US 11,810,576 B2
Personalization of experiences with digital assistants in communal settings through voice and query processing
Ruhi Sarikaya, Redmond, WA (US)
Assigned to Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on Oct. 6, 2020, as Appl. No. 17/064,252.
Application 17/064,252 is a continuation of application No. 15/253,762, filed on Aug. 31, 2016, granted, now 10,832,684.
Prior Publication US 2021/0020182 A1, Jan. 21, 2021
Int. Cl. G10L 15/18 (2013.01); G10L 17/06 (2013.01); G10L 25/63 (2013.01); G10L 17/04 (2013.01); G10L 15/22 (2006.01); G10L 17/00 (2013.01); G10L 15/183 (2013.01); G10L 15/08 (2006.01)
CPC G10L 17/06 (2013.01) [G10L 15/183 (2013.01); G10L 15/1815 (2013.01); G10L 15/22 (2013.01); G10L 17/00 (2013.01); G10L 17/04 (2013.01); G10L 25/63 (2013.01); G10L 15/222 (2013.01); G10L 2015/088 (2013.01); G10L 2015/223 (2013.01); G10L 2015/227 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A computer-implemented method for personalized interaction with a digital assistant, comprising:
receiving a first voice input from a first user;
determining, based on an acoustical analysis of the first voice input, that a user profile exists for the first user, the user profile including user information from previous input of the first user;
obtaining user characteristic information for the first user from the user profile;
categorizing the first voice input into one or more topical categories using the user characteristic information;
identifying one or more keywords from the first voice input using a natural language processing model selected in accordance with the one or more topical categories;
obtaining world knowledge for the first voice input based on the one or more keywords and the user characteristic information;
providing a proactive response to the first user based on the user characteristic information, contextual information corresponding to the first user, the one or more keywords, and the world knowledge obtained for the first voice input;
receiving a second voice input from a second user;
determining based on an acoustical analysis of the second voice input that a user profile does not exist for the second user;
identifying, from the second voice input, a background characteristic of the second user;
categorizing the second voice input into the one or more topical categories using the background characteristic;
identifying one or more keywords from the second voice input, using the natural language processing model selected in accordance with the one or more topical categories, wherein the one or more keywords are the same in the first and second voice inputs;
obtaining world knowledge for the second voice input based on the one or more keywords and the identified background characteristic; and
providing a response to the second user based on the identified background characteristic, the one or more keywords, and the obtained world knowledge for the second voice input, wherein the proactive response provided for the first user is different than the response provided for the second user.