An E2E-ASR-Based Iteratively-Trained Timestamp Estimator.
Runyan YangGaofeng ChengPengyuan ZhangYonghong YanPublished in: IEEE Signal Process. Lett. (2022)
Keyphrases
- automatic speech recognition
- maximum likelihood
- training set
- least squares
- multilayer perceptron
- importance sampling
- image processing
- multi layer perceptron
- variance estimator
- access patterns
- training process
- maximum a posteriori
- speech recognition
- monte carlo
- estimation algorithm
- confidence intervals
- speech signal
- multiscale
- image segmentation
- data records
- learning algorithm
- variance reduction
- real time