Input-aware and input-unaware iterative speech recognition
An interactive voice response (IVR) system including iterative speech recognition with semantic interpretation is deployed to process an audio input in a manner that optimizes and conserves computing resources and facilitates low-latency discovery of start-of-speech events that can be used to support external processes such as barge-in operations. The IVR system can repeatedly receive an audio input at a speech processing component and apply an input-aware recognition process to the audio input. In response to generating a start-of-speech event, the IVR system can apply an input-unaware recognition process to the remaining audio input and determine a semantic meaning in relation to the relevant portion of the audio input.
1. A method for processing an audio input using an iterative speech recognition system, the method comprising:
receiving the audio input at a speech processing component;
applying an input-aware recognition process to the audio input to generate a start-of-speech event;
responsive to generating the start-of-speech event, applying an input-unaware recognition process to a remaining audio input to determine silence, unexpected speech, or expected speech; and
initiating a response in accordance with a result of the input-unaware recognition process.
2. The method of claim 1 , wherein the input-aware recognition process comprises iteratively analyzing the audio input using an overlapping, high-frequency constant size audio window.
3. The method of claim 1 , wherein the input-unaware recognition process comprises iteratively analyzing the audio input using an increasing audio window size.
4. The method of claim 1 , wherein applying the input-unaware recognition process further comprises:
generating a semantic interpretation result describing a semantic meaning of the audio input.
5. The method of claim 4 , further comprising:
using the semantic interpretation result to control speech complete detection.
6. The method of claim 1 , wherein determining a semantic meaning of the audio input further comprises:
generating a confidence value associated with the semantic meaning.
7. The method of claim 6 , further comprising:
determining whether the confidence value meets or exceeds a confidence threshold.
8. A computer system for processing an audio input, the system comprising:
a processor; and
a memory operably coupled to the processor, the memory having computer-executable instructions stored thereon that, when executed by the processor, causes the system to:
receive the audio input at a speech processing component;
apply an input-aware recognition process to the audio input to generate a start-of-speech event;
responsive to generating the start-of-speech event, apply an input-unaware recognition process to a remaining audio input to determine silence, unexpected speech, or expected speech; and
initiate a response in accordance with a result of the input-unaware recognition process.
9. The computer system of claim 8 , wherein the input-aware recognition process comprises iteratively analyzing the audio input using an overlapping, high-frequency constant size audio window.
10. The computer system of claim 8 , wherein the input-unaware recognition process comprises iteratively analyzing the audio input using an increasing audio window size.
11. The computer system of claim 8 , wherein applying the input-unaware recognition process further comprises:
generating a semantic interpretation result describing a semantic meaning of the audio input.
12. The computer system of claim 11 , wherein the computer-executable instructions further include instructions to cause the processor to:
use the semantic interpretation result to control speech complete detection.
13. The computer system of claim 11 , wherein the computer-executable instructions further include instructions to cause the processor to:
generate a confidence value associated with the semantic meaning.
14. The computer system of claim 13 , wherein the computer-executable instructions further include instructions to cause the processor to:
determine whether the confidence value meets or exceeds a confidence threshold.
15. A non-transitory computer readable medium comprising instructions that, when executed by a processor of a processing system, cause the processing system to perform a method for processing an audio input, comprising instructions to:
receive the audio input at a speech processing component;
apply an input-aware recognition process to the audio input to generate a start-of-speech event;
responsive to generating the start-of-speech event, apply an input-unaware recognition process to a remaining audio input to determine silence, unexpected speech, or expected speech; and
initiate a response in accordance with a result of the input-unaware recognition process.
16. The non-transitory computer readable medium of claim 15 , wherein the input-aware recognition process comprises iteratively analyzing the audio input using an overlapping, high-frequency constant size audio window.
17. The non-transitory computer readable medium of claim 15 , wherein the input-unaware recognition process comprises iteratively analyzing the audio input using an increasing audio window size.
18. The non-transitory computer readable medium of claim 15 , wherein applying the input-unaware recognition process further comprises:
generating a semantic interpretation result describing a semantic meaning of the audio input.
19. The non-transitory computer readable medium of claim 18 , wherein the instructions further include instructions to cause the processor to:
generate a confidence value associated with the semantic meaning.
20. The non-transitory computer readable medium of claim 19 , wherein the instructions further include instructions to cause the processor to:
determine whether the confidence value meets or exceeds a confidence threshold.