
| Předmět: |
Acoustics and Ultrasonics
Computer science Speech recognition Speech coding Context (language use) Machine learning computer.software_genre Robust ASR Hybrid ASR Electrical and Electronic Engineering Hidden Markov model Real-time ASR Telecomunicaciones Support vector machines Artificial neural networks Artificial neural network business.industry Compact SVM Speech processing Hidden Markov models (HMM) Support vector machine SVM/HMM ComputingMethodologies_PATTERNRECOGNITION ANN/HMM Computer Science::Sound Additive noise Artificial intelligence Noise (video) business computer Decoding methods |
| Zdroj: | e-Archivo. Repositorio Institucional de la Universidad Carlos III de Madrid instname |
| ISSN: | 1558-7924 1558-7916 2008-0638 |
| DOI: | 10.1109/tasl.2011.2178597 |
| Popis: | In the last years, support vector machines (SVMs) have shown excellent performance in many applications, especially in the presence of noise. In particular, SVMs offer several advantages over artificial neural networks (ANNs) that have attracted the attention of the speech processing community. Nevertheless, their high computational requirements prevent them from being used in practice in automatic speech recognition (ASR), where ANNs have proven to be successful. The high complexity of SVMs in this context arises from the use of huge speech training databases with millions of samples and highly overlapped classes. This paper suggests the use of a weighted least squares (WLS) training procedure that facilitates the possibility of imposing a compact semiparametric model on the SVM, which results in a dramatic complexity reduction. Such a complexity reduction with respect to conventional SVMs, which is between two and three orders of magnitude, allows the proposed hybrid WLS-SVC/HMM system to perform real-time speech decoding on a connected-digit recognition task (SpeechDat Spanish database). The experimental evaluation of the proposed system shows encouraging performance levels in clean and noisy conditions, although further improvements are required to reach the maturity level of current context-dependent HMM based recognizers. Spanish Ministry of Science and Innovation TEC 2008-06382 and TEC 2008-02473 and Comunidad Autónoma de Madrid-UC3M CCG10-UC3M/TIC-5304. Publicado |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|