Method and apparatus for providing global voice-based entry of geographic information in a device

• US 10,249,298 B2
• Filed: 01/11/2017
• Issued: 04/02/2019
• Est. Priority Date: 01/11/2017
• Status: Active Grant

• Current Assignee: HERE Global B.V. (Nokia Corporation)
• Original Assignee: HERE Global B.V. (Nokia Corporation)
• Inventors: Stajner, Tadej, Dousse, Olivier, Vellasques, Eduardo, Hentz, Augusto Henrique
• Primary Examiner(s): Sharma, Neeraj
• Application Number:
  US15/403,891
• Publication Number:
  US 20180197537A1
• Time in Patent Office: 811 Days
• Field of Search: None
• CPC Class Codes: G01C 21/3608 using speech input, e.g. using speech recognition
• CPC Class Codes: G06F 16/9024 Graphs; Linked lists G06F16/9027 takes precedence
• CPC Class Codes: G10L 15/19 Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 2015/088 Word spotting

Abstract

An approach is provided for global voice-based entry of location information. The approach involves partitioning a global speech decoding graph into spatial partitions. The approach also involves determining key entities occurring in each spatial partitions to construct a combined set of key entities. The approach further involves creating a retrieval index to map the key entities in the combined set of key entities to a corresponding partition. A first partition, the combined set of key entities, and the retrieved index are stored in a memory of a device for processing a voice input signal. A second partition that is not in the memory of the device is retrieved based on the combined set of key entities and the retrieval index to automatically re-process the voice input signal when an out-of-vocabulary result is obtained from the first partition.

Claims

1. A method for automatic speech recognition in a device, comprising:

partitioning a global speech decoding graph into one or more spatial partitions according to a geographic topology of one or more geographic entities, one or more geographic terms, or a combination thereof, wherein each of the spatial partitions contains a decoding graph comprising a sub-set of the one or more geographic entities, the one or more geographic terms, or a combination thereof associated with a geographic area;

determining one or more of key entities occurring in each of the one or more spatial partitions, wherein at least one of the one or more of key entities includes at least one of the one or more geographic entities and the one or more geographic terms;

constructing a combined set of key entities comprising the one or more key entities from said each spatial partition;

creating a retrieval index to map the one or more key entities in the combined set of key entities to a corresponding partition from among the one or more spatial partitions,

wherein a voice input signal associated with a request for one or more navigation or mapping related services is processed, using automatic speech recognition and a first partition associated with a first geographic area from among the one or more spatial partitions, the combined set of key entities, and the retrieved index are stored in a memory of the device, and

wherein a second partition that is associated with a second geographic area and not in the memory of the device is retrieved based on the combined set of key entities and the retrieval index to automatically re-process the voice input signal when an out-of-vocabulary result is obtained with respect to the first partition and to provide the one or more navigation or mapping related services via a user interface based, at least in part, on the re-processing.

2. The method of claim 1, further comprising:

processing the voice-based input to detect the out-of-vocabulary result with respect to the first partition;

in response to the out-of-vocabulary result, recognizing at least one keyword in the voice-based input that is in the combined set of key entities;

determining the second partition based on the at least one keyword and the retrieval index; and

initiating a re-processing of the voice input signal using the second partition.

3. The method of claim 2, wherein the re-processing of the voice input signal is performed recursively until a subsequently retrieved partition does not generate the out-of-vocabulary result or until there are no remaining partitions from among the one or more spatial partitions.

4. The method of claim 2, wherein the determining of the second partition is further based on an importance factor of the at least one keyword, a usage pattern of the at least one keyword, a proximity of an area covered by the at least one keyword, or a combination thereof.

5. The method of claim 2, wherein the second partition is loaded into the memory of the device to initiate the re-processing of the voice-based input, the method further comprising:

determining a failure of the second partition to load into the memory of the device; and

returning the at least one keyword as a partial result until the second partition is loaded successfully into the memory of the device.

6. The method of claim 5, further comprising:

presenting a request for a user to provide another voice input signal to augment the partial result when the second partition is loaded successfully into the memory of the device.

7. The method of claim 2, wherein the first partition, the second partition, or combination thereof is stored in the memory of the device, a local storage unit of the device, a remote storage unit accessible by the device or a data network, or a combination thereof based on a cost function.

8. The method of claim 7, wherein the cost function is based on a mobility pattern, a proximity value, a recommendation engine, or a combination thereof.

9. The method of claim 1, further comprising:

transmitting the voice input signal to a network platform from which the second partition can be retrieved,

wherein the voice input signal is re-processed by the network platform using the second partition.

10. The method of claim 9, wherein the voice input signal, a subsequent voice input signal, or a combination thereof is transmitted to the network platform for re-processing until the second partition is downloaded to the memory of the device.

11. The method of claim 1, further comprising:

determining the one or more key entities based on an occurrence frequency in location queries for said each spatial partition, based on whether the one or more key entities can be further sub-divided, or a combination thereof.

12. The method of claim 1, further comprising:

determining the one or more key entities based on a number of contained entities, a popularity of a destination indicated in the one or more key entities, a user mobility history, or a combination thereof.

13. The method of claim 1, wherein the combined set of key entities includes a lexicon and a grammar associated with the one or more key entities.

14. The method of claim 1, wherein the combined set of key entities is combined into said each spatial partition when said each spatial partition is created.

15. The method of claim 1, wherein the combined set of key entities is combined into said each spatial partition when said each spatial partition is loaded in the memory of the device.

16. An apparatus for automatic speech recognition in a device, comprising:

at least one processor; and

at least one memory including computer program code for one or more programs,

the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus to perform at least the following,

partition a global speech decoding graph into one or more spatial partitions according to a geographic topology of one or more geographic entities, one or more geographic terms, or a combination thereof, wherein each of the spatial partitions contains a decoding graph comprising a sub-set of the one or more geographic entities, the one or more geographic terms, or a combination thereof associated with a geographic area;

determine one or more of key entities occurring in each of the one or more spatial partitions, wherein at least one of the one or more of key entities includes at least one of the one or more geographic entities and the one or more geographic terms;

construct a combined set of key entities comprising the one or more key entities from said each spatial partition;

create a retrieval index to map the one or more key entities in the combined set of key entities to a corresponding partition from among the one or more spatial partitions,

wherein a voice input signal associated with a request for one or more navigation or mapping related services is processed, using automatic speech recognition and a first partition associated with a first geographic area from among the one or more spatial partitions, the combined set of key entities, and the retrieved index are stored in a memory of the device, and

wherein a second partition that is associated with a second geographic area and not in the memory of the device is retrieved based on the combined set of key entities and the retrieval index to automatically re-process the voice input signal when an out-of-vocabulary result is obtained with respect to the first partition and to provide the one or more navigation or mapping related services via a user interface based, at least in part, on the re-processing.

17. The apparatus of claim 16, wherein the apparatus is further caused to:

process the voice-based input to detect the out-of-vocabulary result with respect to the first partition in the memory of the device;

in response to the out-of-vocabulary result, recognize at least one keyword in the voice-based input that is in the combined set of key entities in the memory of the device;

determine the second partition that is not in the memory of the device based on the at least one keyword and the retrieval index; and

initiate a re-processing of the voice input signal using the second partition.

18. The apparatus of claim 17, wherein the re-processing of the voice input signal is performed recursively until a subsequently retrieved partition does not generate the out-of-vocabulary result or until there are no remaining partitions from among the one or more spatial partitions.

19. The apparatus of claim 17, wherein the determining of the second partition based on the at least one keyword and the retrieval index is further based on an importance factor of the at least one keyword, a usage pattern of the at least one keyword, a proximity of an area covered by the at least one keyword, or a combination thereof.

20. A non-transitory computer-readable storage medium for automatic speech recognition in a device, carrying one or more sequences of one or more instructions which, when executed by one or more processors, cause an apparatus to at least perform the following steps:

partitioning a global speech decoding graph into one or more spatial partitions according to a geographic topology of one or more geographic entities, one or more geographic terms, or a combination thereof, wherein each of the spatial partitions contains a decoding graph comprising a sub-set of the one or more geographic entities, the one or more geographic terms, or a combination thereof associated with a geographic area;

determining one or more of key entities occurring in each of the one or more spatial partitions, wherein at least one of the one or more of key entities includes at least one of the one or more geographic entities and the one or more geographic terms;

constructing a combined set of key entities comprising the one or more key entities from said each spatial partition;

creating a retrieval index to map the one or more key entities in the combined set of key entities to a corresponding partition from among the one or more spatial partitions,

wherein a voice input signal associated with a request for one or more navigation or mapping related services is processed, using automatic speech recognition and a first partition associated with a first geographic area from among the one or more spatial partitions, the combined set of key entities, and the retrieved index are stored in a memory of the device, and

wherein a second partition that is associated with a second geographic area and not in the memory of the device is retrieved based on the combined set of key entities and the retrieval index to automatically re-process the voice input signal when an out-of-vocabulary result is obtained with respect to the first partition and to provide the one or more navigation or mapping related services via a user interface based, at least in part, on the re-processing.

21. The non-transitory computer-readable storage medium of claim 20, wherein the apparatus is further caused to perform:

processing the voice-based input to detect the out-of-vocabulary result with respect to the first partition in the memory of the device;

in response to the out-of-vocabulary result, recognizing at least one keyword in the voice-based input that is in the combined set of key entities in the memory of the device;

determining the second partition that is not in the memory of the device based on the at least one keyword and the retrieval index; and

initiating a re-processing of the voice input signal using the second partition.

22. The non-transitory computer-readable storage medium of claim 21, wherein the re-processing of the voice input signal is performed recursively until a subsequently retrieved partition does not generate the out-of-vocabulary result or until there are no remaining partitions from among the one or more spatial partitions.