When mismatched training data outperform matched data - Archive ouverte HAL Access content directly
Conference Papers Year :

When mismatched training data outperform matched data

(1)
1

Abstract

My talk will focus on robustness to background noise in distant-microphone speech recordings. I will introduce deep learning based techniques for speech enhancement and for acoustic modeling of speech. I will then report the results of a study on the impact of environment and microphone mismatches on the recognition accuracy. This study reveals that mismatched training data can sometimes outperform matched data. I will suggest a way to optimize the training set in order to exploit this finding.
Fichier principal
Vignette du fichier
vincent_ESI17.pdf (5.04 Mo) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01588876 , version 1 (17-09-2017)

Identifiers

  • HAL Id : hal-01588876 , version 1

Cite

Emmanuel Vincent. When mismatched training data outperform matched data. Systematic approaches to deep learning methods for audio, Sep 2017, Vienna, Austria. ⟨hal-01588876⟩
337 View
140 Download

Share

Gmail Facebook Twitter LinkedIn More