# US 2014/0207451 A1 · Method and Apparatus of Adaptive Textual Prediction of Voice Data > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2014/0207451 A1 - Document ID: US20140207451A1 - Status: Published application - Kind: A1 - Application number: 13/746,847 - Filed: 2013-01-22 - Published: 2014-07-24 - Earliest claimed: 2013-01-22 - Assignees: Nuance Communications, Inc. - Applicants: Nuance Communications, Inc. - Inventors: Diven Topiwala; Uwe Helmut Jost; Lisa Meredith; Daniel Almendro Barreda - Web page: https://patentgrep.com/patent/US20140207451A1 ## CPC - G10L 15/26: Speech to text systems (G10L15/08 takes precedence) - G10L 15/32: Multiple recognisers used in sequence or in parallel; Score combination systems therefor, e.g. voting systems - G10L 15/083: {Recognition networks (G10L15/142, G10L15/16 take precedence)} - G10L 15/183: using context dependencies, e.g. language models - G10L 2015/227: {of the speaker; Human-factor methodology} - G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING ## Abstract Typical textual prediction of voice data employs a predefined implementation arrangement of a single or multiple prediction sources. Using a predefined implementation arrangement of the prediction sources may not provide a good prediction performance in a consistent manner with variations in voice data quality. Prediction performance may be improved by employing adaptive textual prediction. According to at least one embodiment determining a configuration of a plurality of prediction sources, used for textual interpretation of the voice data, is determined based at least in part on one or more features associated with the voice data or one or more a-priori interpretations of the voice data. A textual output prediction of the voice data is then generated using the plurality of prediction sources according to the determined configuration. Employing an adaptive configuration of the text prediction sources facilitates providing more accurate text transcripts of the voice data. ## Claims ### Claim 1 A computerized method of performing adaptive textual prediction of voice data by a server, the computerized method comprising: determining a configuration of a plurality of prediction sources, used for textual interpretation of the voice data, based at least in part on one or more features associated with the voice data or one or more a-priori interpretations of the voice data; and generating a textual output prediction of the voice data using the plurality of prediction sources according to the configuration determined. ### Claim 2 (depends on 1) A computerized method according to claim 1 further comprising extracting the one or more features associated with the voice data or the one or more a-priori interpretations of the voice data. ### Claim 3 (depends on 1) A computerized method according to claim 1, wherein the one or more features include at least one of the following: a signal-to-noise ratio associated with the voice data; a complexity measure of a lattice representing at least one a-priori interpretation of the voice data; and an a-priori interpretation of the voice message generated by a human agent. ### Claim 4 (depends on 1) A computerized method according to claim 1, wherein the multiple prediction sources include at least one of the following: a language model module; a lattice decoder module; and a human agent. ### Claim 5 (depends on 1) A computerized method according to claim 1 further comprising presenting the textual output prediction to a human agent. ### Claim 6 (depends on 1) A computerized method according to claim 1, wherein determining the configuration of a plurality of prediction sources includes at least one of the following: determining an order according to which the multiple prediction sources are to be applied; determining weightings associated with the multiple prediction sources; and determining a subset of the plurality of prediction sources for use in generating the textual output prediction. ### Claim 7 (depends on 1) A computerized method according to claim 1 further comprising at least one of the following: sending a representation of the configuration determined to another device; storing a representation of the configuration determined; and updating a database storing a representation of a previous configuration of the plurality of prediction sources based on the configuration determined. ### Claim 8 (depends on 7) A computerized method according to claim 7, wherein a representation of the configuration determined includes at least one of the following: indication of an order according to which the multiple prediction sources being applied; indication of weightings associated with the multiple prediction sources; and indication of a subset of the plurality of prediction sources for use in generating the textual output prediction. ### Claim 9 A computer server for performing adaptive textual prediction of speech signals, the computer server comprising: a memory storing computer code instructions thereon; and a processor, the memory, with the computer code instructions, and the processor being configured to cause the server to: determine a configuration of a plurality of prediction sources, used for textual interpretation of the voice data, based at least in part on one or more features associated with at least one of the voice data and one or more a-priori interpretations of the voice data; and generate a textual output prediction of the voice data using the plurality of prediction sources according to the determined configuration. ### Claim 10 (depends on 9) A computer server according to claim 9, wherein the memory, with the computer code instructions, and the processor are configured further to extract the one or more features associated with the voice data or the one or more a-priori interpretations of the voice data. ### Claim 11 (depends on 9) A computer server according to claim 9, wherein the one or more features include at least one of the following: a signal-to-noise ratio associated with the voice data; a complexity measure of a lattice representing at least one a-priori interpretation of the voice data; and an a-priori interpretation of the voice message generated by a human agent. ### Claim 12 (depends on 9) A computer server according to claim 9, wherein the multiple prediction sources include at least one of the following: a language model module; a lattice decoder module; and a human agent. ### Claim 13 (depends on 9) A computer server according to claim 9, wherein the memory, with the computer code instructions, and the processor being configured to further present the textual output prediction to a human agent. ### Claim 14 (depends on 9) A computer server according to claim 9, wherein, in determining the configuration of the plurality of prediction sources, the memory, with the computer code instructions, and the processor are configured further to perform at least one of the following: determine an order according to which the multiple prediction sources are to be applied; determine weightings associated with the multiple prediction sources; and determine a subset of the plurality of prediction sources for use in generating the textual output prediction. ### Claim 15 (depends on 9) A computer server according to claim 9, wherein the memory, with the computer code instructions, and the processor are configured further to perform at least one of the following: send a representation of the configuration determined to another device; store a representation of the configuration determined; and update a database storing a representation of a previous configuration of the plurality of prediction sources based on the configuration determined. ### Claim 16 (depends on 9) A computer server according to claim 9, wherein a representation of the configuration determined includes at least one of the following: An indication of an order according to which the multiple prediction sources are to be applied; an indication of weightings associated with the multiple prediction sources; and an indication of a subset of the plurality of prediction sources for use in generating the textual output prediction. ### Claim 17 A non-transitory computer-readable medium including computer code instructions stored thereon, the computer code instructions, when executed by a processor, cause an apparatus to perform at least the following: determining a configuration of a plurality of prediction sources, used for textual interpretation of the voice data, based at least in part on one or more features associated with at least one of the voice data and one or more a-priori interpretations of the voice data; and generating a textual output prediction of the voice data using the plurality of prediction sources according to the determined configuration. ### Claim 18 (depends on 17) A non-transitory computer-readable medium according to claim 17, wherein the computer code instructions, when executed by the processor, cause the apparatus further to extract the one or more features associated with the voice data or the one or more a-priori interpretations of the voice data. ### Claim 19 (depends on 17) A non-transitory computer-readable medium according to claim 17, wherein the one or more features include at least one of the following: a signal-to-noise ratio associated with the voice data; a complexity measure of a lattice representing at least one a-priori interpretation of the voice data; and an a-priori interpretation of the voice message generated by a human agent. ### Claim 20 (depends on 17) A non-transitory computer-readable medium according to claim 17, wherein the multiple prediction sources include at least one of the following: a language model module; a lattice decoder module; and a human agent.