Multi-modal device capable of automated actions

  • US 7,778,632 B2
  • Filed: 10/28/2005
  • Issued: 08/17/2010
  • Est. Priority Date: 10/28/2005
  • Status: Active Grant
  • ×
    • Pin Icon | RPX Insight
    • Pin
First Claim
Patent Images

1. A multi-modal multi-lingual mobile device that facilitates automating an action, comprising:

  • a detection component that employs a plurality of integrated sensors to obtain at least one criterion from an auxiliary act through passive observation, the auxiliary act is a conversation of a user with an entity that is not the mobile device, wherein the at least one criterion is an environmental context that relates to a weather condition, and a schedule manipulation action is performed at least partially in view of at least one of expected travel complications or venue incompatibility with the weather condition; and

    an analyzer component that evaluates the at least one criterion to infer a user intent and automatically implements the action based at least in part upon operation of a rules-based logic component, wherein the rules based logic component automatically allows execution of the action based at least in part upon satisfaction of a defined rule, and based at least in part upon operation of an implementation component configured to identify an individual, the implementation component using an algorithm together with a desired degree of certainty, and based at least in part upon operation of an artificial intelligence component that comprises a classifier function that maps an input attribute vector x=(x1, x2, x3, x4, xn) to a confidence that input associated with the vector belongs to a class, wherein the xi, are input attributes, wherein the confidence that the input belongs to the class is expressed as f(x)=confidence(class), and wherein the class to which an input belongs infers the action that the user desires to be automatically performed;

    wherein the auxiliary act is not for an explicit purpose of implementing the action.

View all claims

    Thank you for your feedback