Combining Bottleneck-BLSTM and Semi-Supervised Sparse NMF for Recognition of Conversational Speech in Highly Instationary Noise.
Felix WeningerMartin WöllmerBjörn W. SchullerPublished in: INTERSPEECH (2012)
Keyphrases
- semi supervised
- conversational speech
- negative matrix factorization
- object recognition
- semi supervised learning
- unsupervised learning
- sparse representation
- automatic speech recognition
- speech retrieval
- nonnegative matrix factorization
- activity recognition
- high dimensional
- pattern recognition
- labeled data
- feature extraction
- missing data
- spectral clustering
- broadcast news
- neural network