Classification by natural language grammar slots across domains
A virtual assistant processes natural language expressions according to grammar rules created by domain providers. The virtual assistant uniquely identifies each of a multiplicity of users and stores values of grammar slots filled by natural language expressions from each user. The virtual assistant stores histories of slot values and computes statistics from the history. The virtual assistant provider, or a classification client, provides values of attributes of users as labels for a machine learning classification algorithm. The algorithm processes the grammar slot values and labels to compute probability distributions for unknown attribute values of users. A network effect of users and domain grammars make the virtual assistant useful and provides increasing amounts of data that improve classification accuracy and usefulness.
1. A computer-implemented method comprising:
determining grammar slot values by parsing natural language expressions from a multiplicity of users according to a grammar;
storing the grammar slot values associated with the multiplicity of users in a database of personal information;
storing parameter values associated with the multiplicity of users in the database;
at a first time, computing a parameter value of a particular user based on the stored grammar slot values and based on a model trained on the parameter values of other users; and
in response to a request for classification of the particular user at a second time,
providing the computed parameter value to a network-connected device;
incrementing a credit count based on the request
incrementing a credit accumulator for each classification request; and
providing a payment to the domain provider in proportion to the value of the credit accumulator;
determining second grammar slot values by parsing the natural language expressions according to a second grammar, the second grammar being provided by a second domain provider;
in response to the classification request, adding a second increment to a second credit accumulator; and
providing a payment to the second domain provider in proportion to the value of the second credit accumulator.
2. The method of claim 1 wherein classification requests are received and parameter values are provided through a web API interface.
3. The method of claim 1 wherein computing the parameter value comprises computing the parameter value based on a probability.
4. The method of claim 1 wherein computing the parameter value comprises computing the parameter value based on a probability for each of multiple attribute labels.
5. The method of claim 1 further comprising:
charging a classification client to increase the credit count.
6. The method of claim 1 wherein the grammar is provided by a domain provider that is not a virtual assistant provider.
7. The method of claim 1 further comprising:
storing, for the particular user, a historical value for the slot.
8. The method of claim 7 further comprising:
storing a timestamp associated with the historical value.
9. A computer system, comprising:
a processor; and
system memory coupled to the processor and storing instructions to cause the processor to:
determine grammar slot values by parsing natural language expressions from a multiplicity of users according to a grammar;
store the grammar slot values associated with the multiplicity of users in a database of personal information;
store parameter values associated with the multiplicity of users in the database;
at a first time, compute a parameter value of a particular user based on the grammar slot values and based on a model trained on the parameter values of other users; and
in response to a request for classification of the particular user at a second time,
provide the computed parameter value to a network-connected device;
increment a credit count based on the request;
increment a credit accumulator for each classification request; and
provide a payment to the domain provider in proportion to the value of the credit accumulator;
determine second grammar slot values by parsing the natural language expressions according to a second grammar, the second grammar being provided by a second domain provider;
in response to the classification request, add a second increment to a second credit accumulator; and
provide a payment to the second domain provider in proportion to the value of the second credit accumulator.
10. The computer system of claim 9 further comprising instructions configured to:
receive the grammar from a domain provider that is not a virtual assistant provider prior to determining grammar slot values.
11. The computer system of claim 9 further comprising instructions configured to store, for the particular user, a historical value for the slot.
12. The computer system of claim 11 further comprising instructions configured to store, for the particular user, a timestamp associated with the historical value.