Publication: Deep Maxout Networks applied to Noise-Robust Speech Recognition
Loading...
Identifiers
ISSN: 0302-9743 (print)
ISSN: 1611-3349 (online)
ISBN: 978-3-319-13622-6 (print)
ISBN: 978-3-319-13623-3 (online)
UXXI: CC/0000022425
Publication date
2014
Defense date
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Springer
Impact
Export
Abstract
Deep Neural Networks (DNN) have become very popular for acoustic modeling due to the improvements found over traditional Gaussian Mixture Models (GMM). However, not many works have addressed the robustness of these systems under noisy conditions. Recently, the machine learning community has proposed new methods to improve the accuracy of DNNs by using techniques such as dropout and maxout. In this paper, we investigate Deep Maxout Networks (DMN) for acoustic modeling in a noisy automatic speech recognition environment. Experiments show that DMNs improve substantially the recognition accuracy over DNNs and other traditional techniques in both clean and noisy conditions on the TIMIT dataset.
Description
Proceedings of: IberSPEECH 2014 "VIII Jornadas en Tecnologías del Habla" and "IV Iberian SLTech Workshop". Las Palmas de Gran Canaria, Spain, November 19-21, 2014.
Keywords
Noise robustness, Deep neural networks, Dropout, Deep maxout networks, Speech recognition, Deep learning
Bibliographic citation
Navarro Mesa, J. L., et al. (eds.) (2014). Advances in Speech and Language Technologies for Iberian Languages: Second International Conference, IberSPEECH 2014, Las Palmas de Gran Canaria, Spain, November 19-21, 2014. Proceedings. (pp. 109-118). (Lecture Notes in Computer Science; 8854). Springer International Publishing.