Improving Acoustic Word Embeddings through Correspondence Training of Self-supervised Speech Representations.
Amit MeghananiThomas HainPublished in: CoRR (2024)
Keyphrases
- speech recognition systems
- speech recognition
- prosodic features
- acoustic models
- speech recognizer
- speech sounds
- speech synthesis
- hearing impaired
- speaker verification
- speech recognizers
- automatic speech recognition
- word sense disambiguation
- text to speech
- word error rate
- broadcast news
- acoustic features
- english text
- training set
- spoken document retrieval
- speech signal
- low dimensional
- speech segments
- word recognition
- co occurrence
- hidden markov models
- lexical features
- bird species
- recognition errors
- manifold learning
- emotional speech
- automatic transcription