Adaptation of a speech recognition system across multiple remote sessions with a speaker
DCFirst Claim
1. A method of adapting a speech recognition system, wherein the method comprises steps of:
- a. obtaining an identification of a speaker;
b. obtaining a sample of a speaker'"'"'s speech during a first remote session;
c. recognizing the speaker'"'"'s speech utilizing the speech recognition system during the first remote session;
d. modifying the speech recognition system by incorporating the sample into the speech recognition system thereby forming a speaker-specific modified speech recognition system;
e. storing a representation of the speaker-specific modified speech recognition system in association with the identification of the speaker; and
f. using the representation of the speaker-specific modified speech recognition system to recognize speech during a subsequent remote session with the speaker.
6 Assignments
Litigations
0 Petitions
Reexaminations
Accused Products
Abstract
A technique for adaptation of a speech recognizing system across multiple remote communication sessions with a speaker. The speaker can be a telephone caller. An acoustic model is utilized for recognizing the speaker'"'"'s speech. Upon initiation of a first remote session with the speaker, the acoustic model is speaker-independent. During the first session, the speaker is uniquely identified and speech samples are obtained from the speaker. In the preferred embodiment, the samples are obtained without requiring the speaker to engage in a training session. The acoustic model is then modified based upon the samples thereby forming a modified model. The model can be modified during the session or after the session is terminated. Upon termination of the session, the modified model is then stored in association with an identification of the speaker. During a subsequent remote session, the speaker is identified and, then, the modified acoustic model is utilized to recognize the speaker'"'"'s speech. Additional speech samples are obtained during the subsequent session and, then, utilized to further modify the acoustic model. In this manner, an acoustic model utilized for recognizing the speech of a particular speaker is cumulatively modified according to speech samples obtained during multiple sessions with the speaker. As a result, the accuracy of the speech recognizing system improves for the speaker even when the speaker only engages in relatively short remote sessions.
272 Citations
55 Claims
-
1. A method of adapting a speech recognition system, wherein the method comprises steps of:
-
a. obtaining an identification of a speaker;
b. obtaining a sample of a speaker'"'"'s speech during a first remote session;
c. recognizing the speaker'"'"'s speech utilizing the speech recognition system during the first remote session;
d. modifying the speech recognition system by incorporating the sample into the speech recognition system thereby forming a speaker-specific modified speech recognition system;
e. storing a representation of the speaker-specific modified speech recognition system in association with the identification of the speaker; and
f. using the representation of the speaker-specific modified speech recognition system to recognize speech during a subsequent remote session with the speaker. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15)
-
-
16. A method of adapting a speech recognition system, wherein the method comprises steps of:
-
a. obtaining an identification of a cluster of speakers;
b. obtaining a sample of a speaker'"'"'s speech during a first remote session;
c. recognizing the speaker'"'"'s speech utilizing the speech recognition system during the first remote session;
d. modifying the speech recognition system by incorporating the sample into the speech recognition system thereby forming a cluster-specific modified speech recognition system;
e. storing a representation of the cluster-specific modified speech recognition system in association with the identification of a cluster of speakers wherein the speaker is a member of the cluster; and
f. using the representation of the cluster-specific modified speech recognition system to recognize speech during a subsequent remote session with a member of the cluster of speakers. - View Dependent Claims (17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29)
-
-
30. A method of adapting a speech recognition system, wherein the method comprises steps of:
-
a. obtaining an identification of each of a plurality of speakers during a corresponding first remote session with each speaker;
b. obtaining a sample of speech made by each of the plurality of speakers during a corresponding first remote session with each speaker;
c. recognizing speech made by each speaker during the corresponding first remote session utilizing the speech recognition system configured to be speaker-independent;
d. modifying the speech recognition system by individually incorporating the sample from each speaker into the speech recognition system thereby forming a speaker-specific modified speech recognition system corresponding to each speaker;
e. storing a representation of the speaker-specific modified speech recognition system corresponding to each speaker in association with the identification of the corresponding speaker; and
f. using the representation of the speaker-specific modified speech recognition system corresponding to a speaker to recognize speech during a subsequent remote session with the speaker. - View Dependent Claims (31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44)
-
-
45. A speech recognition system comprising:
-
a. an interface coupled to receive a remote session from a speaker; and
b. a processing system coupled to the interface to obtain an identification of the speaker and to recognize the speaker'"'"'s speech wherein the processing system is cumulatively modified by incorporating speech samples obtained during a plurality of remote sessions with the speaker into the speech recognition system, thereby forming a speaker-specific modified processing system associated with the identification of the speaker. - View Dependent Claims (46, 47, 48, 49, 50)
-
-
51. A method of adapting an acoustic model utilized for speech recognition, wherein the method comprises steps of:
-
a. obtaining an identification of a speaker;
b. obtaining a speech utterance from the speaker during a remote session;
c. recognizing the speaker'"'"'s speech utilizing an acoustic model during the remote session;
d. making a determination relative to the speech utterance; and
e. only when indicated by the determination, performing steps of;
i. modifying the acoustic model by incorporating the speech utterance into the acoustic model thereby forming a speaker-specific modified acoustic model; and
ii. storing a representation of the speaker-specific modified acoustic model in association with the identification of the speaker. - View Dependent Claims (52, 53, 54, 55)
-
Specification