Learning how to rewrite user-specific input for natural language understanding
Techniques for decreasing (or eliminating) the possibility of a skill performing an action that is not responsive to a corresponding user input are described. A system may train one or more machine learning models with respect to user inputs, which resulted in incorrect actions being performed by skills, and corresponding user inputs, which resulted in the correct action being performed. The system may use the trained machine learning model(s) to rewrite user inputs that, if not rewritten, may result in incorrect actions being performed. The system may implement the trained machine learning model(s) with respect to ASR output text data to determine if the ASR output text data corresponds (or substantially corresponds) to previous ASR output text data that resulted in an incorrect action being performed. If the trained machine learning model(s) indicates the present ASR output text data corresponds (or substantially corresponds) to such previous ASR output text data, the system may rewrite the present ASR output text data to correspond to text data representing a rephrase of the user input that will (or is more likely to) result in a correct action being performed.
1. A computer-implemented method, comprising:
receiving first input data representing a first natural language input;
using a natural language understanding (NLU) component, performing first language processing on the first input data to determine first NLU results data indicating at least an intent of the first natural language input;
using the first NLU results data, determining first output data responsive to the first natural language input;
causing presentation of the first output data;
receiving second input data;
processing the second input data to determine the second input data indicates negative feedback corresponding to the first output data;
based on the second input data indicating the negative feedback, retraining the NLU component to determine an updated NLU component;
after determining the updated NLU component, receiving third input data representing the first natural language input;
using the updated NLU component, performing second language processing on the third input data to determine second NLU results data indicating at least an intent of the first natural language input as represented in the third input data, wherein the first NLU results data is different from the second NLU results data; and
using the second NLU results data, determining second output data responsive to the first natural language input as represented in the third input data, wherein the second output data is different from the first output data.
2. The computer-implemented method of claim 1 , wherein:
the first input data comprises first audio data; and
the second input data comprises second audio data.
3. The computer-implemented method of claim 1 , wherein processing the second input data to determine the second input data indicates negative feedback corresponding to the first output data comprises processing the second input data to determine that the second input data represents a rephrasing of the first natural language input.
4. The computer-implemented method of claim 1 , wherein processing the second input data to determine the second input data indicates negative feedback corresponding to the first output data comprises determining that the second input data corresponds to an interruption of presentation of the first output data.
5. The computer-implemented method of claim 1 , wherein processing the second input data to determine the second input data indicates negative feedback corresponding to the first output data comprises determining that the second input data corresponds to an inquiry corresponding to the first output data.
6. The computer-implemented method of claim 1 , further comprising:
causing presentation of the second output data;
after causing presentation of the second output data, receiving fourth input data; and
processing the fourth input data to determine the fourth input data indicates positive feedback corresponding to the second output data.
7. The computer-implemented method of claim 1 , wherein:
the first input data corresponds to a first profile;
the computer-implemented method further comprises associating the updated NLU component with the first profile; and
the third input data is received from a device associated with the first profile.
8. The computer-implemented method of claim 1 , wherein retraining the NLU component comprises retraining the NLU component using data representing a user sentiment.
9. A system comprising:
at least one processor; and
at least one memory comprising instructions that, when executed by the at least one processor, cause the system to:
receive first input data representing a first natural language input;
using a natural language understanding (NLU) component, perform first language processing on the first input data to determine first NLU results data indicating at least an intent of the first natural language input;
using the first NLU results data, determine first output data responsive to the first natural language input;
cause presentation of the first output data;
receive second input data;
process the second input data to determine the second input data indicates negative feedback corresponding to the first output data;
based on the second input data indicating the negative feedback, retrain the NLU component to determine an updated NLU component;
after determining the updated NLU component, receive third input data representing the first natural language input;
using the updated NLU component, perform second language processing on the third input data to determine second NLU results data indicating at least an intent of the first natural language input as represented in the third input data, wherein the first NLU results data is different from the second NLU results data; and
using the second NLU results data, determine second output data responsive to the first natural language input as represented in the third input data, wherein the second output data is different from the first output.
10. The system of claim 9 , wherein:
the first input data comprises first audio data; and
the second input data comprises second audio data.
11. The system of claim 9 , wherein the instructions that cause the system to process the second input data to determine the second input data indicates negative feedback corresponding to the first output data comprise instructions that, when executed by the at least one processor, cause the system to process the second input data to determine that the second input data represents a rephrasing of the first natural language input.
12. The system of claim 9 , wherein the instructions that cause the system to process the second input data to determine the second input data indicates negative feedback corresponding to the first output data comprise instructions that, when executed by the at least one processor, cause the system to determine that the second input data corresponds to an interruption of presentation of the first output data.
13. The system of claim 9 , wherein the instructions that cause the system to process the second input data to determine the second input data indicates negative feedback corresponding to the first output data comprise instructions that, when executed by the at least one processor, cause the system to determine that the second input data corresponds to an inquiry corresponding to the first output data.
14. The system of claim 9 , wherein the at least one memory further comprises instructions that, when executed by the at least one processor, further cause the system to:
cause presentation of the second output data;
after causing presentation of the second output data, receive fourth input data; and
process the fourth input data to determine the fourth input data indicates positive feedback corresponding to the second output data.
15. The system of claim 9 , wherein:
the first input data corresponds to a first profile;
the at least one memory further comprises instructions that, when executed by the at least one processor, further cause the system to associate the updated NLU component with the first profile; and
the third input data is received from a device associated with the first profile.
16. The system of claim 9 , wherein the instructions that cause the system to retrain the NLU component comprises instructions that, when executed by the at least one processor, cause the system to retrain the NLU component using data representing a user sentiment.