Feedback system
First Claim
1. A computer-implemented method comprising:
- determining a first context signal corresponding to a user-account;
determining a second context signal corresponding to the user-account;
determining, at a computing device, a first context of the user-account based on a combination of two or more context signals, the two or more context signals including the first context signal and the second context signal, wherein the first context comprises a data-based description or characterization of an environment or state that is associated with the user-account;
using a context-to-action mapping of the user-account to determine a first action that corresponds to the first context, wherein the context-to-action mapping maps a plurality of contexts to a plurality of corresponding actions such that the first context maps to the first action;
initiating the first action;
after initiating the first action, receiving a data signal corresponding to a user-action;
analyzing the data signal to determine a feedback relationship between the user-action and the first action; and
based at least in part on the feedback relationship, updating the context-to-action mapping of the user-account.
7 Assignments
0 Petitions
Accused Products
Abstract
Exemplary methods and systems are disclosed that apply feedback-based machine learning in order to more intelligently use context information to automate certain actions. An exemplary method involves: determining a first context based on a combination of two or more context signals, using a context-to-action mapping to determine a first action that corresponds to the first context, initiating the first action, after initiating the first action, receiving a data signal corresponding to a user-action, analyzing the data signal to determine a feedback relationship between the user-action and the first action, and based at least in part on the feedback relationship, updating the context-to-action mapping.
-
Citations
43 Claims
-
1. A computer-implemented method comprising:
-
determining a first context signal corresponding to a user-account; determining a second context signal corresponding to the user-account; determining, at a computing device, a first context of the user-account based on a combination of two or more context signals, the two or more context signals including the first context signal and the second context signal, wherein the first context comprises a data-based description or characterization of an environment or state that is associated with the user-account; using a context-to-action mapping of the user-account to determine a first action that corresponds to the first context, wherein the context-to-action mapping maps a plurality of contexts to a plurality of corresponding actions such that the first context maps to the first action; initiating the first action; after initiating the first action, receiving a data signal corresponding to a user-action; analyzing the data signal to determine a feedback relationship between the user-action and the first action; and based at least in part on the feedback relationship, updating the context-to-action mapping of the user-account. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23)
-
-
24. A computer-implemented method comprising:
-
detecting at least one context signal; determining, at a computing device, a first context based on the at least one context signal; using a context-to-action mapping to determine a first action that corresponds to the first context, wherein the context-to-action mapping maps a plurality of contexts to a plurality of corresponding actions such that the at least one context signal maps to the first action; initiating the first action; after initiating the first action, receiving a data signal corresponding to a user-action; analyzing the data signal to determine a feedback relationship between the user-action and the first action, wherein the feedback relationship indicates a desirability difference between the initiation of the first action and one or more previous initiations of the first action in a substantially same context; responsive to the feedback, determining one or more context signals that have a likelihood of reconciling the desirability difference between the initiation of the first action and one or more previous initiations of the first action in a substantially same context; and updating the context-to-action mapping of the user-account to indicate that the first context should be determined based on both;
(a) the at least one context signal and (b) the one or more context signals that have a likelihood of reconciling the desirability difference. - View Dependent Claims (25, 26)
-
-
27. A system comprising:
-
a non-transitory computer-readable medium; and program instructions stored on the non-transitory computer-readable medium and executable by at least one processor to; determine a first context of a user-account based on a combination of two or more context signals corresponding to the user-account, the two or more context signals including a first context signal and a second context signal, wherein the first context comprises a data-based description or characterization of an environment or state that is associated with the user-account; use a context-to-action mapping of the user-account to determine a first action that corresponds to the first context, wherein the context-to-action mapping maps a plurality of contexts to a plurality of corresponding actions such that the first context maps to the first action; initiate the first action; after initiating the first action, receive a data signal corresponding to a user-action; analyze the data signal to determine a feedback relationship between the user-action and the first action; and based at least in part on the feedback relationship, update the context-to-action mapping of the user-account. - View Dependent Claims (28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40)
-
-
41. A system comprising:
-
a means for determining a first context of a user-account based on a combination of two or more context signals corresponding to the user-account, wherein the first context comprises a data-based description or characterization of an environment or state that is associated with the user-account; a means for using a context-to-action mapping of the user-account to determine a first action that corresponds to the first context, wherein the context-to-action mapping maps a plurality of contexts to a plurality of corresponding home-automation actions, such that the first context maps to the first action; a means for initiating the first action; a means for, after initiating the first action, receiving a data signal corresponding to a user-action; a means for, analyzing the data signal to determine a feedback relationship between the user-action and the first action; and a means for, based at least in part on the feedback relationship, updating the context-to-action mapping of the user-account.
-
-
42. A computer-implemented method comprising:
-
determining a first context signal corresponding to a user-account, wherein the first context signal provides information related to an environment or state that is associated with the user-account; determining a second context signal corresponding to the user-account, wherein the second context signal provides information related to an environment or state that is associated with the user-account; using a context-to-action mapping of the user-account to determine a first action that corresponds to a combination of the first context signal and the second context signal, wherein the context-to-action mapping maps a plurality of context-signal combinations to a plurality of corresponding actions such that the combination of the first context signal and the second context signal maps to the first action; initiating the first action; after initiating the first action, receiving a data signal corresponding to a user-action; analyzing the data signal to determine a feedback relationship between the user-action and the first action; and based at least in part on the feedback relationship, updating the context-to-action mapping of the user-account.
-
-
43. A computer-implemented method comprising:
-
detecting at least one context signal; determining, at a computing device, a first context based on the at least one context signal; using a context-to-action mapping to determine a first action that corresponds to the first context, wherein the context-to-action mapping maps a plurality of contexts to a plurality of corresponding actions such that the at least one context signal maps to the first action; initiating the first action; after initiating the first action, receiving a data signal corresponding to a user-action; analyzing the data signal to determine a feedback relationship between the user-action and the first action, wherein the feedback relationship indicates a desirability difference between the initiation of the first action and one or more previous initiations of the first action in a substantially same context; responsive to the feedback, initiating a prompt for input data to determine one or more context signals that have a likelihood of reconciling the desirability difference between the initiation of the first action and one or more previous initiations of the first action in a substantially same context; and updating the context-to-action mapping to indicate that the first context should be determined based on both;
(a) the at least one context signal and (b) the one or more context signals that have a likelihood of reconciling the desirability difference.
-
Specification