Systems and methods for providing information discovery and retrieval

• US 8,862,615 B1
• Filed: 07/25/2013
• Issued: 10/14/2014
• Est. Priority Date: 11/12/2007
• Status: Expired due to Fees

• Current Assignee: Wyoming Technology Licensing LLC (Leigh M. Rothschild)
• Original Assignee: W. Leo Hoarty
• Inventors: Hoarty, W. Leo
• Primary Examiner(s): LE, UYEN T
• Application Number:
  US13/951,083
• Time in Patent Office: 446 Days
• Field of Search: None
• US Class Current: 707/769
• CPC Class Codes: G06F 16/43 Querying
• CPC Class Codes: G06F 16/48 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• CPC Class Codes: G06F 16/639 using playlists
• CPC Class Codes: G06F 16/686 using information manually generated, e.g. tags, keywords, comments, title or artist information, time, location or usage information, user ratings
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/953 Querying, e.g. by the use of web search engines
• CPC Class Codes: G06F 40/211 Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
• CPC Class Codes: G10L 15/005 Language recognition
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 15/14 using statistical models, e.g. Hidden Markov Models [HMMs] G10L15/18 takes precedence
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

This invention relates generally to software and computers, and more specifically, to systems and methods for providing information discovery and retrieval. In one embodiment, the invention includes a system for providing information discovery and retrieval, the system including a processor module, the processor module configurable to performing the steps of receiving an information request from a consumer device over a communications network; decoding the information request; discovering information using the decoded information request; preparing instructions for accessing the information; and communicating the prepared instructions to the consumer device, wherein the consumer device is configurable to retrieving the information for presentation using the prepared instructions.

Claims

1. A method, comprising:

receiving an utterance captured via a consumer device;

providing a media request to at least one server based at least in part on the utterance, including at least;

performing, by the at least one server, one or more of pitch shifting or time shifting of the utterance via distributing a plurality of test samples of the utterance to a plurality of processing devices for testing in parallel, each test sample differently shifted in at least one of pitch or time;

establishing a plurality of confidence values corresponding to the plurality of test samples based at least in part on the testing in parallel; and

selecting one test sample as the shifted media request based at least in part on the plurality of confidence values;

decoding the shifted media request at least via iteratively decoding portions of the media request;

discovering at least some information using the decoded media request;

preparing one or more instructions for consumer device access to the at least some information; and

communicating the prepared one or more instructions to at least one consumer device.

2. The method of claim 1, wherein providing a media request to at least one server based at least in part on the utterance comprises:

initializing the utterance, including one or more of normalizing volume, removing noise, performing bandpass filtering, detecting clipping in the utterance, or compressing the utterance.

3. The method of claim 2, wherein initializing the utterance, including one or more of normalizing volume, removing noise, performing bandpass filtering, detecting clipping in the utterance, or compressing the utterance comprises:

receiving, at a communications server, the utterance; and

initializing the utterance at the communications server.

4. The method of claim 2, wherein initializing the utterance, including one or more of normalizing volume, removing noise, performing bandpass filtering, detecting clipping in the utterance, or compressing the utterance comprises:

initializing the utterance by at least one of a consumer device or a phone receiving the utterance from a speaker associated with the utterance.

5. The method of claim 2, wherein initializing the utterance, including one or more of normalizing volume, removing noise, performing bandpass filtering, detecting clipping in the utterance, or compressing the utterance comprises:

invoking at least one processing device to at least partially initialize one or more portions of the utterance.

6. The method of claim 5, wherein invoking at least one processing device to at least partially initialize one or more portions of the utterance comprises:

invoking at least two processing devices to at least partially initialize one or more portions of the utterance, wherein at least a first processing device at least partially initializes a first portion of the utterance and at least a second processing device at least partially initializes a second portion of the utterance.

7. The method of claim 5, wherein invoking at least one processing device to at least partially initialize one or more portions of the utterance comprises:

invoking at least two processing devices to at least partially initialize one or more portions of the utterance, wherein at least a first processing device performs a first initialization task, and wherein at least a second processing device performs at least a second initialization task different from the first initialization task, the first and second initialization tasks including one of normalizing volume, removing noise, performing bandpass filtering, or compressing the utterance.

8. The method of claim 1, wherein receiving an utterance captured via a consumer device and providing a media request to at least one server based at least in part on the utterance comprise:

receiving digital audio from a consumer device including at least a multiple microphone array, the multiple microphone array including at least two microphones;

determining at least a portion of the digital audio for providing the media request based at least in part on the utterance, including at least;

performing beam-forming to focus the sensitivity of the multiple microphone array on a speaker associated with the utterance; and

determining at least a portion of the digital audio to isolate from the utterance.

9. The method of claim 1, wherein receiving an utterance captured via a consumer device and providing a media request to at least one server based at least in part on the utterance comprise:

receiving digital audio including at least an utterance and a sample of ambient noise;

determining at least a portion of the digital audio including the sample of ambient noise; and

reducing noise from the utterance associated with the digital audio via at least removing noise from the digital audio similar to the ambient noise.

10. The method of claim 1, wherein providing a media request to at least one server based at least in part on the utterance comprises:

determining a portion of the utterance containing at least one of background or interfering speech via at least in part using a speech recognizer to isolate a portion of the utterance associated with at least one of the speaker of the utterance or a background speaker.

11. The method of claim 1, wherein providing a media request to at least one server based at least in part on the utterance comprises:

determining that a background noise portion of the information utterance interferes with processing the utterance; and

alerting a speaker that a background noise portion of the information utterance interferes with processing the utterance.

12. The method of claim 11, wherein alerting a speaker that a background noise portion of the information utterance interferes with processing the utterance comprises:

alerting a speaker including at least one of alerting the speaker to move to a quieter location, alerting the speaker to reduce background noise, or alerting the speaker to utter the information request again.

13. The method of claim 1, wherein providing a media request to at least one server based at least in part on the utterance comprises:

identifying a speaker associated with the utterance;

receiving one or more of speech pattern characteristics, usage history, or preferences associated with the identified speaker; and

providing the media request based at least in part on the one or more of speech pattern characteristics, usage history, or preferences associated with the identified speaker.

14. The method of claim 1, wherein providing a media request to at least one server based at least in part on the utterance comprises:

performing at least one of pitch shifting or time shifting of one or more portions of the utterance.

15. The method of claim 14, wherein performing at least one of pitch shifting or time shifting of the utterance at least via applying learned digital signal processing comprises:

applying at least one of a reference pitch-shift or time-shift determined during at least one prior decoding of a media request.

16. The method of claim 15, wherein prior decoding of a media request comprises:

determining at least one reference pitch-shift or time-shift, including at least;

receiving an audio sample of at least a known test phrase uttered by a speaker;

transforming the audio sample into a plurality of test samples, each test sample including the audio sample shifted in one or more of pitch and time;

establishing a confidence value for each of the plurality of test samples, wherein the establishing a confidence value includes at least comparing each of the plurality of test samples with the known test phrase;

selecting the test sample with a most favorable confidence value for a reference pitch-shift or time-shift; and

storing the reference pitch-shift or time-shift for subsequent decoding.

17. The method of claim 16, wherein transforming the audio sample into a plurality of test samples, each test sample including the audio sample shifted in one or more of pitch and time comprises:

invoking a processor module for transforming the audio sample into a particular test sample, each particular test sample including the audio sample shifted in one of pitch or time,

wherein a plurality of processor modules each perform the operations of transforming the audio sample and establishing a confidence value for an individual test sample, each of the plurality of processor modules performing the operations in parallel with the other processor modules.

18. The method of claim 17, further comprising:

decoding the media request at least via speech recognition optimized for the reference accent.

19. The method of claim 1, wherein providing a media request to at least one server based at least in part on the utterance comprises:

performing at least one of pitch shifting or time shifting of the utterance at least via applying learned digital signal processing.

20. The method of claim 1, wherein providing a media request to at least one server based at least in part on the utterance comprises:

identifying a reference accent, the reference accent determined during a prior decoding of a media request, the prior decoding of a media request including at least;

receiving an audio sample of at least a known test phrase uttered by a speaker associated with the utterance;

transforming the audio sample into a plurality of test samples;

establishing a confidence value for each of the plurality of test samples, wherein the establishing a confidence value includes at least recognizing the plurality of test samples using a plurality of speech recognition modules, each of the plurality of speech recognition modules optimized for a different accent or dialect of a given language;

selecting the accent or dialect of a given language utilized by the speech recognition module which recognized the test sample with a most favorable confidence value for a reference accent; and

storing the reference accent for subsequent decoding.

21. The method of claim 20, wherein the transforming the audio sample into a plurality of test samples comprises:

invoking a processor module for transforming the audio sample into a particular test sample and establishing a confidence value for that particular test sample,

wherein a plurality of processor modules each perform the operations of transforming the audio sample and establishing a confidence value for a different accent or dialect of a given language, each of the plurality of processor modules performing the operations in parallel with the other processor modules.

22. The method of claim 1, wherein decoding the shifted media request at least via iteratively decoding portions of the media request comprises:

parsing at least some of the media request using a first speech recognition process to provide a first result, including at least a determining a context for further recognition of at least a portion of the media request; and

parsing the at least a portion of the media request using at least a second speech recognition process based at least in part upon the determined context to provide a second result.

23. The method of claim 22, wherein parsing the at least a portion of the media request using at least a second speech recognition process based at least in part upon the determined context to provide a second result comprises:

performing the speech recognition process on the at least a portion of the media request iteratively to determine one or more of a further context or a keyword, the performing the speech recognition process iteratively determining at least a second result.

24. The method of claim 22, wherein parsing at least some of the media request using a first speech recognition process to provide a first result, including at least a determining a context for further recognition of at least a portion of the media request comprises:

performing a first speech recognition process on the entire media request, the first speech recognition process determining at least a first result selected from an open dictionary.

25. The method of claim 22, wherein parsing at least some of the media request using a first speech recognition process to provide a first result, including at least a determining a context for further recognition of at least a portion of the media request comprises:

performing a first speech recognition process including determining a first result using Statistical Language Model Automated Speech Recognition (SLM ASR).

26. The method of claim 22, wherein parsing the at least a portion of the media request using at least a second speech recognition process based at least in part upon the determined context comprises:

performing the second speech recognition process including determining a second result selected from a pre-defined list of keywords determined by the context.

27. The method of claim 22, wherein parsing the at least a portion of the media request using at least a second speech recognition process based at least in part upon the determined context to provide a second result comprises:

performing a second speech recognition process including at least Context Free Grammar Automated Speech Recognition (CFG ASR).

28. The method of claim 22, further comprising:

disambiguating at least the first result and the second result, the disambiguating at least the first result and the second result including at least determining one of the first or second results as a favorable result based upon context.

29. The method of claim 1, wherein discovering at least some information using the decoded media request and preparing one or more instructions for consumer device access to the at least some information comprise:

accessing a database, the database configured at least for containing references to information available from a third party source;

identifying a reference to the information; and

preparing at least some active metadata instructions for accessing the referenced information from the third party source, wherein the at least some active metadata instructions include at least;

one or more identities of the referenced information;

one or more locations of the referenced information;

one or more samples of the referenced information; and

at least some access instructions for the referenced information, the access instructions including any of a login, password, and digital rights management code for use with the third party source.

30. The method of claim 1, wherein communicating the prepared instructions to at least one consumer device comprises:

communicating the prepared instructions to a consumer device other than the device at which the utterance was received.

31. A system, comprising:

circuitry configured for receiving an utterance captured via a consumer device;

circuitry configured for providing a media request to at least one server based at least in part on the utterance, including at least;

performing, by the at least one server, one or more of pitch shifting or time shifting of the utterance via distributing a plurality of test samples of the utterance to a plurality of processing devices for testing in parallel, each test sample differently shifted in at least one of pitch or time;

establishing a plurality of confidence values corresponding to the plurality of test samples based at least in part on the testing in parallel; and

selecting one test sample as the shifted media request based at least in part on the plurality of confidence values;

circuitry configured for decoding the shifted media request at least via iteratively decoding portions of the media request;

circuitry configured for discovering at least some information using the decoded media request;

circuitry configured for preparing one or more instructions for consumer device access to the at least some information; and

circuitry configured for communicating the prepared one or more instructions to at least one consumer device.

32. A computer program product comprising one or more non-transitory computer readable media bearing one or more instructions for:

receiving an utterance captured via a consumer device;

providing a media request to at least one server based at least in part on the utterance, including at least;

performing, by the at least one server, one or more of pitch shifting or time shifting of the utterance via distributing a plurality of test samples of the utterance to a plurality of processing devices for testing in parallel, each test sample differently shifted in at least one of pitch or time;

establishing a plurality of confidence values corresponding to the plurality of test samples based at least in part on the testing in parallel; and

selecting one test sample as the shifted media request based at least in part on the plurality of confidence values;

decoding the shifted media request at least via iteratively decoding portions of the media request;

discovering at least some information using the decoded media request;

preparing one or more instructions for consumer device access to the at least some information; and

communicating the prepared one or more instructions to at least one consumer device.