Acoustic Word Embeddings for Untranscribed Target Languages with Continued Pretraining and Learned Pooling.
Ramon SanabriaOndrej KlejchHao TangSharon GoldwaterPublished in: INTERSPEECH (2023)
Keyphrases
- expressive power
- speech recognition systems
- target language
- language specific
- statistical machine translation
- dimensionality reduction
- manifold learning
- english text
- source localization
- co occurrence
- compound words
- word recognition
- n gram
- grammar induction
- cross lingual
- previously learned
- low dimensional
- prosodic features