Performing speech recognition using a local language context including a set of words with descriptions in terms of components smaller than the words
A method of a local recognition system controlling a host device to perform one or more operations is provided. The method includes receiving, by the local recognition system, a query, performing speech recognition on the received query by implementing, by the local recognition system, a local language context comprising a set of words comprising descriptions in terms of components smaller than the words, and performing speech recognition, using the local language context, to create a transcribed query. Further, the method includes controlling the host device in dependence upon the speech recognition performed on the transcribed query.
1. A method of operating a local recognition system hosted by a host device, and controlling the host device to perform one or more operations, the method comprising:
receiving, by the local recognition system, a query;
routing the query using a control module within the local recognition system to a local speech recognition module;
performing speech recognition on the received query by:
accessing, by the local speech recognition module, a local language context comprising:
(i) a first module comprising a vocabulary that includes a set of words, and
(ii) a second module comprising a language model;
implementing, by the local recognition system, an update module external to the local language context, the update module performing adaptation of the first module of the local language context by adding words to, or removing words from, the vocabulary; and
performing speech recognition, by the local speech recognition module using the local language context, to produce a transcribed output that matches content of the received query; and
in dependence on the transcribed output, issuing, by the control module, a command to control the host device to perform the one or more operations.
2. The method of claim 1 , wherein the local language context includes phonetic strings.
3. The method of claim 2 , wherein the local language context includes one or more of the phonetic strings per pronunciation of a word.
4. The method of claim 1 , wherein the local language context includes phonetic lattices.
5. The method of claim 4 , wherein the local language context includes one phonetic lattice per word of the set of words.
6. The method of claim 1 , wherein the vocabulary comprises words and phrases, and the language model comprises a set of constraints on word sequences expressed as N-grams and grammars.
7. A non-transitory computer-readable recording medium having a program recorded thereon, the program, when executed by a processor of a local recognition system hosted by a host device, causes the processor to perform a method comprising:
receiving, by the local recognition system, a query;
routing the query using a control module within the local recognition system to a local speech recognition module;
performing speech recognition on the received query by:
accessing, by the local speech recognition module, a local language context comprising:
(i) a first module comprising a vocabulary including a set of words, and
(ii) a second module comprising a language model;
implementing, by the local recognition system, an update module external to the local language context, the update module performing adaptation of the first module by adding words to, or removing words from, the vocabulary; and
performing speech recognition, by the local speech recognition module using the local language context, to produce a transcribed output that matches content of the received query; and
in dependence on the transcribed output, issuing, by the control module, a command to control the host device to perform one or more operations.
8. The non-transitory computer-readable recording medium of claim 7 , wherein the local language context includes phonetic strings.
9. The non-transitory computer-readable recording medium of claim 8 , wherein the local language context includes one or more of the phonetic strings per pronunciation of a word.
10. The non-transitory computer-readable recording medium of claim 7 , wherein the local language context includes phonetic lattices.
11. The non-transitory computer-readable recording medium of claim 10 , wherein the local language context includes one phonetic lattice per word of the set of words.
12. The non-transitory computer-readable recording medium of claim 7 , wherein the method further comprises
updating speech recognition functionality of the host device over a wireless network.
13. The non-transitory computer-readable recording medium of claim 7 , wherein the method further comprises
performing garbage collection by the update module when available memory resources are about to run out.
14. A local recognition system hosted by a host device and including one or more processors coupled to memory, the memory being loaded with computer instructions to control the host device to perform one or more operations, the computer instructions, when executed on the one or more processors, causing the one or more processors to implement operations comprising:
receiving a query;
routing the query using a control module within the local recognition system to a local speech recognition module;
performing speech recognition on the received query by:
accessing, by the local speech recognition module from a local database, a local language context comprising:
(i) a first module comprising a vocabulary including a set of words, and
(ii) a second module comprising a language model;
implementing, by the local recognition system, an update module external to the local language context, the update module performing adaptation of the first module by adding words to, or removing words from, the vocabulary; and
performing speech recognition, by the local speech recognition module using the local language context, to produce a transcribed output that matches content of the received query; and
in dependence on the transcribed output, issuing, by the control module, a command to control the host device to perform one or more operations.
15. The local recognition system of claim 14 , wherein
the local language context includes phonetic strings.
16. The local recognition system of claim 15 , wherein
the local language context includes one or more of the phonetic strings per pronunciation of a word.
17. The local recognition system of claim 14 , wherein
the local language context includes phonetic lattices.
18. The local recognition system of claim 17 , wherein
the local language context includes one phonetic lattice per word of the set of words.
19. The method of claim 1 , further comprising
updating speech recognition functionality of the host device over a wireless network.
20. The method of claim 1 , further comprising
performing garbage collection by the update module when available memory resources are about to run out.