Improving Acoustic Word Embeddings through Correspondence Training of Self-supervised Speech Representations.
Amit MeghananiThomas HainPublished in: EACL (1) (2024)
Keyphrases
- speech recognition systems
- speech recognition
- acoustic models
- prosodic features
- speech recognizer
- speech recognizers
- speech sounds
- automatic speech recognition
- hearing impaired
- speech synthesis
- speech segments
- speech signal
- training corpus
- spoken document retrieval
- acoustic features
- training process
- co occurrence
- speaker verification
- text to speech
- speaker independent
- word recognition
- recognition errors
- n gram
- training set
- spoken language
- automatic transcription
- english text
- emotional speech
- spontaneous speech
- translation model
- noisy environments
- word sense disambiguation
- feature vectors