· 2013
Deep convolutional neural networks for LVCSR
- 1,072citations
- 2013year
Short summary
Deep convolutional neural networks (CNNs) achieve 13-30% relative improvement over GMMs and 4-12% over DNNs for large vocabulary continuous speech recognition (LVCSR) tasks.
AI-generated from the title and abstract; the full text is not read.
Key points
- CNNs are proposed as a superior model for LVCSR due to their ability to reduce spectral variations and model spectral correlations in speech signals.
- The study investigates optimal CNN architectures for LVCSR, including the number of convolutional layers, hidden units, pooling strategy, and input features.
- CNNs demonstrated a 13-30% relative improvement compared to GMMs on LVCSR tasks.
- CNNs showed a 4-12% relative improvement compared to DNNs on LVCSR tasks.
AI-generated from the title and abstract; the full text is not read.
Abstract
Convolutional Neural Networks (CNNs) are an alternative type of neural network that can be used to reduce spectral variations and model spectral correlations which exist in signals. Since speech signals exhibit both of these properties, CNNs are a more effective model for speech compared to Deep Neural Networks (DNNs). In this paper, we explore applying CNNs to large vocabulary speech tasks. First, we determine the appropriate architecture to make CNNs effective compared to DNNs for LVCSR tasks. Specifically, we focus on how many convolutional layers are needed, what is the optimal number of hidden units, what is the best pooling strategy, and the best input feature type for CNNs. We then explore the behavior of neural network features extracted from CNNs on a variety of LVCSR tasks, comparing CNNs to DNNs and GMMs. We find that CNNs offer between a 13-30% relative improvement over GMMs, and a 4-12% relative improvement over DNNs, on a 400-hr Broadcast News and 300-hr Switchboard task.
The authors' abstract, as published at the source., 2013 · DOI ↗
Continue with a free account
Ask the paper: 3 free questions a day about this paper; save it, get its citation, new summaries every day for your field. Takeaways are Premium.
Continue free on the webSign in with Google or Apple; no card needed. You come back to this paper.
On your phone:
Field: Signal Processing
Signal ProcessingComputer Science