Human-like response emulator

• US 9,355,092 B2
• Filed: 02/01/2007
• Issued: 05/31/2016
• Est. Priority Date: 02/01/2006
• Status: Active Grant

• Current Assignee: Icommand Ltd.
• Original Assignee: I-Command Limited
• Inventors: Teague, Paul, Sastriawan, Joko
• Primary Examiner(s): Shah, Paras D
• Application Number:
  US12/162,985
• Publication Number:
  US 20090234639A1
• Time in Patent Office: 3,407 Days
• Field of Search: 704/9, 704/275, 704/277, 704/270, 704/231, 704/246, 704/251, 704/257, 704/258
• US Class Current: 1/1
• CPC Class Codes: B60R 16/0373 Voice control in general G10L
• CPC Class Codes: G06F 40/56 Natural language generation
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G06Q 90/00 Systems or methods specially adapted for administrative, commercial, financial, managerial or supervisory purposes, not involving significant data processing
• CPC Class Codes: G09B 19/04 Speaking with audible presentation of the material to be studied G09B5/04
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: H04M 1/271 controlled by voice recognition

Abstract

Human-like response emulator stores a library (14) comprising one or more different subject matter data structures. Each data structure comprising a set of stimuli related to the subject matter of the data structure and one or more output instructions associated with each stimulus. Each output instruction is for producing a human-like response to the associated stimulus. The emulator receives a stimulus (16, 18). The emulator looks up output instructions in each data structure that are associated with the received stimulus. The emulator outputs (18, 20) one or more responses to the received stimulus according to the found output instructions

Claims

1. A method of emulating human-like responses, the method comprising:

storing a library comprising one or more different subject matter data structures, each data structure comprising a plurality of output instructions related to the subject matter of the data structure, each output instruction producing a human-like response and being associated with a received input stimulus, wherein the received input stimuli comprise human inputs and system inputs, the system inputs received via at least one sensor;

associating each event with a tag;

using the tag to determine whether an event corresponds to an important event or a non-important event;

looking up output instructions in each data structure that are associated with the received input stimulus;

outputting, via a human output API, one or more responses to the received stimulus according to a found output instruction when the event corresponds to an important event, wherein the one or more responses are ordered according to a priority rating;

placing non-important events in an event queue;

applying logical rules to the non-important events to determine whether a plurality of non-important events are collectively indicative of an important event; and

outputting, via a human output API, one or more responses when the logical rules determine that the plurality of non-important events are collectively indicative of an important event, wherein the one or more responses are ordered according to a priority rating;

wherein the one or more different subject matter data structures are arranged such that the output instructions which produce the human-like response are grouped hierarchically according to their respective associated stimuli.

2. A method as claimed in claim 1, wherein the method comprises, in response to the looking up output instructions in each data structure that are associated with the received input stimulus, combining the output instructions found, and outputting a produced response to the received input stimulus.

3. A method as claimed in claim 1, wherein the output instructions produce the response in the form of one or more of the following:

language, an animated graphical display, voice characteristics, and/or control signals.

4. A method as claimed in claim 3, wherein where the response is in the form of language, the response is in the form of one or more sentences constructed of one or more sentence segments, wherein each sentence segment conveys a coherent meaningful piece of information.

5. A method as claimed in claim 4, wherein the one or more sentence segments are constructed of one or more words that have a relevant meaning in response to the associated stimulus.

6. A method as claimed in claim 5, wherein the one or more sentence segments are human authored and are of the same type of response that a human would make if the human received the input stimulus.

7. A method as claimed in claim 4, wherein each of the sentence segments is associated with a sentence position category, which indicates where within a complete sentence the sentence segment is permitted to be located.

8. A method as claimed in claim 7, wherein sentences are constructed from the associated sentence segments according to their permitted location with each sentence according to the associated sentence position category/categories.

9. A method as claimed in claim 4, wherein each output instruction is associated with one sentence segment or a scripted sentence.

10. A method as claimed in claim 1, wherein each data structure is arranged in a plurality of levels, where output instructions are grouped in the same level according to contextually related associated input stimuli.

11. A method as claimed in claim 10, wherein matching of an input stimulus to an associated output instruction is initially limited to looking within one level of the data structure and will only search outside the level if a match is not found.

12. A method as claimed in claim 11, wherein a previous level within which a match was found is recorded which initiates the next search for a match within the same level.

13. A method as claimed in claim 11, wherein a previous level within which a match was found is recorded which initiates the next search for a match within the next level down relative to the previous level.

14. A method as claimed in claim 11, wherein in the event a match is not found within the current level, searching will proceed to parallel levels and then up through the levels of the data structure.

15. A method as claimed in claim 1, wherein searching for matches of the received input stimulus to an associated output instruction will progress through the hierarchy if a match is not found.

16. A method as claimed in claim 11, wherein a history of matches is recorded, such that in the event that a match is not found in a level, and there is a choice of which level to search next, the history of matches is used to choose which level to search next.

17. A method as claimed in claim 1, wherein the stimulus is in the form of a human generated stimulus, an environmental stimulus, or a machine generated stimulus.

18. A method as claimed in claim 1, wherein the received input stimulus is in the form of speech.

19. A method as claimed in claim 4, wherein alternative sentence segments having different form but having the same meaning are stored in the data structure and when a match occurs one of the alternatives is chosen in a random manner for output.

20. A method as claimed in claim 1, wherein when no output instructions are found that are associated to the received input stimulus in the data structure, a new received input stimulus is recorded and a new output instruction for producing a response is input in a data structure, the response being taught by a human by providing instructions for producing the response.

21. A method as claimed in claim 20, wherein when the output instruction associated with the new received input stimulus is added to the data structure, a level within the hierarchy is allocated and recorded.

22. A method as claimed in claim 1, wherein the received input stimulus is pre-processed prior to look-up of a match.

23. A method as claimed in claim 22, wherein the pre-processing involves applying mathematical and/or logical operations to the received input stimulus.

24. A method as claimed in claim 1, wherein the library comprises selected data structures such that the human-like responses produced reflect a personality, in that the human-like responses are those which the personality would make.

25. A method as claimed in claim 1, wherein the library comprises selected response instructions that are instructional to a human observer.

26. A method as claimed in claim 1, wherein the library comprises selected response instructions that are constructed to document events reflected by the input stimulus.

27. A human-like response emulator comprising:

a library stored on a storage device, the library comprising one or more different subject matter data structures, each data structure comprising a plurality of output instructions related to the subject matter of the data structure, each output instruction producing a human-like response and being associated with a received input stimulus, wherein the received input stimuli comprise human inputs and system inputs;

an input device for receiving the input stimulus including at least one sensor to receive the system inputs;

a look-up device for looking up output instructions in each data structure that are associated with the received input stimulus; and

an output device for delivering one or more responses to the received stimuli according to the output instructions found by the look-up device;

the emulator associating each event with a tag, using the tag to determine whether an event corresponds to an important event or a non-important event, and placing non-important events in an event queue;

the output device outputting, via a human output API, one or more responses to the received stimulus according to a found output instruction when the event corresponds to an important event, wherein the one or more responses are ordered according to a priority rating;

the emulator applying logical rules to the non-important events to determine whether a plurality of non-important events are collectively indicative of an important event; and

the output device outputting, via a human output API, one or more responses when the logical rules determine that the plurality of non-important events are collectively indicative of an important event, wherein the one or more responses are ordered according to a priority rating;

wherein the one or more different subject matter data structures are arranged such that the output instructions which produce the human-like response are grouped hierarchically according to their respective associated stimuli.

28. An apparatus for emulating human-like responses, the apparatus comprising:

a storage for storing a library comprising one or more different subject matter data structures, each data structure comprising a plurality of output instructions related to the subject matter of the data structure, each output instruction producing a human-like response and being associated with an input stimulus, wherein the received input stimuli comprise human inputs and system inputs;

a receiver for receiving an input stimulus including at least one sensor to receive the system inputs;

a processor for looking up output instructions in each data structure that are associated with the received input stimulus;

an output for outputting, via a human output API, one or more responses to the received stimulus according to found output instructions;

the apparatus associating each event with a tag, using the tag to determine whether an output instruction corresponds to an important event or a non-important event, and placing non-important events in an event queue;

the output outputting one or more responses to the received stimulus according to a found output instruction when the event corresponds to an important event, wherein the one or more responses are ordered according to a priority rating;

the apparatus applying logical rules to the non-important events to determine whether a plurality of non-important events are collectively indicative of an important event; and

the output outputting, via a human output API, one or more responses when the logical rules determine that the plurality of non-important events are collectively indicative of an important event, wherein the one or more responses are ordered according to a priority rating;

wherein the one or more different subject matter data structures are arranged such that the output instructions are grouped hierarchically according to their respective associated stimuli.