Generating representations of acoustic sequences
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating representation of acoustic sequences. One of the methods includes: receiving an acoustic sequence, the acoustic sequence comprising a respective acoustic feature representation at each of a plurality of time steps; processing the acoustic feature representation at an initial time step using an acoustic modeling neural network; for each subsequent time step of the plurality of time steps: receiving an output generated by the acoustic modeling neural network for a preceding time step, generating a modified input from the output generated by the acoustic modeling neural network for the preceding time step and the acoustic representation for the time step, and processing the modified input using the acoustic modeling neural network to generate an output for the time step; and generating a phoneme representation for the utterance from the outputs for each of the time steps.
1. A method comprising:
receiving an acoustic sequence, the acoustic sequence representing an utterance, and the acoustic sequence comprising a respective acoustic feature representation at each of a plurality of time steps; and
generating, for use in a speech recognition system, a phoneme representation for the utterance, comprising:
at a particular time step of the plurality of time steps:
combining the acoustic feature representation at the particular time step and a preceding output generated by an acoustic modeling neural network for a preceding time step to generate a modified input; and
processing the modified input using the acoustic modeling neural network to generate an output for the particular time step.
2. The method of claim 1 , wherein the acoustic modeling neural network is a feed-forward neural network.
3. The method of claim 1 , wherein the acoustic modeling neural network is a recurrent neural network.
4. The method of claim 3 , wherein the acoustic modeling neural network is a long short-term memory (LSTM) neural network.
5. The method of claim 1 , wherein the output for the acoustic feature representation at the particular time step is a set of scores for a set of phonemes or phoneme subdivisions, wherein the score for each phoneme or phoneme subdivision represents a likelihood that the phoneme or phoneme subdivision is a representation of the utterance at the particular time step.
6. The method of claim 5 , wherein generating the modified input comprises appending the set of scores for the preceding time step to the acoustic feature representation for the particular time step.
7. The method of claim 5 , wherein generating the modified input comprises appending data identifying a highest-scoring phoneme or phoneme subdivision according to the set of scores for the preceding time step to the acoustic feature representation for the particular time step.
8. The method of claim 5 , wherein the set of scores defines a probability distribution over a set of Hidden Markov Model (HMM) states.
9. The method of claim 1 , wherein phoneme representation for the utterance is generated from the outputs generated for each of the respective acoustic feature representations.
10. A system comprising one or more computers and one or more storage devices storing instructions that when executed by the one or more computers causes the one or more computers to perform operations comprising:
receiving an acoustic sequence, the acoustic sequence representing an utterance, and the acoustic sequence comprising a respective acoustic feature representation at each of a plurality of time steps; and
generating, for use in a speech recognition system, a phoneme representation for the utterance, comprising:
at a particular time step of the plurality of time steps:
combining the acoustic feature representation at the particular time step and a preceding output generated by an acoustic modeling neural network for a preceding time step to generate a modified input; and
processing the modified input using the acoustic modeling neural network to generate an output for the particular time step.
11. The system of claim 10 , wherein the acoustic modeling neural network is a feed-forward neural network.
12. The system of claim 10 , wherein the acoustic modeling neural network is a recurrent neural network.
13. The system of claim 12 , wherein the acoustic modeling neural network is a long short-term memory (LSTM) neural network.
14. The system of claim 10 , wherein the output for the acoustic feature representation at the particular time step is a set of scores for a set of phonemes or phoneme subdivisions, wherein the score for each phoneme or phoneme subdivision represents a likelihood that the phoneme or phoneme subdivision is a representation of the utterance at the particular time step.
15. The system of claim 14 , wherein generating the modified input comprises appending the set of scores for the preceding time step to the acoustic feature representation for the particular time step.
16. The system of claim 14 , wherein generating the modified input comprises appending data identifying a highest-scoring phoneme or phoneme subdivision according to the set of scores for the preceding time step to the acoustic feature representation for the particular time step.
17. A non-transitory computer storage medium encoded with a computer program, the computer program comprising instructions that when executed by one or more computers cause the one or more computers to perform operations comprising:
receiving an acoustic sequence, the acoustic sequence representing an utterance, and the acoustic sequence comprising a respective acoustic feature representation at each of a plurality of time steps; and
generating, for use in a speech recognition system, a phoneme representation for the utterance, comprising:
at a particular time step of the plurality of time steps:
combining the acoustic feature representation at the particular time step and a preceding output generated by an acoustic modeling neural network for a preceding time step to generate a modified input; and
processing the modified input using the acoustic modeling neural network to generate an output for the particular time step.
18. The non-transitory computer storage medium of claim 17 , wherein the output for the acoustic feature representation at the particular time step is a set of scores for a set of phonemes or phoneme subdivisions, wherein the score for each phoneme or phoneme subdivision represents a likelihood that the phoneme or phoneme subdivision is a representation of the utterance at the particular time step.
19. The non-transitory computer storage medium of claim 18 , wherein generating the modified input comprises appending the set of scores for the preceding time step to the acoustic feature representation for the particular time step.
20. The non-transitory computer storage medium of claim 18 , wherein generating the modified input comprises appending data identifying a highest-scoring phoneme or phoneme subdivision according to the set of scores for the preceding time step to the acoustic feature representation for the particular time step.