Low-Rank Representation of Nearest Neighbor Phone Posterior Probabilities to Enhance DNN Acoustic Modeling
| Type of publication: | Idiap-RR |
| Citation: | Luyet_Idiap-RR-04-2016 |
| Number: | Idiap-RR-04-2016 |
| Year: | 2016 |
| Month: | 3 |
| Institution: | Idiap |
| Abstract: | We hypothesize that optimal deep neural networks (DNN) class-conditional posterior probabilities live in a union of low-dimensional subspaces. In real test conditions, DNN posteriors encode uncertainties which can be regarded as a superposition of unstructured sparse noise to the optimal posteriors. We aim to investigate different ways to structure the DNN outputs exploiting low-rank representation (LRR) techniques. Using a large number of training posterior vectors, the underlying low-dimensional subspace is identified through nearest neighbor analysis, and low-rank decomposition enables separation of the ``optimal'' posteriors from the spurious uncertainties at the DNN output. Experiments demonstrate that by processing subsets of posteriors which possess strong subspace similarity, low-rank representation enables enhancement of posterior probabilities, and lead to higher speech recognition accuracy based on the hybrid DNN-hidden Markov model (HMM) system. |
| Keywords: | automatic speech recognition (ASR), Deep neural network (DNN) , k-nearest neighbor (kNN) search , low-rank representation (LRR) , posterior probability |
| Projects |
Idiap |
| Authors | |
| Crossref by |
Luyet_INTERSPEECH_2016 |
| Added by: | [ADM] |
| Total mark: | 0 |
|
Attachments
|
|
|
|
|
Notes
|
|
|
|
|
