Network based interactive speech recognition system
View Patent ↗A network based interactive speech system responds in real-time to speech-based queries addressed to a set of topic entries. A best matching response is provided based on speech recognition processing and natural language processing performed on recognized speech utterances to identify a selected set of phrases related to the set of topic entries. Another routine converts the selected set of phrases into a search query suitable for identifying a first group of one or more topic entries corresponding to the speech-based query. The words/phrases can be assigned different weightings and can include entries which are not actually in the set of topic entries.
1. A network based interactive speech system adapted for responding to speech-based queries concerning a set of topic entries, the system comprising:
a speech recognition engine adapted to generate recognized speech utterance data from speech data associated with a speech-based query from a speaker concerning one of the set of topic entries;
a first routine executing on a server system and adapted to perform natural language processing on said recognized speech utterance data to identify a selected set of phrases related to the set of topic entries;
a second routine executing on the server system and adapted to convert said selected set of phrases from the first routine into a search query suitable for identifying a first group of one or more topic entries corresponding to said speech-based query;
wherein words and/or phrases in said search query can be assigned different weightings determined by said first routine from said recognized speech utterance data; and
a third routine executing on the server system adapted to evaluate said first group of one or more topic entries and to identify a single topic entry responsive to said speech-based query;
wherein third routine can consider words and/or phrases in said search query which are not in said set of topic entries; and
wherein information corresponding to a single topic entry taken from said first group can be determined and presented in real-time by the interactive speech system automatically as a response best matching said speech-based query.
2. The interactive speech system of claim 1 , wherein said speech recognition engine is distributed between a client device and the server system.
3. The interactive speech system of claim 1 , wherein said single topic entry is selected based on comparing which of said selected set of phrases in said speech-based query also appear as phrases in said first group of one or more topic entries.
4. The interactive speech system of claim 3 , wherein said single topic entry is selected based on determining which of said first group of one or more topic entries has a greatest number of phrases overlapping with said speech-based query.
5. The interactive speech based system of claim 1 , further including an answer routine adapted to control a text-to-speech engine and audibly confirm the single topic entry best matching said speech-based query with the speaker.
6. The interactive speech system of claim 1 , wherein said speech-based query is composed of a list of candidate queries which are concatenated.
7. The interactive speech system of claim 1 , wherein said speech-based query controls an interaction by said speaker with content associated with a web page.
8. The interactive speech system of claim 1 , wherein a confidence level for identifying said topic entry can be specified for said speech-based query.
9. The interactive speech system of claim 1 , wherein said first routine and second routine are part of a natural language engine which performs syntactical processing on said recognized speech utterance data.
10. The interactive speech system of claim 1 , wherein different interactive electronic dialog agents with different control and/or interaction capabilities can be presented by the system to different respective users providing speech-based queries.
11. The interactive speech system of claim 1 , wherein interactive dialog session parameters are adjusted and tailored by the system for individual users.
12. The interactive speech system of claim 1 , wherein characteristics of an interactive electronic dialog agent are adjusted and tailored by the system individually for each speaker in response to configuration data and/or commands from a web page.
13. The interactive speech system of claim 12 , wherein said characteristics of said interactive electronic dialog agent are based on preference data specified by the speaker.
14. The interactive speech system of claim 1 , further including a second server system adapted to assist with a response to said speech-based query, such that said speech based query is recognized at the server system, and multiple databases at different server systems linked to the server system can be considered in responding to queries for said set of topic entries.
15. The interactive speech system of claim 14 , wherein a database at said second server system can be considered for said response to said speech-based query when said server system cannot satisfy a desired confidence level required for said response.
16. The interactive speech system of claim 1 , wherein the system evaluates whether to process said speech query locally at the server system or whether to process said speech query at least in part on another server system.
17. A network based interactive speech system adapted for responding to speech-based queries concerning a set of topic entries, the system comprising:
a speech recognition engine adapted to generate recognized speech utterance data from speech data associated with a speech-based query from a speaker concerning one of the set of topic entries;
a first routine executing on a server system and adapted to perform natural language processing on said recognized speech utterance data to identify a selected set of phrases related to the set of topic entries;
a second routine executing on the server system and adapted to convert said selected set of phrases from the first routine into a search query suitable for identifying a first group of one or more topic entries corresponding to said speech-based query;
wherein words and/or phrases in said search query can be assigned different weightings determined by said first routine from said recognized speech utterance data;
a third routine executing on the server system adapted to evaluate said first group of one or more topic entries and to identify a single topic entry responsive to said speech-based query;
wherein third routine can consider words and/or phrases in said search query which are not in said set of topic entries; and
wherein information corresponding to a single topic entry taken from said first group can be determined and presented in real-time by the interactive speech system automatically as a response best matching said speech-based query; and
a second server system adapted to assist with a response to said speech-based query, such that said speech based query is recognized at the server system, and multiple databases at different server systems linked to the server system can be considered in responding to queries for said set of topic entries, wherein a database at said second server system can be considered for said response to said speech-based query when said server system cannot satisfy a desired confidence level required for said response.
18. The interactive speech system of claim 17 , wherein said speech recognition engine is distributed between a client device and the server system.
19. The interactive speech system of claim 17 , wherein said single topic entry is selected based on comparing which of said selected set of phrases in said speech-based query also appear as phrases in said first group of one or more topic entries.
20. The interactive speech system of claim 19 , wherein said single topic entry is selected based on determining which of said first group of one or more topic entries has a greatest number of phrases overlapping with said speech-based query.
21. The interactive speech based system of claim 17 , further including an answer routine adapted to control a text-to-speech engine and audibly confirm the single topic entry best matching said speech-based query with the speaker.
22. The interactive speech system of claim 17 , wherein said speech-based query is composed of a list of candidate queries which are concatenated.
23. The interactive speech system of claim 17 , wherein said speech-based query controls an interaction by said speaker with content associated with a web page.
24. The interactive speech system of claim 17 , wherein a confidence level for identifying said topic entry can be specified for said speech-based query.
25. The interactive speech system of claim 17 , wherein said first routine and second routine are part of a natural language engine which performs syntactical processing on said recognized speech utterance data.
26. The interactive speech system of claim 17 , wherein different interactive electronic dialog agents with different control and/or interaction capabilities can be presented by the system to different respective users providing speech-based queries.
27. The interactive speech system of claim 17 , wherein interactive dialog session parameters are adjusted and tailored by the system for individual users.
28. The interactive speech system of claim 17 , wherein characteristics of an interactive electronic dialog agent are adjusted and tailored by the system individually for each speaker in response to configuration data and/or commands from a web page.
29. The interactive speech system of claim 28 , wherein said characteristics of said interactive electronic dialog agent are based on preference data specified by the speaker.