| US 7,451,081 B1 | ||
| System and method of performing speech recognition based on a user identifier | ||
| Bojana Gajic, Trondheim (Norway); Shrikanth Sambasivan Narayanan, Los Angeles, Calif. (US); Sarangarajan Parthasarathy, New Providence, N.J. (US); Richard Cameron Rose, Watchung, N.J. (US); and Aaron Edward Rosenberg, Berkeley Heights, N.J. (US) | ||
| Assigned to AT&T Corp., New York, N.Y. (US) | ||
| Filed on Mar. 13, 2007, as Appl. No. 11/685,456. | ||
| Application 11/685456 is a continuation of application No. 10/091689, filed on Mar. 06, 2002, granted, now 7,209,880. | ||
| Claims priority of provisional application 60/277231, filed on Mar. 20, 2001. | ||
| Int. Cl. G10L 15/00 (2006.01) | ||
| U.S. Cl. 704—231 [704/233; 704/243; 704/244] | 20 Claims |
| 1. A method of performing speech recognition comprising:
receiving a voice request from a speaker;
receiving a data field selection by the speaker;
applying one of a plurality of language models to the received voice request for speech recognition based on the selected
data field, wherein the speech recognition uses a different language model for each data field selected by the speaker;
determining an identity of the speaker based, at least in part, on a user identifier;
repeatedly determining parameters of a background model based on sampled information collected at a periodic time interval
during the received voice request;
determining parameters of a transducer model; and
adapting a speech recognition model based on user-specific transformations corresponding to the determined identity of the
speaker and on at least one of the background model or the transducer model.
|