Method and system for robot understanding, knowledge, conversation, volition, planning, and actuation

• US 20140006320A1
• Filed: 06/18/2013
• Published: 01/02/2014
• Est. Priority Date: 06/27/2012
• Status: Active Grant

• Current Assignee: Linker Systems, Incorporated
• Original Assignee: Linker Systems, Incorporated
• Inventors: Linker, Sheldon O., Miller, Michael S. P., Poper, Toni M.
• Granted Patent: US 9,710,763 B2
• Time in Patent Office: Days
• US Class Current: 706/12
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: Y10S 901/03 Teaching system

Abstract

A system for understanding and storing knowledge and plans in a speakable, formal class of languages, allowing querying and retrieving of that knowledge and plan (conversation), and of deciding and acting (volition) upon such plans is provided. The system further allows the users to input data, instructions, questions, queries, imperatives, and plans without structure or rules over and above the actual grammar of the class of languages. The system responds in words, phrases, complete sentences, or actions, as applicable, and its output is bound by the rules of logic, rather than correlation or likely meaning.

Claims

1. A method operational in a processing device, comprising:

inputting data into the processing device in a language, which may be an artificial language or a subset of a natural language, characterized by having a completely defined set of nonambiguous grammar rules;

using an action-capable prover to compare the input data with stored data in a data/knowledge/plan base according to logic operations to determine if the meaning of the parsed input data is executable;

storing results of the evaluation in the data/knowledge/plan base, wherein the input data is obtained from a plurality of different users and/or sensors and the data/knowledge/plan base stores knowledge and instructions for different users and robotic bodies or other actuator systems; and

in reply to input, executing or refusing to execute one or more instructions and/or plans, to actuate one or more robotic body component(s) or actuator(s), and/or otherwise take action, physical and/or simulated.

2. The method of claim 1, wherein the processing device maintains and stores distinct realities for different users and robotic bodies or other actuator systems.

3. The method of claim 1, as limited by claim 2, wherein the processing device acts or decides not to act based on those distinct realities.

4. The method of claim 1, further comprising comparing the use of a logic prover to determine whether an instruction or imperative is ready for execution.

5. The method of claim 1, wherein the language includes a grammar structure defined by:

a characteristic to form imperatives and/or to form sentences, instructions or statements which imply imperatives.

6. The method of claim 1, further comprising retaining the input data if the input data is an imperative or plan which provides new information or plan to the data/knowledge/plan base and adding the imperative or plan to the data/knowledge/plan base.

7. The method of claim 1, wherein executable portions of the plan are executed, in a real or simulated manner.

8. The method of claim 1, as limited by claim 7, wherein a memory of executions made is retained.

9. The method of claim 1, as limited by claim 7, wherein a memory of the reasoning behind execution is maintained.

10. The method of claim 1, wherein all forms of knowledge and planning are stored in the data/knowledge base as predications and/or grammar trees.

11. The method of claim 1, wherein generating the response in a language includes utilizing or causing real or simulated physical action.

12. The method of claim 1, wherein generating the response in a language, time-sensitive context of knowledge and/or plans in the data/knowledge/plan base is considered in formulating the response or action.

13. The method of claim 1, wherein distinct realities for different users allows for the inclusion of contradictory conclusions, results, and/or actions for different users.

14. The method of claim 1, wherein distinct realities for different bodies, real or simulated, allow for the inclusion of contradictory conclusions, results, and/or actions for different bodies, actuators, and/or other devices, real or simulated.

15. The method of claim 1, wherein the data/knowledge/plan base allows for concurrent storage of contradictory knowledge and/or plans for different users and/or robot bodies, actuators, and/or other devices, real or simulated.

16. The mechanism of claim 1 which needs no programming specific to actions to be taken or avoided. (This claim is based on the fact that nothing in the teachings present or referenced describes action-specific programming.)

17. A device for processing a language, which may be an artificial language or a subset of a natural language, selected by a user, comprising:

an input interface configured to receive input data in a language characterized by having a completely defined set of unambiguous grammar rules;

a processing circuit adapted to implement a parsing module adapted to parse the input data from the input module into an internal format;

a prover module adapted to receive the formatted input data from the parse module and comparing the parsed input data with stored data and/or plans in a data/knowledge/plan base to determine if the parsed input data is actionable; and

a memory device configured to store results of the prover module in the data/knowledge/plan base, wherein the input data is obtained from a plurality of different users and/or sensors and the data/knowledge/plan base stores plans;

wherein in execution of an instruction and/or plan, to actuate one or more robotic body component(s) or actuator(s), and/or otherwise take action, physical and/or simulated.

18. The device of claim 17 in which the device recognizes, maintains and stores distinct realities for different users and/or bodies.

19. The device of claim 17, which further acts on an actionable plan item, using distinct realities for one or more bodies, actuators, and/or other devices, real or simulated, each acting alone or in consort.

20. A method operational in a processing device, comprising:

inputting data into the processing device in a language, which may be an artificial language or a subset of a natural language, or in a pre-parsed version thereof, wherein the language is characterized by having a completely defined set of unambiguous grammar rules;

using a prover or determine if the meaning of the parsed input data is actionable;

storing results of the evaluation in the data/knowledge/plan base, wherein the input data is obtained from a plurality of different users and the data/knowledge/plan base stores plans; and

as a consequence, acts only when appropriate, based on accumulated knowledge of the data/knowledge/plan base, wherein the auditory or visual response to a first user includes contradictory realities from different users if first user does not have a stored reality with which to provide the response.

21. A processing device of claim 20 which recognizing, maintaining, and storing distinct realities for different users and bodies.

22. A processing device of claim 20, as limited by claim 21, which acts only when appropriate, based on contradictory realities from different users if first user does not have a stored reality with which to provide the response.

23. The mechanism of claim 20 which needs no programming specific to actions to be taken or avoided. (This claim is based on the fact that nothing in the teachings present or referenced describes action-specific programming.)

24. A method operational in a processing device, comprising:

inputting data into the processing device in a language, which may be an artificial language or a subset of a natural language, characterized by having a completely defined set of nonambiguous grammar rules (which excludes complete natural languages);

using a prover to determine if the input is actionable;

storing results of the evaluation in the data/knowledge/plan base.

25. The processing device of claim 24, wherein the input data is obtained from a plurality of different users and the data/knowledge/plan base stores knowledge and plans while recognizing, maintaining, and acting upon that plan.

26. The mechanism of claim 24 which needs no programming specific to actions to be taken or avoided. (This claim is based on the fact that nothing in the teachings present or referenced describes action-specific programming.)

27. A mechanism providing for robotic volition and/or planning, in which volition and/or planning are provided by a theorem prover.

28. The mechanism of claim 27 in which the robotic system consists of or actuates through one or more robotic bodies and/or actuators, real or simulated.

29. The mechanism of claim 27 in which said theorem prover operates on built-in, stated, observed, and/or derived facts, theorems, and/or rules.

30. The mechanism of claim 27 in which logic states include but are not limited to Imperative, Negative Imperative, Plan, and/or Restriction (a negative Plan).

31. The mechanism of claim 27, as limited by claim 30 in which truth, falsehood, plausibility, and other logic states depend on point of view, in terms of one or more users, one or more referenced entities, and/or robot bodies and/or actuators.

32. The mechanism of claim 27 which needs no programming specific to actions to be taken or avoided. (This claim is based on the fact that nothing in the teachings present or referenced describes action-specific programming.)

33. The mechanism of claim 27 in which a theorem prover proves that one or more actions should be taken at any given time.

34. The mechanism of claim 27 in which a theorem prover proves that one or more actions should not be taken at any given time.

35. A mechanism to accomplish planning which requires no specific programming.

36. The mechanism of claim 35 in which robotics, simulated robotics, or other control and/or actuation systems are so controlled and/or planned.

37. A method operational in a processing device, comprising:

A theorem prover which accepts imperatives, as well as statements of fact, axioms, and/or theorems.

38. The method of claim 37 in which there are negative imperatives, which prevent other imperatives from being deemed as true.

39. The method of claim 37 in which there are as-yet unproven imperatives, stored as a plan.

40. A device comprising:

A theorem prover which accepts imperatives, as well as statements of fact, axioms, and/or theorems.

41. The device of claim 40 in which there are negative imperatives, which prevent other imperatives from being deemed as true.

42. The device of claim 40 in which there are as-yet unproven imperatives, stored as a plan.

43. A mechanism providing for a theorem prover which accepts imperatives, as well as statements of fact, axioms, and/or theorems.

44. The mechanism of claim 43 in which there are negative imperatives, which prevent other imperatives from being deemed as true.

45. The mechanism of claim 43 in which there are as-yet unproven imperatives, stored as a plan.

46. A device comprising:

A planning system which uses a theorem prover to make decisions.

47. The device of claim 46 in which negative plans prevent other plan elements from being scheduled.

48. A mechanism providing for a planning system which uses a theorem prover as its decision maker.

49. The mechanism of claim 48, in which there are negative plans, which prevent other plan elements from being scheduled.