Autonomous humanoid cognitive systems

• US 8,983,889 B1
• Filed: 02/08/2011
• Issued: 03/17/2015
• Est. Priority Date: 03/25/1996
• Status: Expired due to Term

• Current Assignee: Cary D. Stoneman, Dan J. Stoneman
• Original Assignee: Martin L. Stoneman
• Inventors: Stoneman, Martin L.
• Primary Examiner(s): Chaki, Kakali
• Assistant Examiner(s): Tran, Mai T
• Application Number:
  US13/022,839
• Time in Patent Office: 1,498 Days
• Field of Search: 706/46
• US Class Current: 706/46
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06Q 99/00 Subject matter not provided for in other groups of this subclass

Abstract

Disclosed are computer systems with intelligent or autonomous decision systems which include means for determining relevancy, i.e., the threats to and opportunities of the autonomous decision system. Also disclosed are such autonomous decision systems using an efficient ontology to interact sociably with humans, including the use of natural languages and bonding. The desired “whether concrete is included in abstract” computation system is enhanced by the ontology system using categorizing of natural objects using as primitives a set of self tendencies suitable, when hierarchically assigned to objects, to do incremental simulation of “future” situations (including such objects) from a presented situation. Using such primitives and computation system, planning, learning, languaging, etc., are efficiently accomplished.

Claims

1. A computer system for implementing decisions of an autonomous decision system in an environmental situation, comprising:

a) at least one computer processor structured and arranged to process data;

b) at least one computer storage device structured and arranged to store data on a storage medium;

c) at least one input computer processor structured and arranged to provide temporally-incremental input data about a series of said environmental situations; and

d) at least one concrete-situation computer processor structured and arranged to process data regarding said temporally-incremental input data about said series of said environmental situations to provide a temporally-incremental series, respectively, of “

present”

concrete self-situation representations of said respective environmental situations, wherein

i) each said self-situation representation comprises a self representation and a set of event representations, each said event representation being represented specifically spacio-temporally relative to said self representation, and each said event representation including

(1) a behavioral-type designation selected from a set of behavioral-type designations, each said behavioral-type designation of said set of behavioral-type designations being associated with a set of incremental behavioral self-tendencies for determining incrementally-predicted self-situation representations from a said presented self-situation representation; and

(2) a set of current-behavior designations associated with each said event representation specifying the current behaviors of each said event representation; and

e) wherein said system enables implementing decisions of an autonomous decision system in an environmental situation.

2. The computer system according to claim 1 wherein said at least one concrete-situation computer processor comprises at least one representation data computer processor structured and arranged to process data to represent a particular object as part of a collection of object representations, each said object representation of said collection comprising a said behavioral-type designation, and said at least one concrete-situation computer processor further comprising:

a) at least one assignment computer processor structured and arranged to process and store data regarding assigning characteristics to each said object representation of said collection, wherein essentially each of said characteristics comprises a subset of a set of self-tendencies and a corresponding subset of a set of self-tendency markers,

i) said set of self-tendency markers having a 1-to-1 correspondence with said set of self-tendencies, one unique marker from said set of self-tendency markers corresponding respectively with each self-tendency of said set of self-tendencies,

ii) said subset of self-tendencies being constructed and arranged to permit a determination of the expected behavior of each said object representation with respect to any mappable representation of other object representations from said collection of object representations,

iii) each said self-tendency consisting essentially of an instruction for self-behavior (of any first object representation to which said self-tendency may be assigned) on the condition that any mappable representation of object representations from said collection, including said first object representation, from the viewpoint of said first object representation, is included in a specified self-relation selected from a set of self-relations, and

iv) where each self-relation of said set of self-relations comprises a specified space-time relation among a said first object representation and at least one said other object representation, each said other object representation being specified as a subset of a set of self-tendency markers, each said subset of self-tendency markers corresponding to the subset of self-tendencies assigned to each said other object representation.

3. The computer system according to claim 2 wherein:

a) each said self-tendency marker has a first marker part selected from a set of first marker parts and a second marker part selected from a set of second marker parts,

b) said set of first marker parts having a 1-to-1 correspondence with said set of self-relations, and

c) said set of second marker parts having a 1-to-1 correspondence with said set of instructions for self-behavior.

4. A computer system for implementing natural language functions in a humanoid autonomous decision system, comprising:

a) at least one computer processor structured and arranged to process data;

b) at least one computer storage device structured and arranged to store data on a storage medium;

i) wherein said data comprises non-linguistic discrete data-types and, conforming to each of said discrete non-linguistic data-types, a set of non-linguistic discrete data elements; and

c) at least one input computer processor structured and arranged to provide information about current circumstances of the humanoid autonomous decision system;

d) at least one output computer processor structured and arranged to implement decisions of the humanoid autonomous decision system; and

e) at least one relevance computer processor structured and arranged to provide information regarding the relevance to the humanoid autonomous decision system of said current circumstances, comprising

i) at least one self-representation computer processor structured and arranged to process data regarding “

self”

to provide at least one “

self”

representation,

ii) at least one structured-situation computer processor structured and arranged to process data regarding said current circumstances to provide a first non-linguistic structured “

self”

-situation representation,

iii) at least one relational-situation computer storage device structured and arranged to provide data regarding a set of hierarchically-organized, relevant, non-linguistic relational “

self”

-situations, and

iv) at least one inclusional computer processor structured and arranged to process data to determine inclusions of a said first non-linguistic structured “

self”

-situation within said non-linguistic relational “

self”

-situations to determine any relevance of said first structured “

self”

-situation to a said “

self”

of said relevance means,

v) wherein said data regarding said set of hierarchically-organized, relevant, non-linguistic relational “

self”

-situations includes data regarding

(1) a set of hierarchically-organized problem relational “

self”

-situations, and

(2) in association with essentially each of said problem relational “

self”

-situations, a set of hierarchically-organized plan relational “

self”

-situations; and

f) at least one type-linking computer storage device structured and arranged to provide data regarding

i) respectively linking essentially each said discrete data-type of said humanoid autonomous decision system with a respective word/phrase category of a first natural language, and

ii) respectively linking selected words/phrases of each said linked word/phrase category of said first natural language with respective said discrete data elements of each said discrete data-type so linked with a said linked word/phrase category; and

g) at least one data transformation computer processor structured and arranged to process data regarding a first communication to be made by said humanoid autonomous decision system to transform a specified set of non-linguistic data elements into a said first communication in said first natural language, comprising

i) at least one identification computer processor structured and arranged to process data regarding identifying which of said discrete data elements of said discrete data-types is to form part of said first communication,

ii) at least one snippet computer processor structured and arranged to process data regarding selecting natural-language snippets for pointing to the said categories of said natural-language corresponding to whichever of said discrete data-types includes each said discrete data element which is to form part of said first communication,

iii) at least one vocabulary computer processor structured and arranged to process data regarding selecting a word/phrase of said natural-language corresponding to each said discrete data element which is to form part of said first communication, and

iv) at least one grammar computer processor structured and arranged to process data regarding producing from the grammar practices of said natural language and from said snippet selections and from said word/phrase selections said first communication in said natural language.

5. The computer system according to claim 4, further comprising:

a) at least one what-if computer processor comprising

i) at least one trial-decision computer processor structured and arranged to provide data regarding, in association with each of said relevance self-relations, a set of hierarchically-organized “

self”

trial decisions;

ii) at least one selecting computer processor structured and arranged to process data regarding said “

self”

trial decisions to provide data regarding, when a said relevance self-relation has included a said specific circumstance, a current selected said “

self”

trial decision;

iii) at least one trial-decision-testing computer processor structured and arranged to process data regarding using said at least one relevance computer processor to determine the relevance of selected amended structured-situations arising by simulation from using a selected said “

self”

trial decision for the “

self”

object within the then said specific circumstance; and

iv) at least one self-decision-selecting computer processor structured and arranged to process data regarding selecting, depending upon the then specific relevances upon operation of said at least one trial-decision-testing computer processor, a said “

self”

trial decision as a then self-decision of said intelligent system.

6. A computer system for implementing first natural language interpretation functions in a humanoid autonomous decision system interpreting incoming first natural language from another, comprising:

a) at least one computer processor structured and arranged to process data;

b) at least one computer storage device structured and arranged to store data on a storage medium

i) wherein said data comprises non-linguistic discrete data-types and, conforming to each of said discrete non-linguistic data-types, a set of non-linguistic discrete data elements;

c) at least one type-linking computer storage device structured and arranged to provide data regarding

i) respectively linking essentially each said discrete data-type of said humanoid autonomous decision system with a respective word/phrase category of said first natural language, and

ii) respectively linking selected words/phrases of each said linked word/phrase category of said first natural language with respective said discrete data elements of each said discrete data-type so linked with a said linked word/phrase category; and

d) at least one input computer processor structured and arranged to provide input information about characteristics of said incoming natural language sufficient to identify each vocabulary element, snippet type for each said element, and grammatical function for each said element;

e) at least one translation computer processor structured and arranged to process data regarding said input information to provide a non-natural-language concrete circumstance interpretation of said input information; and

f) at least one relevance computer processor structured and arranged to provide information regarding the relevance to the humanoid autonomous decision system of said circumstance interpretation, comprising

i) at least one relational-situation computer storage device structured and arranged to provide data regarding a set of hierarchically-organized, relevant, non-linguistic relational “

self”

-situations, and

ii) at least one inclusional computer processor structured and arranged to process data to determine inclusions of said non-natural-language concrete circumstance interpretation within said non-linguistic relational “

self”

-situations to determine any relevance of said non-natural-language concrete circumstance interpretation a said “

self”

of said at least one relevance computer processor,

iii) wherein said data regarding said set of hierarchically-organized, relevant, non-linguistic relational “

self”

-situations includes data regarding

(1) a set of hierarchically-organized problem relational “

self”

-situations, and

(2) in association with essentially each of said problem relational “

self”

-situations, a set of hierarchically-organized plan relational “

self”

-situations.

7. The computer system according to claim 6 wherein said interpreting humanoid autonomous decision system has abilities to select for use in interpretation similar cognitive, relevancy, and emotion systems to those of the other.

8. A computer system for machine learning from an environmental situation in an autonomous decision system, comprising:

a) at least one computer processor structured and arranged to process data;

b) at least one computer storage device structured and arranged to store data on a storage medium;

c) at least one input computer processor structured and arranged to provide temporally-incremental input data about a series of said environmental situations; and

d) at least one concrete-situation computer processor structured and arranged to process data regarding said temporally-incremental input data about said series of said environmental situations to provide a temporally-incremental series, respectively, of “

present”

concrete self-situation representations of said respective environmental situations, wherein

i) each said self-situation representation comprises a self representation and a set of event representations, each said event representation being represented specifically spacio-temporally relative to said self representation, and each said event representation including

(1) a behavioral-type designation selected from a set of behavioral-type designations, each said behavioral-type designation of said set of behavioral-type designations being associated with a set of incremental behavioral self-tendencies for determining incrementally-predicted self-situation representations from a said presented self-situation representation; and

(2) a set of current-behavior designations associated with each said event representation specifying the current behaviors of each said event representation; and

e) wherein said system provides machine learning from an environmental situation in an autonomous decision system.

9. The computer system according to claim 8 further comprising:

a) at least one predicted-situation computer processor structured and arranged to process data regarding a first of a said series of “

present”

concrete self-situation representations of said respective environmental situations to provide an incremental simulation of a set of “

predicted”

concrete self-situation representations;

b) at least one comparing computer processor structured and arranged to process data regarding, for a corresponding particular time, a said “

predicted”

concrete self-situation representation with a said corresponding “

present”

concrete self-situation representation to identify any selected minimum event differences between said compared concrete representations;

c) at least one difference-storing computer storage device structured and arranged to store data regarding

i) said concrete self-situation representations having at least said identified minimum event difference and

ii) said particular “

present”

concrete self-situation representation from which said identified “

predicted”

concrete self-situation representation was computed;

d) at least one choosing computer processor structured and arranged to process data regarding choosing a modification in a said incremental behavioral self-tendency likely to decrease said event difference between said compared representations;

e) at least one testing computer processor structured and arranged to process data regarding, for said corresponding particular time, a said “

predicted”

concrete self-situation representation and a said corresponding “

present”

concrete self-situation representation to provide a new identified “

predicted”

concrete self-situation representation using said modification chosen for said incremental behavioral self-tendency;

f) at least one new-prediction computer processor structured and arranged to process data regarding comparing said new identified “

predicted”

concrete self-situation representation with said corresponding former “

predicted”

concrete self-situation representation to identify any remaining said event differences; and

g) at least one heuristic computer processor structured and arranged to process data regarding said modification and said event differences to heuristically choose further processing of data among at least the following

i) test a further modification of a said incremental behavioral self-tendency, or

ii) select for storage a modified incremental behavioral tendency, or

iii) combine for storage a said modified incremental behavioral tendency with a corresponding original incremental behavioral tendency as a replacement incremental behavioral tendency.

10. A computer system for an entertainment system, comprising:

a) at least one computer simulation processor structured and arranged to process data regarding a user-selected “

world”

-representation containing user-selected spacio-temporally located “

object”

-representations to provide incremental simulation-stepping of said “

world”

-representation;

b) at least one natural-language computer interface structured and arranged to process data for providing a natural-language interface for user selection of non-natural-language characteristics of a said “

object”

-representation;

c) at least one placement computer interface structured and arranged to process data for providing an interface for user placement of a said “

object”

-representation into said “

world”

-representation;

d) at least one computer interface structured and arranged to process data for providing an interface for placement into said “

world”

-representation, as a said “

object”

-representation, of an autonomous decision system;

e) wherein said autonomous decision system comprises

i) at least one representation computer processor structured and arranged to process data essentially from said “

world”

-representation for presenting a selected self-situation (for said autonomous decision system) representation, said presented self-situation representation comprising a self representation and a set of event representations, each said event representation being represented specifically spacio-temporally relative to said self representation, and

ii) at least one prediction computer processor structured and arranged to process data for determining the representations of a set of incrementally-predicted self-situations, predicted as incremental consequences from said presented self-situation representation;

iii) wherein said at least one prediction computer processor comprises said computer simulation processor.

11. A machine computational-processing method, for providing current simulated-emotion expression in at least one simulated-humanoid autonomous decision system having at least one ability to assess a set of environmental circumstances, comprising the steps of:

a) storing in such at least one simulated-humanoid autonomous decision system planning data providing plan capability to such at least one simulated-humanoid autonomous decision system;

b) using information about such set of environmental circumstances of such at least one simulated-humanoid autonomous decision system and such plan capability, computing at least one current planning selection;

c) using information about such at least one current planning selection, computing at least one current planning status;

d) using information about such at least one current planning status, computing current simulated-emotion source data; and

e) using such current simulated-emotion-source data, computing current simulated-emotion status.

12. The machine computational-processing method according to claim 11 further comprising the step of using information about such current simulated-emotion status, computationally providing output signals to operate effectors to provide current simulated-emotion expression of such at least one simulated-humanoid autonomous decision system.

13. A machine computational-processing method, for implementing decisions of at least one autonomous decision system in environmental situations, such at least one autonomous decision system having at least one input system for providing temporally-incremental input data about a series of such environmental situations, comprising the steps of:

a) computationally processing such temporally-incremental input data about such series of such environmental situations to provide a temporally-incremental series, respectively, of “

present”

concrete self-situation representations of such respective environmental situations, wherein

i) each such “

present”

concrete self-situation representation comprises a self representation and a set of event representations, each such event representation being represented spacio-temporally relative to each such self representation, wherein each such event representation comprises

(1) at least one behavioral-type designation selected from a set of behavioral-type designations, each such at least one behavioral-type designation of such set of behavioral-type designations being associated with a set of incremental behavioral self-tendencies for determining incrementally-predicted self-situation representations from each such “

present”

concrete self-situation representation, and

(2) a set of current-behavior designations associated with each such event representation specifying the current behaviors of each such event representation;

b) computationally processing data regarding at least one such “

present”

concrete self-situation representation of such respective environmental situation to determine the representations of a set of incrementally-predicted self-situations, predicted as incremental consequences from such at least one “

present”

concrete self-situation representation;

c) storing data for hierarchical planning comprising

i) a hierarchical set of n problem representations, and

ii) m plan-sets of hierarchical subgoal representations, each said plan-set of hierarchical subgoal representations being associated with at least one of said set of n problem representations; and

d) computationally comparing such data for hierarchical planning with such data about each such “

present”

concrete self-situation representation and each such incrementally-predicted self-situation to determine

i) at least one self-relevancy of each such presented self-situation representation, and

ii) at least one self-relevancy of each such incrementally-predicted self-situation representation;

e) wherein any threat to and any opportunity of such at least one autonomous decision system may be determined.

14. The machine computational-processing method according to claim 13 further comprising the step of, from such information determined about any such threat to and any such opportunity of such at least one autonomous decision system, computationally determining at least one decision of such autonomous decision system.

15. The machine computational-processing method according to claim 13 wherein such computational comparing step comprises:

a) computationally determining which of such hierarchical set of n problems includes at least one such “

present”

concrete self-situation representation;

b) computationally determining which of such hierarchical subgoal representations includes at least one such “

present”

concrete self-situation representation; and

c) computationally determining which of such problem representations and which of such subgoal representations include at least one such incrementally-predicted self-situation representation.

16. The machine computational-processing method according to claim 13 further comprising the steps of:

a) computationally specifying at least one self trial decision for use in what-iffing; and

b) computationally determining, using such at least one self trial decision, at least one set of such incrementally predicted self-situation representations.

17. A machine computational-processing method, for implementing natural language functions in at least one simulated humanoid autonomous decision system, comprising the steps of:

a) storing in at least one computational storage system data comprising non-linguistic discrete data-types and, conforming to each of such discrete non-linguistic data-types, a set of non-linguistic discrete data elements;

b) storing in at least one computational storage system data

i) respectively linking essentially each such discrete data-type of such simulated-humanoid autonomous decision system with a respective word/phrase category of at least one first natural language, and

ii) respectively linking selected words/phrases of each such linked word/phrase category of such at least one first natural language with respective such discrete data elements of each such discrete data-type so linked with a such linked word/phrase category;

c) using information about at least one set of current circumstances of such at least one simulated-humanoid autonomous decision system, computationally determining at least one relevance to the simulated-humanoid autonomous decision system of such current circumstances;

d) using information about such at least one relevance, computationally specifying at least one set of relevant such non-linguistic discrete data elements; and

e) using such specification of such at least one set of relevant such non-linguistic discrete data elements, computationally determining at least one first communication to be made by such simulated-humanoid autonomous decision system to transform such specified set of non-linguistic discrete data elements into such at least one first communication in such first natural language;

f) wherein such step of computational determining comprises

i) computationally processing data regarding identifying which of such discrete data elements of such discrete data-types is to form part of such at least one first communication,

ii) computationally processing data regarding selecting natural-language snippets for pointing to the such categories of such natural-language corresponding to whichever of such discrete data-types includes each such discrete data element which is to form part of such at least one first communication,

iii) computationally processing data regarding selecting a word/phrase of such natural-language corresponding to each such discrete data element which is to form part of such at least one first communication, and

iv) computationally processing data regarding producing from grammar practices of such natural language and from such snippet selections and from such word/phrase selections such at least one first communication in such natural language.

18. A machine computational-processing method for implementing first natural language interpretation functions in at least one simulated-humanoid autonomous decision system interpreting incoming first natural language from at least one other, comprising the steps of:

a) storing in at least one computational storage system data comprising non-linguistic discrete data-types and, conforming to each of such discrete non-linguistic data-types, a set of non-linguistic discrete data elements;

b) storing in at least one computational storage system data

i) respectively linking essentially each such discrete data-type of such simulated-humanoid autonomous decision system with a respective word/phrase category of at least one first natural language, and

ii) respectively linking selected words/phrases of each such linked word/phrase category of such at least one first natural language with respective such discrete data elements of each such discrete data-type so linked with a such linked word/phrase category;

c) computationally processing such incoming natural language sufficiently to provided data identifying each vocabulary element, snippet type for each such element, and grammatical function for each such element;

d) computationally processing such identifying data to provide a non-natural-language concrete circumstance interpretation of such incoming natural language; and

e) computationally determining at least one relevance to such at least one simulated-humanoid autonomous decision system of such circumstance interpretation;

f) wherein such computational determining comprises

i) storing data regarding a set of hierarchically-organized, relevant, non-linguistic relational “

self”

-situations, and

ii) computationally processing to determine inclusions of such non-natural-language concrete circumstance interpretation within such non-linguistic relational “

self”

-situations to determine any relevance of such non-natural-language concrete circumstance interpretation to such at least one simulated-humanoid autonomous decision system,

iii) wherein such data regarding such set of hierarchically-organized, relevant, non-linguistic relational “

self”

-situations includes data regarding

(1) a set of hierarchically-organized problem relational “

self”

-situations, and

(2) in association with essentially each of said problem relational “

self”

-situations, a set of hierarchically-organized plan relational “

self”

-situations.

19. The machine computational-processing method according to claim 18 wherein such interpreting simulated-humanoid autonomous decision system possesses abilities to select for use in interpretation similar cognitive, relevancy, and emotion systems to those of the at least one other.

20. The machine computational-processing method according to claim 18 wherein such step of computational processing of such incoming natural language comprises natural-language default-selecting to process data regarding selection of non-natural-language data types and data for correspondence with such incoming information.

21. The machine computational-processing method according to claim 18, further comprising the steps of:

a) computationally processing data regarding at least one story-series of such incoming information to provide at least one story-series of such non-natural-language concrete circumstance interpretation; and

b) computationally processing data regarding such at least one story-series of such non-natural-language concrete circumstance interpretation to provide at least one learned modification of at least one such nonlinguistic discrete data element,

c) wherein such story-series of such non-natural-language concrete circumstance interpretation is treated as at least one temporally-incremental series, respectively, of “

present”

concrete self-situation representations of at least one temporally-incremental series of respective environmental situations.