Extracting Domain Invariant Features by Unsupervised Learning for Robust Automatic Speech Recognition
This addresses robustness issues in ASR systems for speech processing applications, but it is incremental as it applies an existing model to a specific problem.
The paper tackled the problem of automatic speech recognition (ASR) performance degradation due to mismatched training and testing distributions by extracting domain invariant features using an unsupervised Factorized Hierarchical Variational Autoencoder (FHVAE), resulting in absolute word error rate reductions of 41% on Aurora-4 and 27% on CHiME-4.
The performance of automatic speech recognition (ASR) systems can be significantly compromised by previously unseen conditions, which is typically due to a mismatch between training and testing distributions. In this paper, we address robustness by studying domain invariant features, such that domain information becomes transparent to ASR systems, resolving the mismatch problem. Specifically, we investigate a recent model, called the Factorized Hierarchical Variational Autoencoder (FHVAE). FHVAEs learn to factorize sequence-level and segment-level attributes into different latent variables without supervision. We argue that the set of latent variables that contain segment-level information is our desired domain invariant feature for ASR. Experiments are conducted on Aurora-4 and CHiME-4, which demonstrate 41% and 27% absolute word error rate reductions respectively on mismatched domains.