Rapid speaker adaptation in eigenvoice space

By Roland Kuhn, Jean-Claude Junqua, Patrick Nguyen, Nancy Niedzielski. IEEE Transactions on Speech and Audio Processing, 2000

569 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Signal Processing

View Paper: https://www.semanticscholar.org/paper/82c3b15d146850b9b8b320e40f68ecbe35c08c6e

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Dong Yu
2. Li Deng
3. Jinyu Li
4. Junichi Yamagishi
5. Heiga Zen
6. Shinji Watanabe
7. Yoshua Bengio
8. Yifan Gong

In Signal Processing

1. Dong Yu
2. Jinyu Li
3. Li Deng
4. Junichi Yamagishi
5. Heiga Zen
6. Yifan Gong
7. Daniel Povey
8. Shinji Watanabe

Citations in the corpus

• Machine Learning Paradigms for Speech Recognition: An Overview (2013, unknown): “In a similar spirit, eigenvoice [214] con-”
• Comparison of discriminative input and output transformations for speaker adaptation in the hybrid NN/HMM systems (2010, methodology): “Adaptation approaches such as “eigenvoices”[11] construct a set of bases from training speakers and for each new speaker, the adapted model is given by W = WSI + ∑N k=1 αkBk, whereWSI is the SI model and {Bk|k = 1, 2, .”
• Improving DNN speaker independence with I-vector inputs (2014, methodology): “In experiments on the TIMIT dataset, they used backpropagation to learn a separate code for each speaker.”
• Porting concepts from DNNs back to GMMs (2013, background): “This assures that the resulting recognition system can still rely on the vast array of GMM-based techniques developed in the last decades to cope with complex tasks such as speaker and environment adaptation [5, 6] and model-based noise robustness [7, 8, 9].”
• The subspace Gaussian mixture model - A structured model for speech recognition (2011, methodology): “This method has similarities to Joint Factor Analysis as used in speaker recognition [19] and to Eigenvoices [21] and Cluster Adaptive Training (CAT) [10] as proposed for speech recognition.”
• Speech Synthesis Based on Hidden Markov Models (2013, methodology): “The framework of probabilistic PCA can similarly be applied to HMM-based speech synthesis systems [28] to improve acoustic modeling.”
• Statistical Parametric Speech Synthesis (2007, methodology): “applied the eigenvoice technique (Kuhn et al., 2000) to HMM-based speech synthesis (Shichiri et al.”
• An Overview of Noise-Robust Automatic Speech Recognition (2014, methodology): “By maximizing the auxiliary function, the combination weight can be solved as (69) This model combination method is very similar to general speaker adaptation methods such as cluster adaptive training (CAT) [153] and eigenvoice [154].”
• Eigentriphones for Context-Dependent Acoustic Modeling (2013, methodology): “The procedure is called maximum-likelihood eigen-decomposition (MLED) in [22].”