Acoustic-to-articulatory Speech Inversion with Multi-task Learning.
Yashish M. SiriwardenaGanesh SivaramanCarol Y. Espy-WilsonPublished in: INTERSPEECH (2022)
Keyphrases
- multi task learning
- acoustic features
- formant frequencies
- speech recognition
- speech signal
- multi task
- vocal tract
- automatic speech recognition
- multiple tasks
- learning tasks
- gaussian processes
- multitask learning
- speech synthesis
- learning problems
- visual features
- theoretical analysis
- audio features
- transfer learning
- high order
- multi label image annotation
- learning algorithm
- inductive learning
- image reconstruction
- learning models
- feature vectors
- pairwise
- metric learning
- learning experience
- semi supervised learning
- text mining
- visual speech
- training data