Convolutional, long short-term memory, fully connected deep neural networks
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for identifying the language of a spoken utterance. One of the methods includes receiving input features of an utterance; and processing the input features using an acoustic model that comprises one or more convolutional neural network (CNN) layers, one or more long short-term memory network (LSTM) layers, and one or more fully connected neural network layers to generate a transcription for the utterance.
1. A method comprising:
receiving input features, wherein the input features include respective segment features for each of a plurality of segments; and
processing the input features using a model comprising one or more convolutional neural network (CNN) layers and one or more long short-term memory (LSTM) network layers, wherein the processing comprises:
for each of the plurality of segments of the received input features:
providing the respective segment features for the respective segment to:
the one or more CNN layers, and
the one or more LSTM network layers;
generating first features for the respective segment by processing the respective segment features for the respective segment using the one or more convolutional neural network (CNN) layers, wherein the convolutional neural network (CNN) layers perform spatial modeling on the input feature;
generating second features for the respective segment by processing both the respective segment features for the respective segment and the first features generated for the respective segment using the one or more long short-term memory network (LSTM) layers to perform temporal modeling over the first features and the respective segment features, wherein a first layer of the one or more LSTM layers is configured to receive, as input, both the respective segment features for the respective segment and the first features generated for the respective segment; and
determining an output feature based on at least the second features for the plurality of segments.
2. The method of claim 1 , wherein processing the first features using the one or more LSTM layers to generate the second features comprises:
processing the first features using a linear layer to generate reduced features having a reduced dimension from a dimension of the first features; and
processing the reduced features using the one or more LSTM layers to generate the second features.
3. The method of claim 1 , wherein the one or more CNN layers, the one or more LSTM layers, and the one or more fully connected neural network layers have been jointly trained to determine trained values of parameters of the one or more CNN layers, the one or more LSTM layers, and the one or more fully connected neural network layers.
4. The method of claim 1 , wherein the input features include log-mel features having multiple dimensions.
5. A system comprising one or more computers and one or more storage devices storing instructions that when executed by the one or more computers cause the one or more computers to perform operations comprising:
receiving input features, wherein the input features include respective segment features for each of a plurality of segments; and
processing the input features using a model comprising one or more convolutional neural network (CNN) layers and one or more long short-term memory (LSTM) network layers, wherein the processing comprises:
for each of the plurality of segments of the received input features:
providing the respective segment features for the respective segment to:
the one or more CNN layers, and
the one or more LSTM network layers;
generating first features for the respective segment by processing the respective segment features for the respective segment using one or more convolutional neural network (CNN) layers, wherein the convolutional neural network (CNN) layers perform spatial modeling on the input feature;
generating second features for the respective segment by processing both the respective segment features for the respective segment and the first features generated for the respective segment using one or more long short-term memory network (LSTM) layers to perform temporal modeling over the first features and the respective segment features, wherein a first layer of the one or more LSTM layers is configured to receive, as input, both the respective segment features for the respective segment and the first features generated for the respective segment; and
determining an output feature based on at least the second features for the plurality of segments.
6. The system of claim 5 , wherein processing the first features using the one or more LSTM layers to generate the second features comprises:
processing the first features using a linear layer to generate reduced features having a reduced dimension from a dimension of the first features; and
processing the reduced features using the one or more LSTM layers to generate the second features.
7. The system of claim 5 , wherein the one or more CNN layers, the one or more LSTM layers, and the one or more fully connected neural network layers have been jointly trained to determine trained values of parameters of the one or more CNN layers, the one or more LSTM layers, and the one or more fully connected neural network layers.
8. A computer program product encoded on one or more non-transitory computer storage media, the computer program product comprising instructions that when executed by one or more computers cause the one or more computers to perform operations comprising:
receiving input features, wherein the input features include respective segment features for each of a plurality of segments; and
processing the input features using a model comprising one or more convolutional neural network (CNN) layers and one or more long short-term memory (LSTM) network layers, wherein the processing comprises:
for each of the plurality of segments of the received input features:
providing the respective segment features for the respective segment to:
the one or more CNN layers, and
the one or more LSTM network layers;
generating first features for the respective segment by processing the respective segment features for the respective segment using one or more convolutional neural network (CNN) layers, wherein the convolutional neural network (CNN) layers perform spatial modeling on the input feature;
generating second features for the respective segment by processing both the respective segment features for the respective segment and the first features generated for the respective segment using one or more long short-term memory network (LSTM) layers to perform temporal modeling over the first features and the respective segment features, wherein a first layer of the one or more LSTM layers is configured to receive, as input, both the respective segment features for the respective segment and the first features generated for the respective segment; and
determining an output feature based on at least the second features for the plurality of segments.
9. The computer program product of claim 8 , wherein processing the first features using the one or more LSTM layers to generate the second features comprises:
processing the first features using a linear layer to generate reduced features having a reduced dimension from a dimension of the first features; and
processing the reduced features using the one or more LSTM layers to generate the second features.
10. The computer program product of claim 8 , wherein the one or more CNN layers, the one or more LSTM layers, and the one or more fully connected neural network layers have been jointly trained to determine trained values of parameters of the one or more CNN layers, the one or more LSTM layers, and the one or more fully connected neural network layers.