Robust features in deep-learning-based speech recognition | ||
| dc.conicet.organizacion | SRI International | |
| dc.conicet.organizacion | SRI International | |
| dc.conicet.organizacion | University of Carnegie Mellon | |
| dc.conicet.organizacion | SRI International | |
| dc.conicet.organizacion | Consejo Nacional de Investigaciones Científicas y Técnicas. Oficina de Coordinación Administrativa Ciudad Universitaria. Instituto de Investigación en Ciencias de la Computación. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales. Instituto de Investigación en Ciencias de la Computación | |
| dc.conicet.organizacion | SRI International | |
| dc.conicet.organizacion | SRI International | |
| dc.conicet.organizacion | SRI International | |
| dc.conicet.organizacion | University of California at Los Angeles | |
| dc.conicet.organizacion | University of Texas | |
| dc.conicet.paginas | 436 | es |
| dc.contributor.author | Mitra, Vikramjit | |
| dc.contributor.author | Franco, Horacio | |
| dc.contributor.author | Stern, Richard M. | |
| dc.contributor.author | Van Hout, Julien | |
| dc.contributor.author | Ferrer, Luciana | |
| dc.contributor.author | Graciarena, Martin | |
| dc.contributor.author | Wang, Wen | |
| dc.contributor.author | Vergyri, Dimitra | |
| dc.contributor.author | Alwan, Abeer | |
| dc.contributor.author | Hansen, John H. L. | |
| dc.contributor.other | Watanabe, Shinji | |
| dc.contributor.other | Delcroix, Marc | |
| dc.contributor.other | Metze, Florian | |
| dc.contributor.other | Hershey, John R. | |
| dc.date.issued | 2017 | es |
| dc.description.abstract | Recent progress in deep learning has revolutionized speech recognition research, with Deep Neural Networks (DNNs) becoming the new state of the art for acoustic modeling. DNNs offer significantly lower speech recognition error rates compared to those provided by the previously used Gaussian Mixture Models (GMMs). Unfortunately, DNNs are data sensitive, and unseen data conditions can deteriorate their performance. Acoustic distortions such as noise, reverberation, channel differences, etc. add variation to the speech signal, which in turn impact DNN acoustic model performance. A straightforward solution to this issue is training the DNN models with these types of variation, which typically provides quite impressive performance. However, anticipating such variation is not always possible; in these cases, DNN recognition performance can deteriorate quite sharply. To avoid subjecting acoustic models to such variation, robust features have traditionally been used to create an invariant representation of the acoustic space. Most commonly, robust feature-extraction strategies have explored three principal areas: (a) enhancing the speech signal, with a goal of improving the perceptual quality of speech; (b) reducing the distortion footprint, with signal-theoretic techniques used to learn the distortion characteristics and subsequently filter them out of the speech signal; and finally (c) leveraging knowledge from auditory neuroscience and psychoacoustics, by using robust features inspired by auditory perception. In this chapter, we present prominent robust feature-extraction strategies explored by the speech recognition research community, and we discuss their relevance to coping with data-mismatch problems in DNN-based acoustic modeling. We present results demonstrating the efficacy of robust features in the new paradigm of DNN acoustic models. And we discuss future directions in feature design for making speech recognition systems more robust to unseen acoustic conditions. Note that the approaches discussed in this chapter focus primarily on single channel data. | es |
| dc.description.fil | Fil: Mitra, Vikramjit. SRI International; Estados Unidos | |
| dc.description.fil | Fil: Franco, Horacio. SRI International; Estados Unidos | |
| dc.description.fil | Fil: Stern, Richard M.. University of Carnegie Mellon; Estados Unidos | |
| dc.description.fil | Fil: Van Hout, Julien. SRI International; Estados Unidos | |
| dc.description.fil | Fil: Ferrer, Luciana. Consejo Nacional de Investigaciones Científicas y Técnicas. Oficina de Coordinación Administrativa Ciudad Universitaria. Instituto de Investigación en Ciencias de la Computación. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales. Instituto de Investigación en Ciencias de la Computación; Argentina | |
| dc.description.fil | Fil: Graciarena, Martin. SRI International; Estados Unidos | |
| dc.description.fil | Fil: Wang, Wen. SRI International; Estados Unidos | |
| dc.description.fil | Fil: Vergyri, Dimitra. SRI International; Estados Unidos | |
| dc.description.fil | Fil: Alwan, Abeer. University of California at Los Angeles; Estados Unidos | |
| dc.description.fil | Fil: Hansen, John H. L.. University of Texas; Estados Unidos | |
| dc.format | application/pdf | es |
| dc.identifier.citation | Mitra, Vikramjit; Franco, Horacio; Stern, Richard M.; Van Hout, Julien; Ferrer, Luciana; et al.; Robust features in deep-learning-based speech recognition; Springer Nature Switzerland AG; 2017; 183-212 | es |
| dc.identifier.institucional | CONICET | |
| dc.identifier.isbn | 978-3-319-64679-4 | es |
| dc.identifier.repo | CONICET Digital | |
| dc.identifier.uri | http://hdl.handle.net/11336/163169 | |
| dc.journal.ciudad | Cham | es |
| dc.journal.pagination | 183-212 | es |
| dc.journal.pais | Suiza | es |
| dc.language.iso | eng | es |
| dc.publisher | Springer Nature Switzerland AG | es |
| dc.relation.alternativeid | info:eu-repo/semantics/altIdentifier/url/https://link.springer.com/chapter/10.1007/978-3-319-64680-0_8 | |
| dc.relation.alternativeid | info:eu-repo/semantics/altIdentifier/doi/https://doi.org/10.1007/978-3-319-64680-0_8 | |
| dc.rights | info:eu-repo/semantics/restrictedAccess | * |
| dc.rights.uri | https://creativecommons.org/licenses/by-nc-sa/2.5/ar/ | * |
| dc.source.titulo | New era for robust speech recognition: exploiting deep learning | es |
| dc.subject | SPEECH RECOGNITION | es |
| dc.subject | ROBUST FEATURES | es |
| dc.subject | DEEP LEARNING | es |
| dc.subject.classification | Ciencias de la Información y Bioinformática | es |
| dc.subject.classification | Ciencias de la Computación e Información | es |
| dc.subject.classification | CIENCIAS NATURALES Y EXACTAS | es |
| dc.title | Robust features in deep-learning-based speech recognition | es |
| dc.type | info:eu-repo/semantics/publishedVersion | |
| dc.type | info:eu-repo/semantics/bookPart | |
| dc.type | info:ar-repo/semantics/parte de libro | |
Archivos
Bloque original
Cargando...
- Nombre:
- CONICET_Digital_Nro.f8bf955a-4393-4c9d-8275-2cea1cd73fba_A.pdf
- Tamaño:
- 284.39 KB
- Formato:
- Adobe Portable Document Format