Distributed, predictive, dichotomous decision engine for an electronic personal assistant
AssigneeTelepathy Labs, Inc.
InventorsDamien Phelan Stolarz, David Joseph Diaz, James Rossfeld, Scott Raven, Chris O'Malley
Life of this patent
Estimating the term…
Abstract
A system, method and user interface are described for providing a personal assistant functionality using a predictive, adaptive, dichotomous (two choices) decision engine that proactively prompts the user for decisions on matters deemed relevant by the decision engine based on past user decisions and activities.
Claims42
A method of operating a personal assistant comprising:
- (a) Accepting input from the user on a personal electronics device;
- (b) Attempting to process the user input;
- (c) Establishing communication between the personal electronics device and a remote device;
- (d) Sending the user input to the remote device for further processing.
An electronic personal assistant system comprising:
- (a) A first personal electronics device in communication with one or more remote devices;
- (b) A software program running on the first personal electronics device that:
- (i) Accepts input from the user,
- (ii) Attempts to process the input, and
- (iii) Delivers the input to a remote device via a communications link for further processing.
A system comprising one or more remote servers running a software program that:
- (a) Receives input which has been partially processed by a personal electronics device;
- (b) Further processes the input;
- (c) Performs commands or looks up information based upon the input;
- and
- (d) Delivers the results of performing the commands or looking up information to the personal electronic device.
A method of using a computing system to query a user and arrive at a conclusion, comprising:
- (a) Establishing a first question to be answered;
- (b) Establishing a first unknown datum that must be known to answer the first question;
- (c) Formulating a first dichotomous or multiple-choice question to determine the first unknown datum;
- (d) Presenting the first dichotomous or multiple-choice question to the user;
- (e) Receiving a user response;
- (f) Evaluating the relevance of the user's response;
- (g) If the user's response is understood, adding it to the list of known data.
A method of using a computing system to detect and adapt to a user's current mood, comprising:
- (a) Analyzing the user's voice tone, velocity of speech, velocity of responses, or physical movement;
- (b) Classifying their general emotional state according to a database of emotional state characteristics;
- (c) Storing the determined emotional state in an historical database maintained for the user;
- (d) Determining whether the user's emotional state has changed, and whether it is different from the user's usual emotional state;
- (e) If the analyzed emotional state is different from the user's usual emotional state, changing the behavior of the computing system.
A method of using a computing system to predict the data a user might need or want and the actions a user might need or want to undertake, and providing the predicted desired data or performing the predicted actions, comprising:
- (a) Retrieving information about a user from one or more data sources;
- (b) Estimating the user's current circumstances, state or activities;
- (c) Using the user's estimated current circumstances, state or activities to determine likely actions the user might want to undertake or data the user might want to be presented with;
- (d) Prompting the user with an offer to perform an action or to receive data, the answers to said prompts being dichotomous or multiple-choice.
Dependencies are inferred from the claim text.