Direct enhancement of pre-trained speech embeddings for speech processing in noisy conditions.
Mohamed Nabih AliAlessio BruttiDaniele FalavignaPublished in: Comput. Speech Lang. (2023)
Keyphrases
- speech processing
- pre trained
- speech recognition
- signal processing
- speaker identification
- noisy environments
- natural language processing
- multimedia systems
- english text
- image processing
- training data
- artificial intelligence
- dimensionality reduction
- speech signal
- machine learning
- training examples
- pattern recognition
- variable length
- multi modal
- appearance variations
- language model
- small number
- hidden markov models
- data sets