METHOD FOR CHANGING OVER FROM A FIRST ADAPTIVE DATA PROCESSING VERSION TO A SECOND ADAPTIVE DATA PROCESSING VERSION

• US 20090125899A1
• Filed: 05/09/2007
• Published: 05/14/2009
• Est. Priority Date: 05/12/2006
• Status: Active Grant

• Current Assignee: Nuance Communications Austria Gmbh (Microsoft Corporation)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Unfried, Johannes
• Granted Patent: US 9,009,695 B2
• Time in Patent Office: Days
• US Class Current: 717/168
• CPC Class Codes: G06F 8/656 while running
• CPC Class Codes: G10L 15/065 Adaptation
• CPC Class Codes: G10L 15/28 Constructional details of speech recognition systems
• CPC Class Codes: G10L 15/32 Multiple recognisers used in sequence or in parallel; Score combination systems therefor, e.g. voting systems

Abstract

The invention relates to a method and to a system for changing over from a first adaptive data processing version (V1) on data processing means using at least one data model (dm) which is continuously adapted on the basis of data processing results to a second adaptive data processing version (V2) also using at least one data model (DM) to be continuously adapted, characterized in that, in a first phase, the second adaptive data processing version (V2) is used in parallel to the first data processing version (V1), thereby continuously adapting said at least one data model (dm) related to the first version (V1) as well as that data model (DM) related to the second version (V2), and in that the performance of data processing by means of the second version (V2) in checked to comply with a quality criterion, where after in a second phase, as soon as said criterion is met, the results of the data processing by means of the second version (V2) are outputted to be used. The invention further relates to a computer program product having a computer program recorded thereon which is adapted to carry out such a method.

Claims

1. A method for changing over from a first adaptive data processing version (V1) on a data processor using at least one first data model (dm) which is continuously adapted on the basis of data processing results to a second adaptive data processing version (V2) also using at least one second data model (DM) to be continuously adapted, characterized in that, in a first phase, the second adaptive data processing version (V2) is used in parallel to the first data processing version (V1), thereby continuously adapting the at least one first data model (dm) related to the first version (V1) as well as the at least second data model (DM) related to the second version (V2), and in that the performance of data processing by means of the second version (V2) in checked to comply with a quality criterion, whereafter in a second phase, as soon as said quality criterion is met, the results of the data processing by means of the second version (V2) are outputted to be used.

2. The method according to claim 1, characterized in that a given amount of data trained into the second data model (DM) related to the second adaptive data processing version (V2) is used as the quality criterion, in that the amount of adapted data is compared with a given amount of data, and when reaching said given amount of data, an automatic change over to the use of the results of the data processing by means of the second version (V2) is caused.

3. The method according to claim 1, characterized in that the performance of the second data processing version is estimated in relation to the first version results, and in the case of adequate performance change over to the use of the second version results is forced.

4. The method according to claim 1, characterized in that the results of the data processing by means of the second version (V2) are automatically compared with the results of the data processing by means of the first version (V1) with respect to performance, and when the second version results are equal to or better than the first version results, it is automatically changed over to the use of the second version results.

5. The method according to claim 1, characterized in that, by the adaptive data processing, speech data are recognized and automatically converted into text files on the basis of an acoustic reference data model, of a language data model, of a grammar data model and/or of a lexicon data model, said automatically converted text files being checked, and corrected, if necessary, on the basis of the speech data, thereby causing a feedback for continuous adaption of the data model(s).

6. A computer program product having a computer program recorded thereon which is adapted to carry out the method according to claim 1.

7. The computer program product according to claim 6, characterized in that the computer program additionally includes software adapted to carry out said second adaptive data processing version.

8. A system comprising a data processor (4;

42) having a first data processing version (V1) for processing data using at least one first data model (dm) which is continuously adapted on the basis of data processing results, characterized in that the dataprocessor is arranged to run a second data processing version (V2) in parallel to the first data processing version and using at least one second data model (DM) which is continuously adapted on the basis of data processing results, and that the dataprocessor is arranged to switch over from outputting the data processing results of the first data processing version to the results of the second data processing version as soon as an adequate quality of the results of the second data processing version is achieved by the continuous adaption of the respective at least one data model.

9. The system of claim 8, characterized in that a given amount of data trained into the second data model (DM) related to the second adaptive data processing version (V2) is used as a predetermined quality criterion, with a comparator (37, 38) for comparing the amount of adapted data with a given amount of data, and for automatically changing over to the use of the results of the data processing by means of the second version when reaching the given amount of data.

10. The system of claim 8, characterized by a second comparator (37′

) for comparing the results of the data processing by means of the second version with the results of the data processing by means of the first version, and for automatically changing over to the use of the second version results when said second version results are better than the first version results.