Arccosine kernels: Acoustic modeling with infinite neural networks

Chih-Chieh Cheng, Brian KingsburyICASSP 2011 · OpenReview

Abstract

Neural networks are a useful alternative to Gaussian mixture models for acoustic modeling; however, training multilayer networks involves a difficult, nonconvex optimization that requires some "art" to make work well in practice. In this paper we investigate the use of arccosine kernels for speech recognition, using these kernels in a hybrid support vector machine/hidden Markov model recognition system. Arccosine kernels approximate the computation in a certain class of infinite neural networks using a single kernel function, but can be used in learners that require only a convex optimization for training. Phone recognition experiments on the TIMIT corpus show that arccosine kernels can outperform radial basis function kernels.
BibTeX
@inproceedings{cheng2011arccosine,
  author = {Chih-Chieh Cheng and Brian Kingsbury},
  title = {{Arccosine kernels: Acoustic modeling with infinite neural networks}},
  booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing},
  year = {2011},
  url = {https://openreview.net/forum?id=BBSfD-C08lq},
}

Loading related content...

Figure