IP Library › Granted Patent US 12,314,834

Granted Patent B1

US 12,314,834 · App. 18/810,460 · Granted May 27, 2025

Iterative attention-based neural network training and processing

Inventors: Steven Dennis Flinn (Sugar Land, TX); Naomi Felina Moneypenny (Bellevue, WA)

Assignee: Steven D. Flinn

G06N3/043 G06F40/211 G06F40/216 G06F40/30 G06N3/045 G06N5/048

View Patent ↗

Loading inventors, assignments & file history…

Monitor This Case

Get email alerts when status or documents change.

Order Certified Copies

Most orders are placed with the USPTO same day — all within 24 business hours.

Order via The Patent Place →

Pre-filled with this patent's details

Quick Facts

Patent No.: US 12,314,834
App. No.: 18/810,460
Granted: May 27, 2025
Kind: B1

External Links

USPTO Patent Center↗ Full Patent PDF↗

Abstract

An iterative attention-based neural network training and processing method and system iteratively applies a focus of attention of a trained neural network on syntactical elements and generates probabilities associated with representations of the syntactical elements, which in turn inform a subsequent focus of attention of the neural network, resulting in updated probabilities. The updated probabilities are then applied to generate syntactical elements for delivery to a user. The user may respond to the delivered syntactical elements, providing additional training information to the trained neural network.

Claims (525)

1. A system, comprising:

one or more processors;

one or more memories in communication with the one or more processors; and

one or more programs, wherein the one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, the one or more programs including instructions for causing:

access to a first neural network that is trained by automatic internal application of a first plurality of attentions to representations of different subsets of a first plurality of items of content;

generation, by automatic internal application of a second plurality of attentions and utilizing the trained first neural network, of a second plurality of items of content;

training of a second neural network by automatic internal application of a third plurality of attentions to representations of different subsets of the second plurality of items of content;

identification of a plurality of syntactical elements;

generation, by automatic internal application of a fourth plurality of attentions to representations of different subsets of the plurality of syntactical elements and utilizing the trained second neural network, of a resulting item of content; and

the resulting item of content to be sent to a user.

2. The system of claim 1 , wherein the system is configured such that the first neural network and the second neural network are components of the system.

3. The system of claim 1 , wherein the system is configured such that at least a portion of the second plurality of items of content is automatically input into the second neural network for training of the second neural network, in response to being generated.

4. The system of claim 1 , wherein the system is configured such that the second plurality of items of content include both syntactical elements and images.

5. The system of claim 1 , wherein the system is configured such that the plurality of syntactical elements is identified in response to being received from the user.

6. The system of claim 1 , wherein the system is configured such the trained first neural network is a component of the system, and the second neural network is external to the system.

7. The system of claim 1 , wherein the system is configured such that the second plurality of items of content is generated by the trained first neural network based on a composition that results in a chain including linked representations of each of a plurality of semantic chains.

8. The system of claim 1 , wherein the system is configured such that the second plurality of items of content is generated by the trained first neural network based on operations including:

one or more searches in at least one database;

a composition that results in a chain including linked representations of each of a plurality of semantic chains; and

generating an explanation for an application of the trained first neural network.

9. The system of claim 8 , wherein the system is configured such that one or more of the operations is iteratively performed by application of the trained first neural network based on different pluralities of items of content, until an expected net information value, that is calculated based on an expected value and an expected cost and that indicates an expected affect on an output from the trained first neural network, is insufficient.

10. The system of claim 1 , wherein the system is configured such that the first plurality of attentions are prioritized.

11. The system of claim 10 , wherein the system is configured such that the first plurality of attentions are prioritized utilizing a single data structure representing all of the first plurality of items of content and a position aspect as well as a non-position aspect thereof.

12. The system of claim 10 , wherein the system is configured such that the first plurality of attentions are prioritized utilizing a single data structure representing all of the first plurality of items of content that constitute an entirety of a syntactical element portion of a training prompt.

13. The system of claim 10 , wherein the system is configured such that the first plurality of attentions are prioritized utilizing a single matrix representing all of the first plurality of items of content.

14. The system of claim 10 , wherein the system is configured such that the first plurality of attentions are prioritized, before any usage of the prioritization of any of the first plurality of attentions.

15. The system of claim 1 , wherein the system is configured such that the first plurality of attentions are prioritized in a single operation.

16. The system of claim 10 , wherein the system is configured such that the first plurality of attentions are prioritized by an application of a first process that prioritizes the different subsets of the first plurality of items of content based on a first relational aspect among the different subsets of the first plurality of items of content and an application of a second process that prioritizes the different subsets of the first plurality of items of content based on a second relational aspect among the different subsets of first plurality of items of content.

17. The system of claim 16 , wherein the system is configured such that the application of the first process and the application of the second process of the first prioritization are completed before any utilization of the prioritization of any of the first plurality of attentions via the first neural network.

18. The system of claim 10 , wherein the system is configured such that the first plurality of attentions are prioritized to generate a result based on a relative importance or relevance of a plurality of relationships among the first plurality of items of content, and then another plurality of attentions are prioritized based on the result.

19. The system of claim 10 , wherein the system is configured such that the first plurality of attentions are prioritized utilizing first weights that reflect a relative importance or relevance of a plurality of relationships among the first plurality of items of content.

20. The system of claim 19 , wherein the system is configured such that an additional plurality of attentions are prioritized utilizing second weights that reflect the relative importance or relevance of the plurality of relationships among the first plurality of items of content, as updated based on probabilities generated as a result of the prioritization of the first plurality of attentions.

21. The system of claim 10 , wherein the system is configured such that the prioritization results in only a subset of the first plurality of attentions corresponding to only a subset of the different subsets of the first plurality of items of content being a basis for the training of the first neural network.

22. The system of claim 1 , wherein the system is configured such that the first neural network is trained on a training set including: images associated with corresponding syntactical elements to infer one or more syntactical elements corresponding to pixel patterns within the images, and training syntactical elements, including training natural language syntactical elements and training computer-executable code syntactical elements, that are utilized by computing hardware that processes the training syntactical elements to direct training attentions to representations of different subsets of the training syntactical elements.

23. The system of claim 1 , wherein the one or more programs further include instructions for causing:

a determination whether a first scenario described by the resulting item of content represents objective reality; and

based on a determination that the first scenario does not represent objective reality, generation, by applying the trained second neural network utilizing the resulting item of content, of another resulting item of content that describes a second scenario that represents objective reality.

24. The system of claim 1 , wherein the system is configured such that one or more user preferences is identified, where the resulting item of content is caused to be generated, based on the one or more user preferences.

25. The system of claim 24 , wherein the system is configured such that the one or more user preferences is based on at least one aspect of the plurality of syntactical elements and at least one previous prompt that are received from the user that includes a human user.

26. The system of claim 1 , wherein the system includes:

one or more servers including a first memory of the one or more memories with a first program of the one or more programs stored therein that, when a first portion of the instructions of the first program is executed by at least one first processor of the one or more processors of the one or more servers of the system, cause the one or more servers of the system to perform the training of the first neural network and the second neural network; and

a mobile device including a second memory of the one or more memories with a second program of the one or more programs stored therein that, when a second portion of the instructions of the second program is executed by at least one second processor of the one or more processors of the mobile device of the system, cause the mobile device of the system to generate the resulting item of content.

27. A computer-implemented method, comprising:

within a system:

causing access to a first neural network that is trained by automatic internal application of a first plurality of attentions to representations of different subsets of a first plurality of items of content, and that generates, by automatic internal application of a second plurality of attentions and utilizing the trained first neural network, a second plurality of items of content;

causing training of a second neural network by automatic internal application of a third plurality of attentions to representations of different subsets of the second plurality of items of content;

causing identification of a plurality of syntactical elements;

causing generation, by automatic internal application of a fourth plurality of attentions to representations of different subsets of the plurality of syntactical elements and utilizing the trained second neural network, of a resulting item of content; and

causing the resulting item of content to be sent to a user.

28. The method of claim 27 , wherein the first neural network and the second neural network are components of the system.

29. The method of claim 27 , wherein at least a portion of the second plurality of items of content is automatically input into the second neural network for training of the second neural network, in response to being generated.

30. The method of claim 27 , wherein the second plurality of items of content include both syntactical elements and images.

31. The method of claim 27 , wherein the plurality of syntactical elements is identified in response to being received from the user that includes a human user.

32. The method of claim 27 , wherein the trained first neural network is a component of the system, and the second neural network is external to the system.

33. The method of claim 27 , wherein the second plurality of items of content is generated by the trained first neural network based on a composition that results in a chain including linked representations of each of a plurality of semantic chains.

34. The method of claim 27 , wherein the second plurality of items of content is generated by the trained first neural network based on operations including:

one or more searches in at least one database;

a composition that results in a chain including linked representations of each of a plurality of semantic chains; and

generating an explanation for at least one application of the trained first neural network.

35. The method of claim 34 , wherein one or more of the operations is iteratively performed by application of the trained first neural network based on different pluralities of items of content, until an expected net information value, that indicates an expected affect on an output from the trained first neural network, is insufficient.

36. The method of claim 27 , wherein the first plurality of attentions are prioritized.

37. The method of claim 36 , wherein the first plurality of attentions are prioritized utilizing a single data structure representing all of the first plurality of items of content and a position aspect as well as a non-position aspect thereof.

38. The method of claim 36 , wherein the first plurality of attentions are prioritized utilizing a single data structure representing all of the first plurality of items of content that constitute an entirety of a syntactical element portion of a training data set.

39. The method of claim 36 , wherein the first plurality of attentions are prioritized utilizing a single matrix representing all of the first plurality of items of content.

40. The method of claim 36 , wherein the first plurality of attentions are prioritized, before any usage of the prioritization of any of the first plurality of attentions.

41. The method of claim 36 , wherein the first plurality of attentions are prioritized in a single operation.

42. The method of claim 36 , wherein the first plurality of attentions are prioritized based on a first relational aspect among the different subsets of the first plurality of items of content and based on a second relational aspect among the different subsets of the first plurality of items of content.

43. The method of claim 42 , wherein the prioritization is completed before any utilization of any prioritization of any of the first plurality of attentions via the first neural network.

44. The method of claim 36 , wherein the first plurality of attentions are prioritized to generate a result based on a relative importance or relevance of a plurality of relationships among the first plurality of items of content, and then another first plurality of attentions are prioritized based on the result.

45. The method of claim 36 , wherein the first plurality of attentions are prioritized utilizing first weights that reflect a relative importance or relevance of a plurality of relationships among the first plurality of items of content.

46. The method of claim 45 , wherein an additional first plurality of attentions are prioritized utilizing second weights that reflect the relative importance or relevance of the plurality of relationships among the first plurality of items of content, as updated based on probabilities generated as a result of the prioritization of the first plurality of attentions utilizing the first weights.

47. The method of claim 36 , wherein the prioritization results in only a subset of the first plurality of attentions corresponding to only a subset of the different subsets of the first plurality of items of content being a basis for the training of the first neural network.

48. The method of claim 27 , wherein the first neural network is trained on a training set including: images associated with corresponding syntactical elements to infer one or more syntactical elements corresponding to pixel patterns within the images, and training syntactical elements, including training natural language syntactical elements and training computer-executable code syntactical elements, that are utilized by computing hardware that processes the training syntactical elements to direct training attentions to representations of different subsets of the training syntactical elements.

49. The method of claim 27 , and further comprising causing:

a determination whether a first scenario described by the second plurality of items of content represents objective reality; and

based on a determination that the first scenario does not represent objective reality, an update, by applying the trained first neural network utilizing the second plurality of items of content, to the second plurality of items of content to describe a second scenario that represents objective reality, for use in training the second neural network.

50. The method of claim 27 , wherein one or more user preferences is identified, where the resulting item of content is caused to be generated, based on the one or more user preferences.

51. The method of claim 50 , wherein the plurality of syntactical elements is part of a prompt, and the one or more user preferences is based on at least one aspect of the prompt and at least one previous prompt that are received from the user that includes a human user.

52. The method of claim 27 , wherein the first plurality of items of content and the second plurality of items of content include different executable computer code.

53. The method of claim 27 , wherein the second plurality of items of content is generated by the trained first neural network based on at least two operations including at least two of:

one or more searches in at least one database;

a composition that results in a chain including linked representations of each of a plurality of semantic chains; and

generating an explanation for at least one application of the trained first neural network.

54. The method of claim 53 , wherein one or more of the at least two operations is iteratively performed by application of the trained first neural network based on different pluralities of items of content, until an expected net information value, that indicates an expected affect on an output from the trained first neural network, is insufficient.

55. The method of claim 53 , wherein all of the at least two operations are iteratively performed by application of the trained first neural network based on different pluralities of items of content, until an expected net information value, that indicates an expected affect on an output from the trained first neural network, is insufficient.

56. The method of claim 27 , wherein the second plurality of items of content is generated by the trained first neural network based on one or more searches in at least one database to augment input to the trained first neural network.

57. The method of claim 27 , wherein the second plurality of items of content is generated by the trained first neural network based on a composition that results in a chain including linked representations of each of a plurality of semantic chains.

58. The method of claim 27 , wherein the first plurality of items of content includes at least one chain including linked representations of each of a plurality of semantic chains.

59. The method of claim 27 , wherein the first neural network is trained based on at least one operation including generating a natural-language explanation for at least one application of the first neural network.

60. The method of claim 59 , wherein the at least one operation is iteratively performed by application of the first neural network based on different pluralities of items of content, until an expected net information value, that indicates an expected affect on an output from the first neural network, is insufficient.

61. The method of claim 27 , wherein the first neural network is trained by applying one or more additional attentions to a mathematical representation that represents a rationale in connection with the training of the first neural network based on the first plurality of items of content.

62. The method of claim 61 , wherein another one or more additional attentions are applied to one or more other mathematical representations, until an expected net information value, that indicates an expected affect on an output from the first neural network, is insufficient.

63. The method of claim 27 , wherein the second plurality of items of content is generated by the trained first neural network based on an explanation generated by the trained first neural network for a representation of a third plurality of items of content generated by the trained first neural network after being trained by applying the first plurality of attentions.

64. The method of claim 63 , wherein the second plurality of items of content and the explanation are generated automatically without the explanation being directly manually initiated after the trained first neural network is manually prompted to initiate content generation, and the explanation for the representation of the third plurality of items of content includes a natural-language explanation for the generation of the third plurality of items of content.

65. The method of claim 27 , wherein a representation of the second plurality of items of content corresponds with a description of a first scenario that is updated, based on a mathematical representation that is generated by the trained first neural network in connection with a representation of a third plurality of items of content that corresponds with the first scenario and that is generated by the trained first neural network after being trained by applying the first plurality of attentions.

66. The method of claim 65 , wherein the representation of the second plurality of items of content and the mathematical representation are generated automatically without either being directly manually initiated after the trained first neural network is manually prompted to initiate content generation, and the trained first neural network is trained by applying the first plurality of attentions in an unsupervised manner after being manually initiated.

67. The method of claim 65 , wherein the mathematical representation generated in connection with the representation of the third plurality of items of content, represents a rationale for one or more of the third plurality of items of content themselves, or the generation thereof.

68. The method of claim 27 , wherein the second plurality of items of content includes a description of a first scenario that is updated, based on an explanation generated by the trained first neural network as to why a third plurality of items of content that describe the first scenario were generated by the trained first neural network.

69. The method of claim 68 , wherein the second plurality of items of content, the explanation, and the third plurality of items of content, are generated automatically without being directly manually initiated after the trained first neural network is manually prompted to initiate content generation, and the trained first neural network is trained, at least in part, by applying the first plurality of attentions without human supervision in response the training of the first neural network being manually started.

70. The method of claim 27 , and further comprising causing:

a determination whether a first scenario described by the resulting item of content represents objective reality; and

71. The method of claim 27 , wherein the second plurality of items of content is generated by the application of the second plurality of attentions to representations of different subsets of a third plurality of items of content that is generated by the trained first neural network after being trained by applying the first plurality of attentions, and further comprising causing:

a determination whether a first scenario described by the third plurality of items of content represents objective reality; and

based on a determination that the first scenario does not represent objective reality, generation, by applying the trained first neural network utilizing the third plurality of items of content, of the second plurality of items of content that describes a second scenario that represents objective reality.

72. The method of claim 71 , and further comprising:

before causing the determination whether the first scenario represents objective reality, causing a communication including the third plurality of items of content to be sent to a human training user;

receiving, from the human training user, a response to the communication; and

in response to receiving the response to the communication, causing the determination whether the first scenario represents objective reality based on the response, such that, based on the determination that the first scenario does not represent objective reality, the second plurality of items of content is caused to be automatically generated based on the response.

73. The method of claim 72 , wherein the determination whether the first scenario represents objective reality is based on the response, by applying the trained first neural network to at least one semantic change of the response.

74. The method of claim 71 , wherein the determination is automatically performed and the second plurality of items of content is caused to be automatically generated without human intervention.

75. The method of claim 74 , wherein, before causing the determination whether the first scenario represents objective reality, information is automatically retrieved based on a prompt, such that third plurality of items of content is caused to be automatically generated utilizing the information.

76. The method of claim 75 , wherein the prompt is automatically generated.

77. The method of claim 76 , wherein only one of:

each instance of the generation is performed by hardware;

each instance of the generation is performed by software;

each instance of the generation is caused by software and performed by hardware;

each instance of the generation is caused and performed by software;

only a subset of each instance of the generation is performed by hardware;

only a subset of each instance of the generation is performed by software;

only a subset of each instance of the generation is caused by software and performed by hardware;

only a subset of each instance of the generation is caused and performed by software;

the causing includes a direct causation;

the causing includes an indirect causation;

the system is portable;

the system is not portable;

the system includes a user device;

the system includes a user device to which the resulting item of content is caused to be sent;

the system does not include a user device;

the system does not include a user device to which the resulting item of content is caused to be sent;

the system, further comprising: a user device;

the system, further comprising: a user device to which the resulting item of content is caused to be sent;

the automatic internal application of the first plurality of attentions is automatic by not being caused by the user;

the automatic internal application of the first plurality of attentions is automatic by not being directly caused by the user;

the automatic internal application of the first plurality of attentions is automatic by being indirectly caused by a human;

the automatic internal application of the first plurality of attentions is automatic by being caused without any human intervention;

the automatic internal application of the first plurality of attentions is automatic by being caused without human intervention after the training is manually initiated by a human;

the automatic internal application of the first plurality of attentions is automatic by being caused without human intervention after the training is manually initiated by a training human user;

the automatic internal application of the first plurality of attentions is automatic by being caused without human intervention after the first neural network is caused to be accessed by a human;

the automatic internal application of the first plurality of attentions is automatic by being caused without human intervention after the first neural network is caused to be accessed by a training human user;

the automatic internal application of the first plurality of attentions to the representations of different subsets of the first plurality of items of content;

the automatic internal application of the first plurality of attentions is automatic by not being applied by the user;

the automatic internal application of the first plurality of attentions is automatic by not being directly applied by the user;

the automatic internal application of the first plurality of attentions is automatic by being indirectly applied by a human;

the automatic internal application of the first plurality of attentions is automatic by being applied without any human intervention;

the automatic internal application of the first plurality of attentions is automatic by being applied without human intervention after the training is manually initiated by a human;

the automatic internal application of the first plurality of attentions is automatic by being applied without human intervention after the training is manually initiated by a training human user;

the automatic internal application of the first plurality of attentions is automatic by being applied without human intervention after the first neural network is caused to be accessed by a human;

the automatic internal application of the first plurality of attentions is automatic by being applied without human intervention after the first neural network is caused to be accessed by a training human user;

the automatic internal application of the first plurality of attentions is automatic by not being caused to be applied by the user;

the automatic internal application of the first plurality of attentions is automatic by not being directly caused to be applied by the user;

the automatic internal application of the first plurality of attentions is automatic by being indirectly caused to be applied by a human;

the automatic internal application of the first plurality of attentions is automatic by being caused to be applied without any human intervention;

the automatic internal application of the first plurality of attentions is automatic by being caused to be applied without human intervention after the training is manually initiated by a human;

the automatic internal application of the first plurality of attentions is automatic by being caused to be applied without human intervention after the training is manually initiated by a training human user;

the automatic internal application of the first plurality of attentions is automatic by being caused to be applied without human intervention after the first neural network is caused to be accessed by a human;

the automatic internal application of the first plurality of attentions is internal by being internally caused;

the automatic internal application of the first plurality of attentions is internal by being internally caused within the system;

the automatic internal application of the first plurality of attentions is internal by being caused from within the system;

the automatic internal application of the first plurality of attentions is internal by being caused from within the system after the training is caused by a human user;

the automatic internal application of the first plurality of attentions is internal by being caused from within the system after the training is externally initiated;

the automatic internal application of the first plurality of attentions is internal by being caused from within the system in response to the training being externally caused;

the automatic internal application of the first plurality of attentions is internal by being internally caused within the system, but external to the first neural network;

the automatic internal application of the first plurality of attentions is internal by being caused from within the system, but external to the first neural network;

the automatic internal application of the first plurality of attentions is internal by being internally caused within the first neural network;

the automatic internal application of the first plurality of attentions is internal by being caused from within the first neural network;

the automatic internal application of the first plurality of attentions is internal by being caused by at least one of one or more processors;

the automatic internal application of the first plurality of attentions is internal by being caused by at least one of one or more programs;

the automatic internal application of the first plurality of attentions is internal by being caused by at least one of a plurality of instructions;

the automatic internal application of the first plurality of attentions is internal by the first plurality of attentions not being human attentions;

the automatic internal application of the first plurality of attentions is internal by being internally applied;

the automatic internal application of the first plurality of attentions is internal by being internally applied within the system;

the automatic internal application of the first plurality of attentions is internal by being applied within the system;

the automatic internal application of the first plurality of attentions is internal by being applied within the system after the training is caused by a human user;

the automatic internal application of the first plurality of attentions is internal by being applied within the system after the training is externally initiated;

the automatic internal application of the first plurality of attentions is internal by being applied within the system in response to the training being externally caused;

the automatic internal application of the first plurality of attentions is internal by being internally applied within the system, but external to the first neural network;

the automatic internal application of the first plurality of attentions is internal by being applied from within the system, but external to the first neural network;

the automatic internal application of the first plurality of attentions is internal by being internally applied within the first neural network;

the automatic internal application of the first plurality of attentions is internal by being applied from within the first neural network;

the automatic internal application of the first plurality of attentions is internal by being applied by at least one of one or more processors;

the automatic internal application of the first plurality of attentions is internal by being applied by at least one of one or more programs;

the automatic internal application of the first plurality of attentions is internal by being applied by at least one of a plurality of instructions;

the automatic internal application of the first plurality of attentions is internal by being internally caused to be applied;

the automatic internal application of the first plurality of attentions is internal by being internally caused to be applied within the system;

the automatic internal application of the first plurality of attentions is internal by being caused to be applied from within the system;

the automatic internal application of the first plurality of attentions is internal by being caused to be applied within the system after the training is caused by a human user;

the automatic internal application of the first plurality of attentions is internal by being caused to be applied within the system after the training is externally initiated;

the automatic internal application of the first plurality of attentions is internal by being caused to be applied within the system in response to the training being externally caused;

the automatic internal application of the first plurality of attentions is internal by being internally caused to be applied within the system, but external to the first neural network;

the automatic internal application of the first plurality of attentions is internal by being caused to be applied from within the system, but external to the first neural network;

the automatic internal application of the first plurality of attentions is internal by being internally caused to be applied within the first neural network;

the automatic internal application of the first plurality of attentions is internal by being caused to be applied from within the first neural network;

the automatic internal application of the first plurality of attentions is internal by being caused to be applied by at least one of one or more processors;

the automatic internal application of the first plurality of attentions is internal by being caused to be applied by at least one of one or more programs;

the automatic internal application of the first plurality of attentions is internal by being caused to be applied by at least one of a plurality of instructions;

the first plurality of items of content includes syntactical elements;

the first plurality of items of content includes syntactical elements, including computer code;

the first plurality of items of content includes images;

the first plurality of items of content includes only syntactical elements;

the first plurality of items of content includes only computer code;

the first plurality of items of content includes only images;

the first plurality of items of content includes different forms of content;

the first plurality of items of content includes syntactical elements, images, and computer code;

the representations of the different subsets of the first plurality of items of content, includes the different subsets of the first plurality of items of content themselves;

the representations of the different subsets of the first plurality of items of content, includes a processed version of the different subsets of the first plurality of items of content;

the representations of the different subsets of the first plurality of items of content, includes a processed form of the different subsets of the first plurality of items of content;

the representations of the different subsets of the first plurality of items of content, includes a mathematical representation of the different subsets of the first plurality of items of content;

the representations of the different subsets of the first plurality of items of content, includes a digital format of the different subsets of the first plurality of items of content;

the automatic internal application of the first plurality of attentions to the representations of different subsets of the first plurality of items of content, includes application of: a first one of the first plurality of attentions to a first one of the representations of different subsets of the first plurality of items of content, and a second one of the first plurality of attentions to a second one of the representations of different subsets of the first plurality of items of content;

the automatic internal application of the second plurality of attentions and the utilizing the trained first neural network, includes the automatic internal application of the second plurality of attentions before utilizing the trained first neural network;

the second plurality of items of content includes at least a portion of the first plurality of items of content;

the second plurality of items of content does not include any portion of the first plurality of items of content;

the identification of the plurality of syntactical elements includes receiving the plurality of syntactical elements;

the identification of the plurality of syntactical elements includes occurs after receiving the plurality of syntactical elements;

the identification of the plurality of syntactical elements includes automatic identification of the plurality of syntactical elements;

the access to the first neural network includes direct access;

the access to the first neural network includes indirect access;

the causing of the access to the first neural network includes direct causing;

the causing of the access to the first neural network includes indirect causing;

the user includes a human user;

the user includes a non-human user;

the first neural network is capable of multiple iterations utilizing a same hardware;

the first neural network is capable of multiple iterations utilizing a same software;

the first neural network is capable of multiple iterations utilizing a same process;

the first neural network is capable of multiple iterations each utilizing different hardware;

the first neural network is capable of multiple iterations each utilizing different software;

the first neural network is capable of multiple iterations each utilizing a same process;

the first neural network is capable of multiple iterations each utilizing a same process and a different hardware;

the first neural network is capable of multiple iterations each utilizing a same hardware and a different process;

the first neural network includes a single computer-implemented neural network;

the first neural network includes a plurality of computer-implemented neural networks;

the first neural network includes a plurality of computer-implemented neural networks, where a same one or more of the plurality of computer-implemented neural networks is utilized for each instance of the causing;

the first neural network includes a plurality of computer-implemented neural networks, where a different one or more of the plurality of computer-implemented neural networks is utilized for each instance of the causing;

the first neural network is implemented utilizing the system;

the first neural network is implemented utilizing another system other than the system;

the first neural network is implemented utilizing one or more processors;

the first neural network is implemented utilizing one or more other processors other than the one or more processors;

the first neural network is implemented utilizing one or more other processors other than the one or more processors, such that the one or more processors causes operation of the first neural network utilizing one or more other processors;

the application of the first neural network includes a utilization of the first neural network;

the application of the first neural network includes a direct utilization of the first neural network;

the application of the first neural network includes an indirect utilization of the first neural network;

the second neural network is capable of multiple iterations utilizing a same hardware;

the second neural network is capable of multiple iterations utilizing a same software;

the second neural network is capable of multiple iterations utilizing a same process;

the second neural network is capable of multiple iterations each utilizing different hardware;

the second neural network is capable of multiple iterations each utilizing different software;

the second neural network is capable of multiple iterations each utilizing a same process;

the second neural network is capable of multiple iterations each utilizing a same process and a different hardware;

the second neural network is capable of multiple iterations each utilizing a same hardware and a different process;

the second neural network includes a single computer-implemented neural network;

the second neural network includes a plurality of computer-implemented neural networks;

the second neural network includes a plurality of computer-implemented neural networks, where a same one or more of the plurality of computer-implemented neural networks is utilized for each instance of the causing;

the second neural network includes a plurality of computer-implemented neural networks, where a different one or more of the plurality of computer-implemented neural networks is utilized for each instance of the causing;

the second neural network is implemented utilizing the system;

the second neural network is implemented utilizing another system other than the system;

the second neural network is implemented utilizing the one or more processors;

the second neural network is implemented utilizing one or more other processors other than the one or more processors;

the second neural network is implemented utilizing one or more other processors other than the one or more processors, such that the one or more processors causes operation of the second neural network utilizing one or more other processors;

the application of the second neural network includes a utilization of the second neural network;

the application of the second neural network includes a direct utilization of the second neural network;

the application of the second neural network includes an indirect utilization of the second neural network;

the second neural network is another instance of the first neural network;

the second neural network is another instance that is the same as the first neural network, but is trained differently;

the second neural network is different from the first neural network;

the second neural network is part of the system and the first neural network is not part of the system;

the first neural network is part of the system and the second neural network is not part of the system;

the first neural network and the second neural network are each part of the system;

the first neural network and the second neural network are both not part of the system;

at least one of the syntactical elements includes at least one of: a word, a phrase, a sentence, a punctuation symbol, or a chain;

at least one of the syntactical elements includes at least one of the syntactical element itself, or a mathematical representation thereof;

at least one of the syntactical elements includes at least one of the syntactical element itself, or a representation thereof;

at least one of the syntactical elements includes at least one of the syntactical element itself, or a symbol thereof;

at least one of the syntactical elements includes a syntactical structure;

at least one of the syntactical elements does not include a syntactical structure;

each representation includes a mathematical representation;

each representation includes a symbolic representation;

each representation includes a processed form of a corresponding one or more syntactical elements;

the resulting item of content is sent directly to the user;

the resulting item of content is sent indirectly to the user;

the resulting item of content is sent to the user via a network;

the resulting item of content is sent to the user via a wide area network;

the resulting item of content is not sent to the user via a network;

the resulting item of content is in written form;

the resulting item of content is in written form, by visually showing syntactical elements;

the resulting item of content is in visual form;

the resulting item of content is in audible form, by converting syntactical elements to audible versions of the syntactical elements;

the plurality of syntactical elements are included in a single chain;

the plurality of syntactical elements are included in a single composite chain;

the plurality of syntactical elements are included in a plurality of chains;

the first plurality of attentions includes a first attention that is part of a multiple-attention direction, which includes a direction of a plurality of attentions;

the first plurality of attentions includes a first attention that is part of a multiple-attention direction, which includes a direction of a plurality of attentions to different syntactical elements;

the first plurality of attentions includes a first attention that is part of a multiple-attention direction, which includes a direction of a plurality of attentions each to one of a plurality of different syntactical elements;

the first plurality of attentions includes a first attention that is part of a multiple-attention direction that includes different attentions from a multiple-attention direction of which the second attention is part;

the first plurality of attentions includes a first attention that is part of a multiple-attention direction that includes directions that are different from those associated with a multiple-attention direction of which the second attention is part;

the first plurality of attentions includes a first attention that is an initial attention;

the first plurality of attentions includes a first attention that is not an initial attention;

the first plurality of attentions includes a first attention that includes a reflection;

the first plurality of attentions includes a first attention that does not include a reflection;

at least one of the first plurality of attentions includes a first attention that or the second attention, includes monitored attention;

at least one of the first plurality of attentions includes a first attention that includes inferred attention;

at least one of the first plurality of attentions includes a first attention that includes a consciousness;

at least one of the first plurality of attentions includes a first attention that includes an awareness;

at least one of the first plurality of attentions includes a first attention that is represented by an identified one or more of the plurality of syntactical elements;

at least one of the first plurality of attentions includes a first attention that is identified by identification of a representative one or more of the plurality of syntactical elements;

at least one of the first plurality of attentions includes a first attention that is based on at least one of: processing input from a sensor, processing input from externally sourced content, processing input from internally sourced content, a value of information, or a probabilistic selection process;

at least one of the first plurality of attentions includes a first attention that is not based on at least one of: processing input from a sensor, processing input from externally sourced content, processing input from internally sourced content, a value of information, or a probabilistic selection process;

at least one of the first plurality of attentions includes a first attention that includes system attention;

at least one of the first plurality of attentions includes a first attention that does not include user attention;

at least one of the first plurality of attentions includes a first attention that includes a focus of attention;

at least one of the first plurality of attentions includes a first attention that includes a potential attention;

at least one of the first plurality of attentions includes a first attention that includes a stream of attention;

at least one of the first plurality of attentions includes a first attention that is not generated by the first neural network;

at least one of the first plurality of attentions is based on processing input from a sensor;

at least one of the first plurality of attentions is based on processing input from externally sourced content;

at least one of the first plurality of attentions is based on processing input from internally sourced content;

at least one of the first plurality of attentions is based on a value of information;

at least one of the first plurality of attentions is based on a probabilistic selection process;

at least one of the first plurality of attentions is based on processing input from a sensor, processing input from externally or internally sourced content, a value of information, and a probabilistic selection process;

the first plurality of attentions includes a first attention that is directed to a representation of only a single subset of the first plurality of items of content;

the first plurality of attentions includes a first attention that is directed to a representation of a subset of the first plurality of items of content, including multiple items of content;

the first plurality of attentions is collectively directed to representations of all of the first plurality of items of content;

each instance of based on, includes directly based on;

at least one instance of based on, includes directly based on;

each instance of based on, includes indirectly based on;

at least one instance of based on, includes indirectly based on;

each instance of in response, includes a direct response;

each instance of in response, includes an indirect response;

each instance of based, includes directly based;

at least one instance of based, includes directly based;

each instance of based, includes indirectly based;

at least one instance of based, includes indirectly based;

each instance of utilizing, includes directly utilizing;

at least one instance of utilizing, includes directly utilizing;

each instance of utilizing, includes indirectly utilizing;

at least one instance of utilizing, includes indirectly utilizing;

each instance of the causing is an act;

each instance of the causing is not a step;

the causing are acts;

the causing are not steps;

the access to the first neural network, the training of the second neural network, the identification, the generation, and the sending, are acts;

the access to the first neural network, the training of the second neural network, the identification, the generation, and the sending, are not steps;

the first neural network includes a neural network-based system capable of a first operation including neural network-independent attention prioritization, and a second operation to direct the first attention by causing processing of the first neural network to be based on the neural network-independent attention prioritization, by affecting a particular manner of the processing by the first neural network;

the system is further configured such that the first neural network is caused to be accessed and the second plurality of items of content is caused to be generated utilizing a first processor of one or more processors of a first apparatus of the system that executes a first program of one or more programs that is stored in a first memory of one or memories of the first apparatus of the system, and the plurality of syntactical elements is caused to be identified and the resulting item of content is caused to be generated utilizing a second processor of the one or more processors of a second apparatus of the system that executes a second program of the one or more programs that is stored in a second memory of the one or memories of the second apparatus of the system;

the first neural network is caused to be accessed, the second plurality of items of content is caused to be generated, the plurality of syntactical elements is caused to be identified, and the resulting item of content is caused to be generated, utilizing a same processor of the one or more processors of a same apparatus of the system that executes a same program of the one or more programs that is stored in a same memory of the one or memories of the same apparatus of the system;

the first neural network is caused to be accessed and the second plurality of items of content is caused to be generated, by a first one of the one or more processors; and the plurality of syntactical elements is caused to be identified and the resulting item of content is caused to be generated, by a second one of the one or more processors;

the system includes a non-transitory computer-readable media storing a plurality of instructions, including a first storage including a first portion of the instructions and a second storage including a second portion of the instructions.

78. The method of claim 75 , wherein the prompt is received from a human user.

79. The method of claim 75 , wherein the information is automatically retrieved from a source other than the trained first neural network.

80. The method of claim 75 , wherein the third plurality of items of content is caused to be automatically generated utilizing the information, for reducing a probability that the third plurality of items of content describe the first scenario that does not represent objective reality.

81. The method of claim 75 , wherein the third plurality of items of content is caused to be automatically generated utilizing the information, by editing the prompt based on the information, such that the third plurality of items of content is caused to be automatically generated based on the edited prompt.

82. The method of claim 27 , wherein a setting is received, for causing a tuning of the generation of the second plurality of items of content based on the setting, such that the setting causes the tuning of the generation of the second plurality of items of content, by causing a selection of a first item of content with a first associated probability to be involved with the second plurality of items of content, instead of a second item of content with a second associated probability, even though the first associated probability is less than the second associated probability.

83. The method of claim 27 , wherein the first plurality of items, the second plurality of items, and the resulting item of content each includes a different one or more of a plurality of computer-executable instruction elements.

84. The method of claim 83 , wherein the plurality of syntactical elements does not include any program code.

85. The method of claim 83 , wherein the plurality of syntactical elements includes a first portion that does not include any program code, and a second portion that does include program code.

86. The method of claim 83 , wherein the plurality of syntactical elements includes at least a portion of an executable computer program.

87. The method of claim 86 , wherein the at least portion of executable computer program is received from the user that includes a human user.

88. The method of claim 86 , wherein the at least portion of executable computer program is automatically generated.

89. The method of claim 86 , wherein the at least portion of executable computer program includes at least a portion of at least one of one or more programs of the system.

90. The method of claim 89 , wherein the at least portion of the at least one of the one or more programs is identified by the user that includes a human user.

91. The method of claim 89 , wherein the at least portion of the at least one of the one or more programs is automatically identified.

92. The method of claim 83 , wherein a representation of the second plurality of items of content, including a representation of a first one or more of the plurality of computer-executable instruction elements, is generated by the trained first neural network based on a logical evaluation by the trained first neural network for a representation of a third plurality of items of content, including a representation of a second one or more of the plurality of computer-executable instruction elements generated by the trained first neural network.

93. The method of claim 92 , wherein the logical evaluation is performed on a mathematical representation of the third plurality of items of content themselves, or the generation thereof.

94. The method of claim 92 , wherein the representation of the second plurality of items of content and the representation of the third plurality of items of content, are generated automatically without being directly manually initiated after the trained first neural network is manually prompted to initiate content generation, and the first neural network is trained, at least in part, by applying the first plurality of attentions without human supervision in response to the training of the first neural network being manually started.

95. The method of claim 92 , wherein the first one or more of the plurality of computer-executable instruction elements includes a modified version of the second one or more of the plurality of computer-executable instruction elements.

96. The method of claim 92 , wherein a setting is received, for causing a tuning of the generation of the second plurality of items of content based on the setting, such that the setting causes the tuning of the generation of the second plurality of items of content, by causing a selection of a first item of content with a first associated probability to be involved with the second plurality of items of content, instead of a second item of content with a second associated probability, even though the first associated probability is less than the second associated probability.

97. The method of claim 83 , wherein a setting is received, for causing a tuning of the generation of the second plurality of items of content based on the setting, such that the setting causes the tuning of the generation of the second plurality of items of content, by causing a selection of a first item of content with a first associated probability to be involved with the second plurality of items of content, instead of a second item of content with a second associated probability, even though the first associated probability is less than the second associated probability.

98. The method of claim 83 , wherein an action is caused to be automatically taken on the resulting item of content including a first one or more of the plurality of computer-executable instruction elements.

99. The method of claim 83 , wherein the resulting item of content including a first one or more of the plurality of computer-executable instruction elements, is automatically executed.

100. The method of claim 83 , wherein an action is caused to be automatically taken on the resulting item of content including a first one or more of the plurality of computer-executable instruction elements, without causing a communication to be sent to a human user beforehand.

101. The method of claim 83 , wherein the resulting item of content including a first one or more of the plurality of computer-executable instruction elements, is caused to be executed, without causing a communication to be sent to a human user beforehand.

102. The method of claim 27 , wherein:

the first plurality of items of content includes a syntactical element portion of a training set, and a single matrix that represents the syntactical element portion of the training set is generated, based on a position of each of the first plurality of items of content within the syntactical element portion of the training set and further based on a relationship, other than position, of at least one of the first plurality of items of content with at least one other of the first plurality of items of content;

the first plurality of attentions, that is applied to the representations of the different subsets of the first plurality of items of content, is subject to a first prioritization utilizing first weights that reflect a relative importance or relevance of a plurality of relationships among the first plurality of items of content, and utilizing the single matrix that represents the syntactical element portion of the training set, the first prioritization of the first plurality of attentions being a function of an application of a first process that is based on a first relational aspect among the different subsets of the first plurality of items of content and an application of a second process that is based on a second relational aspect among the different subsets of the first plurality of items of content, such that the first prioritization is completed in a single operation before any utilization of the first prioritization of any of the first plurality of attentions;

a first plurality of probabilities is generated, utilizing one or more cognitive computing-based processors, based on at least a portion of the first prioritization and by application of the first neural network, such that the first prioritization results in only a subset of the first plurality of attentions corresponding to only a subset of the different subsets of the first plurality of items of content being a basis for the generation of the first plurality of probabilities;

an additional first plurality of attentions, that is applied to representations of at least a portion of the different subsets of the first plurality of items of content, is subject to a second prioritization utilizing second weights that reflect the relative importance or relevance of at least a portion of the plurality of relationships as updated based on the first plurality of probabilities, the second prioritization of the additional first plurality of attentions being a function of another application of the first process that is based on the first relational aspect and another application of the second process that is based on the second relational aspect, such that the second prioritization is completed in another single operation before any utilization of the second prioritization of any of the additional first plurality of attentions; and

a second plurality of probabilities is generated, utilizing the one or more cognitive computing-based processors, based on at least a portion of the second prioritization and by application of the first neural network, such that the second prioritization results in only a subset of the additional first plurality of attentions being a basis for the generation of the second plurality of probabilities.

103. The method of claim 27 , wherein:

the plurality of syntactical elements includes an entirety of a syntactical element portion of a prompt, and a single matrix that represents the entirety of the syntactical element portion of the prompt is generated, based on a position of each of the plurality of syntactical elements within the entirety of the syntactical element portion of the prompt and further based on a relationship, other than position, of at least one of the plurality of syntactical elements with at least one other of the plurality of syntactical elements;

the fourth plurality of attentions, that is applied to the representations of the different subsets of the plurality of syntactical elements, is subject to a first prioritization utilizing first weights that reflect a relative importance or relevance of a plurality of relationships among the plurality of syntactical elements, and utilizing the single matrix that represents the entirety of the syntactical element portion of the prompt, the first prioritization of the fourth plurality of attentions being a function of an application of a first process that is based on a first relational aspect among the different subsets of the plurality of syntactical elements and an application of a second process that is based on a second relational aspect among the different subsets of the plurality of syntactical elements, such that the first prioritization is completed in a single operation before any utilization of the first prioritization of any of the fourth plurality of attentions;

a first plurality of probabilities is generated, utilizing one or more cognitive computing-based processors, based on at least a portion of the first prioritization and by application of the second neural network, such that the first prioritization results in only a subset of the fourth plurality of attentions corresponding to only a subset of the different subsets of the plurality of syntactical elements being a basis for the generation of the first plurality of probabilities;

an additional fourth plurality of attentions, that is applied to representations of at least a portion of the different subsets of the plurality of syntactical elements, is subject to a second prioritization utilizing second weights that reflect the relative importance or relevance of at least a portion of the plurality of relationships as updated based on the first plurality of probabilities, the second prioritization of the additional fourth plurality of attentions being a function of another instance of the first process that is based on the first relational aspect and another instance of the second process that is based on the second relational aspect, such that the second prioritization is completed in another single operation before any utilization of the second prioritization of any of the additional fourth plurality of attentions; and

a second plurality of probabilities is generated, utilizing the one or more cognitive computing-based processors, based on at least a portion of the second prioritization and by application of the second neural network, such that the second prioritization results in only a subset of the additional fourth plurality of attentions being a basis for the generation of the second plurality of probabilities.

104. The method of claim 27 , wherein the system includes one or more servers and a mobile device, wherein the mobile device of the system is configured to generate the resulting item of content and the one or more servers of the system is configured to perform the training of the first neural network and the second neural network, where the one or more servers is configured such that:

a first plurality of probabilities is generated, utilizing one or more cognitive computing-based processors of the one or more servers, based on at least a portion of the first prioritization and by application of the first neural network, such that the first prioritization results in only a subset of the first plurality of attentions corresponding to only a subset of the different subsets of the first plurality of items of content being a basis for the generation of the first plurality of probabilities;

a second plurality of probabilities is generated, utilizing the one or more cognitive computing-based processors of the one or more servers, based on at least a portion of the second prioritization and by application of the first neural network, such that the second prioritization results in only a subset of the additional first plurality of attentions being a basis for the generation of the second plurality of probabilities.

105. The method of claim 27 , wherein the system includes one or more servers and a mobile device, wherein the one or more servers of the system is configured to perform the training of the first neural network and the second neural network and the mobile device of the system is configured to generate the resulting item of content, where the mobile device is configured such that:

the plurality of syntactical elements includes a syntactical element portion of a prompt, and at least one vector that represents the syntactical element portion of the prompt is generated, based on a position of each of the plurality of syntactical elements within the syntactical element portion of the prompt;

a first plurality of probabilities is generated, based on at least a portion of the first prioritization and by application of the second neural network;

a second plurality of probabilities is generated, based on at least a portion of the second prioritization and by application of the second neural network.

106. The method of claim 27 , wherein the system includes one or more servers that: performs the training of the first neural network and the second neural network, and generates the resulting item of content.

107. The method of claim 27 , and further comprising:

automatically causing, without requiring human user input after the identification of the plurality of syntactical elements and before the resulting item of content being caused to be sent to the user that includes a human user:

generation, utilizing the trained second neural network, of a plurality of sequences of elements, where each of the plurality of sequences of elements includes multiple elements;

evaluation of the plurality of sequences of elements; and

selection of at least one of the plurality of sequences of elements to be included in the resulting item of content, based on the evaluation.

108. The method of claim 107 , wherein each of the plurality of sequences of elements represents a different scenario.

109. The method of claim 107 , wherein each of the plurality of sequences of elements represents a different explanation for content generation.

110. The method of claim 107 , wherein the plurality of sequences of elements is generated by automatic internal application of the fourth plurality of attentions to the representations of the different subsets of the plurality of syntactical elements and utilizing the trained second neural network.

111. The method of claim 107 , wherein the plurality of sequences of elements is generated by automatic internal application of a fifth plurality of attentions to the representations of the different subsets of the plurality of syntactical elements and utilizing the trained second neural network.

112. The method of claim 107 , wherein one or more of the plurality of sequences of elements is generated based on a probabilistic exploration.

113. The method of claim 107 , wherein one or more of the plurality of sequences of elements is generated based on an expected net information value, that indicates an expected affect on an output from the trained second neural network.

114. The method of claim 107 , wherein the plurality of sequences of elements includes at least one of: a plurality of chains of elements, or a plurality of streams of elements.

115. The method of claim 107 , wherein the plurality of sequences of elements are iteratively linked.

116. The method of claim 107 , wherein the plurality of sequences of elements include a first subset of the plurality of sequences of elements and a second subset of the plurality of sequences of elements, such that the first subset of the plurality of sequences of elements is generated and evaluated during a first iteration to select at least one of the first subset of the plurality of sequences of elements for being a basis for generating and evaluating a second subset of the plurality of sequences of elements during a second iteration, where the at least one of the plurality of sequences of elements includes one or more of the second subset of the plurality of sequences of elements.

117. The method of claim 116 , wherein the at least one of the plurality of sequences of elements includes all of the second subset of the plurality of sequences of elements.

118. The method of claim 116 , wherein an element of the first subset of the plurality of sequences of elements and an element of the second subset of the plurality of sequences of elements are linked.

119. The method of claim 116 , wherein the identification of the plurality of syntactical elements is in response to a receipt thereof from the human user.

120. The method of claim 107 , wherein elements of the at least one of the plurality of sequences of elements includes one or more representations of multiple syntactical-based elements.

121. The method of claim 107 , wherein elements of the at least one of the plurality of sequences of elements includes one or more representations of multiple computer executable instruction elements.

122. The method of claim 107 , wherein elements of the at least one of the plurality of sequences of elements includes one or more representations of multiple pixel pattern elements.

123. The method of claim 107 , wherein the evaluation is caused by performing an internally-generated interrogative.

124. The method of claim 123 , wherein the internally-generated interrogative includes a request by the system for the system to generate an explanation for the generation of the plurality of sequences of elements by the system.

125. The method of claim 107 , wherein the evaluation includes a logical evaluation.

126. The method of claim 107 , wherein the evaluation includes performing a search.

127. The method of claim 107 , wherein the evaluation includes performing a probabilistic assessment.

128. The method of claim 107 , wherein the evaluation includes applying a heuristic rule.

129. The method of claim 107 , wherein the evaluation is performed by another system.

130. The method of claim 107 , wherein the evaluation is performed by automatic internal application of a fifth plurality of attentions to representations of different subsets of the plurality of sequences of elements and utilizing the trained second neural network.

131. The method of claim 107 , wherein the evaluation utilizes an expected net information value, that is based on an expected value and an expected cost and that indicates an expected affect on an output from the trained second neural network.

132. The method of claim 107 , wherein the evaluation includes a hierarchical structuring of the plurality of syntactical elements.

133. The method of claim 107 , wherein a setting is received, for causing a tuning of the generation of the plurality of sequences of elements based on the setting, such that the setting causes the tuning of the generation of the plurality of sequences of elements, by causing a selection of a first item of content with a first associated probability to be involved with the plurality of sequences of elements, instead of a second item of content with a second associated probability, even though the first associated probability is less than the second associated probability.

134. The method of claim 133 , wherein a setting is for causing a tuning of a creativity associated with the generation of the plurality of sequences of elements.

135. The method of claim 107 , wherein the plurality of sequences of elements is associated with one or more probabilities.

136. The method of claim 107 , wherein the plurality of sequences of elements is associated with one or more probabilities and is included in the resulting item of content.

137. The method of claim 27 , and further comprising causing:

generation, by automatic internal application of a fifth plurality of attentions to representations of different subsets of at least one item of content and utilizing the trained second neural network, of a plurality of sequences of elements, where each of the plurality of sequences of elements includes a plurality of elements;

evaluation of the plurality of sequences of elements; and

selection of at least one of the plurality of sequences of elements to be included in another resulting item of content, based on the evaluation.

138. The method of claim 137 , wherein the at least one item of content includes at least one image and the another resulting item of content includes at least one other image.

139. The method of claim 27 , and further comprising causing:

generation, utilizing the trained first neural network, of a plurality of sequences of elements, where each of the plurality of sequences of elements includes multiple elements; and

evaluation of the plurality of sequences of elements such that one or more of the evaluated plurality of sequences of elements is included in the second plurality of items of content, so that the second neural network is trained based thereon.

140. The method of claim 139 , and further comprising causing:

selection of at least one of the plurality of sequences of elements, based on the evaluation, such that a result of the selection is also included in the second plurality of items of content, so that the second neural network is trained based thereon.

141. The method of claim 139 , wherein the second neural network is trained to have expertise in a predetermined domain.

142. The method of claim 139 , wherein each of the plurality of sequences of elements represents a different scenario.

143. The method of claim 139 , wherein each of the plurality of sequences of elements represents a different explanation for content generation.

144. The method of claim 139 , wherein the plurality of sequences of elements is generated by automatic internal application of the second plurality of attentions to the representations of the different subsets of the first plurality of items of content and utilizing the trained first neural network.

145. The method of claim 139 , wherein the plurality of sequences of elements is generated by automatic internal application of a fifth plurality of attentions to the representations of the different subsets of the first plurality of items of content and utilizing the trained first neural network.

146. The method of claim 139 , wherein at least one of the plurality of sequences of elements is generated based on a probabilistic exploration.

147. The method of claim 139 , wherein at least one of the plurality of sequences of elements is generated based on an expected net information value, that indicates an expected affect on an output from the trained second neural network.

148. The method of claim 139 , wherein the plurality of sequences of elements are iteratively linked.

149. The method of claim 139 , wherein elements of at least one of the plurality of sequences of elements includes one or more representations of multiple syntactical-based elements.

150. The method of claim 139 , wherein elements of at least one of the plurality of sequences of elements includes one or more representations of multiple computer executable instruction elements.

151. The method of claim 139 , wherein elements of at least one of the plurality of sequences of elements includes one or more representations of multiple pixel pattern elements.

152. The method of claim 139 , wherein the evaluation is caused by performing an internally-generated interrogative.

153. The method of claim 152 , wherein the internally-generated interrogative includes a request by the system for the system to generate an explanation for the generation of the plurality of sequences of elements by the system.

154. The method of claim 139 , wherein the evaluation includes a logical evaluation.

155. The method of claim 139 , wherein the evaluation includes performing a search.

156. The method of claim 139 , wherein the evaluation includes performing a probabilistic assessment.

157. The method of claim 139 , wherein the evaluation includes applying a heuristic rule.

158. The method of claim 139 , wherein the evaluation is performed by another system.

159. The method of claim 139 , wherein the evaluation is performed by automatic internal application of a fifth plurality of attentions to the representations of the different subsets of the first plurality of items of content and utilizing the trained first neural network.

160. The method of claim 139 , wherein the evaluation includes utilizing an expected net information value, that is based on an expected value and an expected cost and that indicates an expected affect on an output from the trained second neural network.

161. The method of claim 139 , wherein the evaluation includes a hierarchical structuring of the plurality of syntactical elements.

162. The method of claim 139 , wherein a setting is received, for causing a tuning of the generation of the plurality of sequences of elements based on the setting, such that the setting causes the tuning of the generation of the plurality of sequences of elements, by causing a selection of a first item of content with a first associated probability to be involved with the plurality of sequences of elements, instead of a second item of content with a second associated probability, even though the first associated probability is less than the second associated probability.

163. The method of claim 162 , wherein the setting is for causing a tuning of a creativity associated with the generation of the plurality of sequences of elements.

164. The method of claim 139 , wherein the plurality of sequences of elements is associated with one or more probabilities.

165. The method of claim 139 , wherein the plurality of sequences of elements is associated with one or more probabilities and is included in the second plurality of items of content.

166. The method of claim 27 , wherein:

a first plurality of probabilities is generated, utilizing one or more cognitive computing-based processors, based on at least a portion of the first prioritization and by application of the second neural network, such that the first prioritization results in only a subset of the fourth plurality of attentions corresponding to only a subset of the different subsets of the plurality of syntactical elements being a basis for the generation of the first plurality of probabilities, which are utilized for the generation of the resulting item of content.

167. The method of claim 166 , wherein the resulting item of content is generated by the trained second neural network based on at least one operation including generating an explanation for at least one application of the trained second neural network.

168. The method of claim 167 , wherein the resulting item of content is generated by the trained second neural network based on one or more searches in at least one database to augment the plurality of syntactical elements.

169. The method of claim 168 , wherein the resulting item of content is generated by the trained second neural network based on a composition that results in a chain including linked representations of each of a plurality of semantic chains.

170. The method of claim 168 , wherein one or more user preferences is identified, where the resulting item of content is caused to be generated, based on the one or more user preferences.

171. The method of claim 170 , wherein the plurality of syntactical elements is part of a prompt, and the one or more user preferences is based on at least one aspect of the prompt and at least one previous prompt that are received from the user that includes a human user.

172. The method of claim 167 , wherein one or more user preferences is identified, where the resulting item of content is caused to be generated, based on the one or more user preferences.

173. The method of claim 166 , wherein the entirety of the syntactical element portion of the prompt includes an entirety of at least one document and an entirety of a human query.

174. The method of claim 27 , wherein the system comprises:

one or more servers performing the training of the first neural network and the second neural network; and

a client device generating the resulting item of content.

175. The method of claim 174 , wherein the one or more servers is operable such that:

176. The method of claim 27 , wherein the plurality of syntactical elements includes an entirety of a document.

177. The method of claim 27 , wherein the plurality of syntactical elements includes an entirety of a document and a human user query in connection with the document.

178. The method of claim 27 , wherein the plurality of syntactical elements includes an entirety of a plurality of documents.

179. A computer-implemented method, comprising:

within a system:

causing access to a first neural network that is trained by automatic internal application of a first plurality of attentions to representations of different subsets of a first plurality of computer-executable instruction elements, and that generates, by automatic internal application of a second plurality of attentions and utilizing the trained first neural network, a second plurality of computer-executable instruction elements;

causing training of a second neural network by automatic internal application of a third plurality of attentions to representations of different subsets of the second plurality of computer-executable instruction elements;

causing identification of a plurality of syntactical elements;

causing a communication to be sent to a user.

180. The method of claim 179 , wherein:

a first plurality of probabilities is generated, utilizing one or more cognitive computing-based processors, based on at least a portion of the first prioritization and by application of the second neural network, such that the first prioritization results in only a subset of the fourth plurality of attentions corresponding to only a subset of the different subsets of the plurality of syntactical elements being a basis for the generation of the first plurality of probabilities; and

a second plurality of probabilities is generated based on the first plurality of probabilities, such that the third plurality of computer-executable instruction elements is generated based on the second plurality of probabilities.

181. The method of claim 180 , wherein the third plurality of computer-executable instruction elements, including a first one or more computer-executable instruction elements, is generated by the trained second neural network based on a logical evaluation by the trained second neural network for generation, by the trained second neural network, of a fourth plurality of computer-executable instruction elements, including a second one or more computer-executable instruction elements, after the second neural network is trained by applying the third plurality of attentions.

182. The method of claim 181 , wherein the third plurality of computer-executable instruction elements are generated and the logical evaluation is performed automatically without either being directly manually initiated after the trained second neural network is prompted to initiate content generation.

183. The method of claim 182 , wherein third plurality of computer-executable instruction elements includes a modified version of the fourth plurality of computer-executable instruction elements.

184. The method of claim 182 , wherein a setting is received, for causing a tuning of the generation of the third plurality of computer-executable instruction elements based on the setting, such that the setting causes the tuning of the generation of the third plurality of computer-executable instruction elements, by causing a selection of the first one or more computer-executable instruction elements with a first associated probability to be involved with the third plurality of computer-executable instruction elements, instead of another one or more computer-executable instruction elements with a second associated probability, even though the first associated probability is less than the second associated probability.

185. The method of claim 184 , wherein the plurality of syntactical elements does not include any program code.

186. The method of claim 184 , wherein the plurality of syntactical elements includes at least a portion of an executable computer program.

187. The method of claim 186 , wherein the at least portion of executable computer program includes at least a portion of at least one program of the system.

188. The method of claim 186 , wherein the at least portion of executable computer program is automatically generated.

189. The method of claim 182 , wherein the first prioritization of the fourth plurality of attentions is a function of an application of a first process that is based on a first relational aspect among the different subsets of the plurality of syntactical elements and an application of a second process that is based on a second relational aspect among the different subsets of the plurality of syntactical elements, such that the first prioritization is completed in a single operation before any utilization of the first prioritization of any of the fourth plurality of attentions.

190. A system, comprising:

means for causing access to a first neural network that is trained by automatic internal application of a first plurality of attentions to representations of different subsets of a first plurality of items of content, and that generates, by automatic internal application of a second plurality of attentions and utilizing the trained first neural network, a second plurality of items of content;

means for causing training of a second neural network by automatic internal application of a third plurality of attentions to representations of different subsets of the second plurality of items of content; and

means for:

causing identification of a plurality of syntactical elements,

causing the resulting item of content to be sent to a user.

Assignments (4)

ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Feb 12, 2025

From: FLINN, STEVEN D.; MONEYPENNY, NAOMI F.

To: MANYWORLDS, INC.

Reel/Frame 070200/0915 →

ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Feb 12, 2025

From: MANYWORLDS, INC.

To: FLINN, STEVEN D.

Reel/Frame 070200/0924 →

ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Sep 20, 2024

From: FLINN, STEVEN D.; MONEYPENNY, NAOMI F.

To: MANYWORLDS, INC.

Reel/Frame 068645/0201 →

ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Sep 7, 2024

From: MANYWORLDS, INC.

To: FLINN, STEVEN D., MR.

Reel/Frame 068519/0897 →

Continuity (4)

Continuation 18101612 · Jan 26, 2023

Continuation 16660908 · Oct 23, 2019

Continuation 15000011 · Jan 18, 2016

Continuation In Part 14816439 · Aug 3, 2015

References Cited (295)

US 6421655B1 · Horvitz et al. · 2002 [cited by applicant]

US 6795826B2 · Flinn et al. · 2004 [cited by applicant]

US 7375838B2 · Moneypenny et al. · 2008 [cited by applicant]

US 7493294B2 · Flinn et al. · 2009 [cited by applicant]

US 7526458B2 · Flinn et al. · 2009 [cited by applicant]

US 7526459B2 · Flinn et al. · 2009 [cited by applicant]

US 7526464B2 · Flinn et al. · 2009 [cited by applicant]

US 7539652B2 · Flinn et al. · 2009 [cited by applicant]

US 7606772B2 · Flinn et al. · 2009 [cited by applicant]

US 7739231B2 · Flinn et al. · 2010 [cited by applicant]

US 7831535B2 · Flinn et al. · 2010 [cited by applicant]

US 7860811B2 · Flinn et al. · 2010 [cited by applicant]

US 7904341B2 · Flinn et al. · 2011 [cited by applicant]

US 8060462B2 · Flinn et al. · 2011 [cited by applicant]

US RE43768E · Flinn et al. · 2012 [cited by applicant]

US 8380579B2 · Flinn et al. · 2013 [cited by applicant]

US 8458119B2 · Flinn et al. · 2013 [cited by applicant]

US 8458120B2 · Flinn et al. · 2013 [cited by applicant]

US 8478716B2 · Flinn et al. · 2013 [cited by applicant]

US 8515900B2 · Flinn et al. · 2013 [cited by applicant]

US 8515901B2 · Flinn et al. · 2013 [cited by applicant]

US RE44559E · Flinn et al. · 2013 [cited by applicant]

US 8566263B2 · Flinn et al. · 2013 [cited by applicant]

US 8600920B2 · Flynn et al. · 2013 [cited by applicant]

US 8600926B2 · Flinn et al. · 2013 [cited by applicant]

US 8615484B2 · Flinn et al. · 2013 [cited by applicant]

US 8645292B2 · Flinn et al. · 2014 [cited by applicant]

US 8645312B2 · Flinn et al. · 2014 [cited by applicant]

US 8650149B2 · Flinn et al. · 2014 [cited by applicant]

US 8655829B2 · Flinn et al. · 2014 [cited by applicant]

US 8676742B2 · Flinn et al. · 2014 [cited by applicant]

US 8694442B2 · Flinn et al. · 2014 [cited by applicant]

US 8694457B2 · Flinn et al. · 2014 [cited by applicant]

US 8700620B1 · Lieberman et al. · 2014 [cited by applicant]

US 8719213B2 · Flinn et al. · 2014 [cited by applicant]

US RE44966E · Flinn et al. · 2014 [cited by applicant]

US RE44967E · Flinn et al. · 2014 [cited by applicant]

US RE44968E · Flinn et al. · 2014 [cited by applicant]

US 8818917B2 · Flinn et al. · 2014 [cited by applicant]

US 8818931B2 · Flinn et al. · 2014 [cited by applicant]

US 8843433B2 · Flinn et al. · 2014 [cited by applicant]

US 9026488B2 · Flinn et al. · 2015 [cited by applicant]

US 9047245B2 · Flinn et al. · 2015 [cited by applicant]

US RE45770E · Flinn et al. · 2015 [cited by applicant]

US 9159027B2 · Flinn et al. · 2015 [cited by applicant]

US 9165259B2 · Flinn et al. · 2015 [cited by applicant]

US 9171262B2 · Flinn et al. · 2015 [cited by applicant]

US 9171263B2 · Flinn et al. · 2015 [cited by applicant]

US 9436913B2 · Flinn et al. · 2016 [cited by applicant]

US 9454729B2 · Flinn et al. · 2016 [cited by applicant]

US 9454730B2 · Flinn et al. · 2016 [cited by applicant]

US 9906899B2 · Flinn et al. · 2018 [cited by applicant]

US 9924306B2 · Flinn et al. · 2018 [cited by applicant]

US 10510018B2 · Flinn et al. · 2019 [cited by applicant]

US 10699202B2 · Flinn et al. · 2020 [cited by applicant]

US 10733251B2 · Flinn et al. · 2020 [cited by applicant]

US 10783464B2 · Flinn et al. · 2020 [cited by applicant]

US 11481459B2 · Flinn et al. · 2022 [cited by applicant]

US 11593708B2 · Flinn et al. · 2023 [cited by applicant]

US 11715132B2 · Flinn et al. · 2023 [cited by applicant]

US 11960557B2 · Flinn et al. · 2024 [cited by applicant]

US 12093983B2 · Flinn et al. · 2024 [cited by applicant]

US 20010047358A1 · Flinn et al. · 2001 [cited by applicant]

US 20020083025A1 · Robarts et al. · 2002 [cited by applicant]

US 20030234812A1 · Drucker et al. · 2003 [cited by applicant]

US 20050105134A1 · Moneypenny et al. · 2005 [cited by applicant]

US 20060015380A1 · Flinn et al. · 2006 [cited by applicant]

US 20060015381A1 · Flinn et al. · 2006 [cited by applicant]

US 20060155681A1 · Chiang et al. · 2006 [cited by applicant]

US 20060184482A1 · Flinn et al. · 2006 [cited by applicant]

US 20060200432A1 · Flinn et al. · 2006 [cited by applicant]

US 20060200433A1 · Flinn et al. · 2006 [cited by applicant]

US 20060200434A1 · Flinn et al. · 2006 [cited by applicant]

US 20060200435A1 · Flinn et al. · 2006 [cited by applicant]

US 20060259360A1 · Flinn et al. · 2006 [cited by applicant]

US 20060287937A1 · Flinn et al. · 2006 [cited by applicant]

US 20070073553A1 · Flinn et al. · 2007 [cited by applicant]

US 20070073554A1 · Flinn et al. · 2007 [cited by applicant]

US 20070094157A1 · Flinn et al. · 2007 [cited by applicant]

US 20070094216A1 · Mathias et al. · 2007 [cited by applicant]

US 20070112712A1 · Flinn et al. · 2007 [cited by applicant]

US 20070156614A1 · Flinn et al. · 2007 [cited by applicant]

US 20070174220A1 · Flinn et al. · 2007 [cited by applicant]

US 20070203589A1 · Flinn et al. · 2007 [cited by applicant]

US 20070203872A1 · Flinn et al. · 2007 [cited by applicant]

US 20080052288A1 · Flinn et al. · 2008 [cited by applicant]

US 20080059395A1 · Flinn et al. · 2008 [cited by applicant]

US 20080243817A1 · Chan et al. · 2008 [cited by applicant]

US 20080249967A1 · Flinn et al. · 2008 [cited by applicant]

US 20080249968A1 · Flinn et al. · 2008 [cited by applicant]

US 20080288354A1 · Flinn et al. · 2008 [cited by applicant]

US 20090012842A1 · Srinivasan et al. · 2009 [cited by applicant]

US 20090018918A1 · Moneypenny et al. · 2009 [cited by applicant]

US 20090144075A1 · Flinn et al. · 2009 [cited by applicant]

US 20090248599A1 · Hueter et al. · 2009 [cited by applicant]

US 20090319267A1 · Kurki-Suonio · 2009 [cited by applicant]

US 20100217722A1 · Flinn et al. · 2010 [cited by applicant]

US 20100228582A1 · King et al. · 2010 [cited by applicant]

US 20100228693A1 · Dawson et al. · 2010 [cited by applicant]

US 20100235313A1 · Rea et al. · 2010 [cited by applicant]

US 20100274635A1 · Flinn et al. · 2010 [cited by applicant]

US 20110022564A1 · Flinn et al. · 2011 [cited by applicant]

US 20110055017A1 · Solomon et al. · 2011 [cited by applicant]

US 20110066497A1 · Gopinath et al. · 2011 [cited by applicant]

US 20110082825A1 · Sathish et al. · 2011 [cited by applicant]

US 20110093341A1 · Flinn et al. · 2011 [cited by applicant]

US 20110112986A1 · Flinn et al. · 2011 [cited by applicant]

US 20110125700A1 · Funada · 2011 [cited by applicant]

US 20110137849A1 · Flinn et al. · 2011 [cited by applicant]

US 20110153414A1 · Elvekrog et al. · 2011 [cited by applicant]

US 20110153452A1 · Flinn et al. · 2011 [cited by applicant]

US 20110219011A1 · Carmel et al. · 2011 [cited by applicant]

US 20110295663A1 · Flinn et al. · 2011 [cited by applicant]

US 20110295664A1 · Flinn et al. · 2011 [cited by applicant]

US 20110295665A1 · Flinn et al. · 2011 [cited by applicant]

US 20110295696A1 · Flinn et al. · 2011 [cited by applicant]

US 20110295699A1 · Flinn et al. · 2011 [cited by applicant]

US 20110295703A1 · Flinn et al. · 2011 [cited by applicant]

US 20110302121A1 · Flinn et al. · 2011 [cited by applicant]

US 20110302134A1 · Flinn et al. · 2011 [cited by applicant]

US 20110302208A1 · Flinn et al. · 2011 [cited by applicant]

US 20110302209A1 · Flinn et al. · 2011 [cited by applicant]

US 20120030028A1 · Flinn et al. · 2012 [cited by applicant]

US 20120030029A1 · Flinn et al. · 2012 [cited by applicant]

US 20120030030A1 · Flinn et al. · 2012 [cited by applicant]

US 20120030063A1 · Flinn et al. · 2012 [cited by applicant]

US 20120030064A1 · Flinn et al. · 2012 [cited by applicant]

US 20120030620A1 · Flinn et al. · 2012 [cited by applicant]

US 20120030621A1 · Flinn et al. · 2012 [cited by applicant]

US 20120035969A1 · Flinn et al. · 2012 [cited by applicant]

US 20120035970A1 · Flinn et al. · 2012 [cited by applicant]

US 20120035976A1 · Flinn et al. · 2012 [cited by applicant]

US 20120035985A1 · Flinn et al. · 2012 [cited by applicant]

US 20120036020A1 · Flinn et al. · 2012 [cited by applicant]

US 20120054148A1 · Flinn et al. · 2012 [cited by applicant]

US 20120089915A1 · Flinn et al. · 2012 [cited by applicant]

US 20120096155A1 · Flinn et al. · 2012 [cited by applicant]

US 20120101976A1 · Flinn et al. · 2012 [cited by applicant]

US 20120254074A1 · Flinn et al. · 2012 [cited by applicant]

US 20120254075A1 · Flinn et al. · 2012 [cited by applicant]

US 20120254079A1 · Flinn et al. · 2012 [cited by applicant]

US 20120254094A1 · Flinn et al. · 2012 [cited by applicant]

US 20120254095A1 · Flinn et al. · 2012 [cited by applicant]

US 20120254096A1 · Flinn et al. · 2012 [cited by applicant]

US 20120254097A1 · Flinn et al. · 2012 [cited by applicant]

US 20120254098A1 · Flinn et al. · 2012 [cited by applicant]

US 20120254099A1 · Flinn et al. · 2012 [cited by applicant]

US 20120290518A1 · Flinn · 2012 [cited by examiner]

US 20120324565A1 · Cohen et al. · 2012 [cited by applicant]

US 20130024464A1 · Berner et al. · 2013 [cited by applicant]

US 20130165234A1 · Hall et al. · 2013 [cited by applicant]

US 20130191214A1 · Flinn et al. · 2013 [cited by applicant]

US 20130311501A1 · Flinn et al. · 2013 [cited by applicant]

US 20130339450A1 · Flinn et al. · 2013 [cited by applicant]

US 20140025660A1 · Mohammed et al. · 2014 [cited by applicant]

US 20140082082A1 · Flinn et al. · 2014 [cited by applicant]

US 20140108325A1 · Flinn et al. · 2014 [cited by applicant]

US 20140129504A1 · Soon-Shiong · 2014 [cited by applicant]

US 20140195464A1 · Flinn et al. · 2014 [cited by applicant]

US 20140207582A1 · Flinn et al. · 2014 [cited by applicant]

US 20140358832A1 · Flinn et al. · 2014 [cited by applicant]

US 20140358842A1 · Flinn et al. · 2014 [cited by applicant]

US 20140365421A1 · Flinn et al. · 2014 [cited by applicant]

US 20150095278A1 · Flinn et al. · 2015 [cited by applicant]

US 20150213367A1 · Flinn et al. · 2015 [cited by applicant]

US 20150230052A1 · Flinn et al. · 2015 [cited by applicant]

US 20150339573A1 · Flinn et al. · 2015 [cited by applicant]

US 20160004966A1 · Flinn et al. · 2016 [cited by applicant]

US 20160012343A1 · Flinn et al. · 2016 [cited by applicant]

US 20160012344A1 · Flinn et al. · 2016 [cited by applicant]

US 20160132789A1 · Flinn et al. · 2016 [cited by applicant]

US 20160232440A1 · Gregor et al. · 2016 [cited by applicant]

US 20160342907A1 · Flinn et al. · 2016 [cited by applicant]

US 20160364651A1 · Flinn et al. · 2016 [cited by applicant]

US 20170039174A1 · Strope et al. · 2017 [cited by applicant]

US 20170076222A1 · Khapra et al. · 2017 [cited by applicant]

US 20170124433A1 · Chandraker et al. · 2017 [cited by applicant]

US 20170140271A1 · Reed et al. · 2017 [cited by applicant]

US 20180075459A1 · Flinn et al. · 2018 [cited by applicant]

US 20180146328A1 · Flinn et al. · 2018 [cited by applicant]

US 20180199025A1 · Holzer et al. · 2018 [cited by applicant]

US 20180225598A1 · Flinn et al. · 2018 [cited by applicant]

US 20180225599A1 · Flinn et al. · 2018 [cited by applicant]

US 20180232669A1 · Flinn et al. · 2018 [cited by applicant]

US 20180232670A1 · Flinn et al. · 2018 [cited by applicant]

US 20180232671A1 · Flinn et al. · 2018 [cited by applicant]

US 20180314992A1 · Flinn et al. · 2018 [cited by applicant]

US 20180341860A1 · Shazeer et al. · 2018 [cited by applicant]

US 20190325358A1 · Flinn et al. · 2019 [cited by applicant]

US 20190325359A1 · Flinn et al. · 2019 [cited by applicant]

US 20190378060A1 · Flinn et al. · 2019 [cited by applicant]

US 20200065709A1 · Flinn et al. · 2020 [cited by applicant]

US 20200074335A1 · Flinn et al. · 2020 [cited by applicant]

US 20200286009A1 · Flinn · 2020 [cited by applicant]

US 20200327359A1 · Blundell et al. · 2020 [cited by applicant]

US 20200334310A1 · Flinn et al. · 2020 [cited by applicant]

US 20230004609A1 · Flinn et al. · 2023 [cited by applicant]

US 20230146960A1 · Flinn et al. · 2023 [cited by applicant]

US 20240046311A1 · Flinn et al. · 2024 [cited by applicant]

CN 105159890A · 2015 [cited by applicant]

CN 105868829A · 2016 [cited by applicant]

CN 106462802A · 2017 [cited by applicant]

EP 1311980A2 · 2003 [cited by applicant]

EP 1397252A1 · 2004 [cited by applicant]

EP 1397252A4 · 2005 [cited by applicant]

EP 1630700A1 · 2006 [cited by applicant]

EP 1630701A1 · 2006 [cited by applicant]

EP 1311980B1 · 2006 [cited by applicant]

KR 1020150037986 · 2015 [cited by applicant]

TW 501074B · 2002 [cited by applicant]

You, Quanzeng, et al. “Robust image sentiment analysis using progressively trained and domain transferred deep networks.” Proceedings of the AAAI conference on Artificial Intelligence. vol. 29. No. 1. Feb. 9, 2015. (Yea… [cited by examiner]

Luong et al., “Effective Approaches to Attention-based Neural Machine Translation,” (Sep. 20, 2015) Conf. on Empirical Methods in Natural Language Processing at p. 1412-1421. [cited by applicant]

Sukhbaatar et al., “End-to-End Memory Networks,” (2015) available from Internet < https://arxiv.org/pdf/1503.08895.pdf> at p. 1-11. [cited by applicant]

Cho et al. “Learning phrase representations using rnn encode··decoder for statistical machine translation,” (Sep. 3, 2014), available from Internet < //efaidnbmnnnibpcajpcglclefindmkaj/https://arxiv.org/pdf/1406.1078 > … [cited by applicant]

Chung et al. “Empirical evaluation of gated recurrent neural networks on sequence modeling,” available from Internet arXiv 1412.3555vl, Dec. 11, 2014, 9 pages. [cited by applicant]

Hochreiter et al. “Gradient flow in recurrent nets: the difficulty of learning long-term dependencies, A field Guide to Dynamical Recurrent Neural Networks,” IEEE Press, 2001, 15 pages. [cited by applicant]

Hochreiter et al. “Long short term memory,” Neural Computation 9(8), Nov. 1997, 46 pages. [cited by applicant]

Kingma et al. “Adam: A method for stochastic optimization,” available from Internet < arXiv 1412.6980v8 > Jul. 23, 2015, 15 pages. [cited by applicant]

Luong et al. “Effective approaches to attention based neural machine translation,” arXiv 1508.04025v2, Sep. 20, 2015, 11 pages. [cited by applicant]

Srivastava et al. “Dropout: a simple way to prevent neural network from overfitting,” Journal of Machine Learning Research, 15(1), Jan. 2014, 30 pages. [cited by applicant]

Sutskever et al. “Sequence to sequence learning with neural networks,” Advances in Neural Information Processing Systems, Dec. 2014, 9 pages. [cited by applicant]

US Final Rejection Office Action in Publication No. 2015-0339573A1, dated Jan. 9, 2020, 32 Pages. [cited by applicant]

US Non-Final Rejection Office Action in Publication No. 2015-0339573A1, dated May 23, 2019, 30 Pages. [cited by applicant]

US Non-Final Rejection Office Action in Publication No. 2015-0095278A1 dated Nov. 1, 2017, 66 Pages. [cited by applicant]

US Final Rejection Office Action in Publication No. 2015-0095278A1 dated Apr. 30, 2018, 64 Pages. [cited by applicant]

US Examiner's Answer to Appeal Brief Office Action in Publication No. 2015-0095278A1 dated Sep. 18, 2018, 86 Pages. [cited by applicant]

US Patent Board Decision Office Action in Publication No. 2015-0095278A1 dated Feb. 26, 2020, 20 Pages. [cited by applicant]

D. Bahdanau, K. Cho, and Y. Bengio, “Neural machine translation by jointly learning to align and translate,” May 19, 2016, 15 pages., retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://arxiv.org/pdf/1409.0473. [cited by applicant]

Christian Buck, Kenneth Heafield, and Bas van Ooyen, “N-gram counts and language models from the common crawl,” Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC'14), May 2014,… [cited by applicant]

Kyunghyun Cho, Bart van Merrienboer, Caglar Gulcehre, Fethi Bougares, Holger Schwenk, and Yoshua Bengio., “Learning phrase representations using RNN encoder-decoder for statistical machine translation,” Sep. 3, 2014, 15… [cited by applicant]

Alexander Fraser and Daniel Marcu, “Measuring word alignment quality for statistical machine translation,” Computational Linguistics, 33(3):293-303, Sep. 2007, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://a… [cited by applicant]

Karol Gregor, Ivo Danihelka, Alex Graves, Danilo Jimenez Rezende, and Daan Wierstra, “DRAW: A recurrent neural network for image generation,” In ICML, May 20, 2015, 10 pages, retrieved from //efaidnbmnnnibpcajpcglclefin… [cited by applicant]

Sebastien Jean, Kyunghyun Cho, Roland Memisevic, and Yoshua Bengio., “On using very large target vocabulary for neural machine translation,” Mar. 18, 2015, In ACL, 10 pages, retrieved from //efaidnbmnnnibpcajpcglclefind… [cited by applicant]

N. Kalchbrenner and P. Blunsom, “Recurrent continuous translation models,” EMNLP, 2013., 10 pages, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://aclanthology.org/D13-1176.pdf. [cited by applicant]

Philipp Koehn, Franz Josef Och, and Daniel Marcu, “Statistical phrase-based translation,” In NAACL., 2003, 7 pages, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://aclanthology.org/N03-1017.pdf. [cited by applicant]

P. Liang, B. Taskar, and D. Klein, “Alignment by agreement,” In NAACL, 2006, 8 pages, retrieved from c//efaidnbmnnnibpcajpcglclefindmkaj/https://aclanthology.org/N06-1014.pdf. [cited by applicant]

M.-T. Luong, I. Sutskever, Q. V. Le, O. Vinyals, and W. Zaremba, “Addressing the rare word problem in neural machine translation,” 2015, In ACL, 9 pages, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://arxiv.o… [cited by applicant]

Volodymyr Mnih, Nicolas Heess, Alex Graves, and Koray Kavukcuoglu, “Recurrent models of visual attention,” In NIPS, 2014, 9 pages, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://proceedings.neurips.cc/paper_f… [cited by applicant]

Kishore Papineni, Salim Roukos, Todd Ward, and Wei jing Zhu, “Bleu: a method for automatic evaluation of machine translation,” In ACL, 2002, 8 pages, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://aclantholog… [cited by applicant]

I. Sutskever, O. Vinyals, and Q. V. Le, “Sequence to sequence learning with neural networks,” In NIPS, 2014, 9 pages, retrieved from c//efaidnbmnnnibpcajpcglclefindmkaj/https://arxiv.org/pdf/1409.3215. [cited by applicant]

Kelvin Xu, Jimmy Ba, Ryan Kiros, Kyunghyun Cho, Aaron C. Courville, Ruslan Salakhutdinov, Richard S. Zemel, and Yoshua Bengio, “Show, attend and tell: Neural image caption generation with visual attention,” In ICML, 201… [cited by applicant]

Wojciech Zaremba, Ilya Sutskever, and Oriol Vinyals, “Recurrent neural network regularization,” In ICLR, 2015, 8 pages, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://arxiv.org/pdf/1409.2329. [cited by applicant]

Bengio, Y., Ducharme, R., Vincent, P., and Janvin, C, “A neural probabilistic language model,” The Journal of Machine Learning Research, 3:1137-1155, 2003, 19 pages, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/htt… [cited by applicant]

Hochreiter, S. and Schmidhuber, J, “Long short-term memory,” Neural Computation, 1997, 32 pages, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://www.bioinf.jku.at/publications/older/2604.pdf. [cited by applicant]

Jurafsky, D. and Martin, J, “Speech and language processing,” Pearson International, 2009, 599 pages, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://web.stanford.edu/˜jurafsky/slp3/ed3book.pdf. [cited by applicant]

Lester, J., Branting, K., and Mott, B. “Conversational agents,” In Handbook of Internet Computing. Chapman & Hall, 2004, 19 pages. [cited by applicant]

Mikolov, T. “Statistical Language Models based on Neural Networks,” PhD thesis, Brno University of Technology, 2012, 59 pages, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://www.fit.vut.cz/person/imikolov/pub… [cited by applicant]

Mikolov, T., Karafiat, M., Burget, L., Cernocky, J., and Khudanpur, S. “Recurrent neural network based language model,” In INTERSPEECH, pp. 1045-1048, 2010. [cited by applicant]

Shang, L., Lu, Z., and Li, H. “Neural responding machine for short-text conversation,” In Proceedings of ACL, 2015, 12 pages, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://arxiv.org/pdf/1503.02364. [cited by applicant]

Sordoni, A., Galley, M., Auli, M., Brockett, C., Ji, Y., Mitchell, M., Gao, J., Dolan, B., and Nie, J.-Y. “A neural network approach to context-sensitive generation of conversational responses,” In Proceedings of NAACL,… [cited by applicant]

Tiedemann, J. “News from OPUS—A collection of multilingual parallel corpora with tools and interfaces,” In Advances in Natural Language Processing (pp. 237-248), Jan. 2009, Chapter V, Publisher: John Benjamins, Amsterda… [cited by applicant]

Turing, A. M. “Computing machinery and intelligence,” Mind, pp. 433-460, 1950. [cited by applicant]

Vinyals, O., Kaiser, L., Koo, T., Petrov, S., Sutskever, I., and Hinton, G. “Grammar as a foreign language.,” 2015, 10 pages, retrieved from //efaidnbmnnnibpcajpcglclefindmkaj/https://arxiv.org/pdf/1412.7449. [cited by applicant]

Vinyals, O., Toshev, A., Bengio, S., and Erhan, D. “Show and tell: A neural image caption generator,” Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, 9 pages, retrieved from //efaidn… [cited by applicant]

Will, T. “Creating a Dynamic Speech Dialogue.,” VDM Verlag Dr, 2007, 1 page. [cited by applicant]

Cao et al., Object Spatial Recognition for Service Robots: Where is the Front?, IEEE Int'l Conference on Mechatronics and Automation (2012) at pp. 875-880 (Year: 2011).* [cited by applicant]

Krizhevsky et al., ImageNet Classification with Deep Convolutional Neural Networks, In NIPS, pp. 1106-1114 (2012) (Year: 2012).* [cited by applicant]

Gupta et al., Observing Human-Object Interactions: Using Spatial and Functional Compatibility for Recognition, Transactions on Pattern Analysis and Mach. Intel., vol. 31 at pp. 1775-1789 (2009) (Year: 2009).* [cited by applicant]

Berenson et al., Manipulation Planning with Workspace Goal Regions, Int'l Conference on Robotics and Automation (2009) at pp. 618-624 (Year: 2009).* [cited by applicant]

Rissland & Ashley, Hypotheticals as Heuristic Device, Proc. of the 5th Annual Conference on Artificial Intelligence (1986) at pp. 165-178 (Year: 1986).* [cited by applicant]

Wallace, Richard S. “The anatomy of ALICE.” Parsing the Turing Test. Springer, Dordrecht, 2009. 181-210. (Year: 2009).* [cited by applicant]

Using weak ties to understand resource usage behaviors in an online community of educators Ogheneovo Dibie; Tamara Sumner 2015 IEEE/ACM International Conference on Advances in Social Network Analysis and Mining (ASONAM)… [cited by applicant]

Uberveillance and the Internet of Things and People M.G. Michael; Katina Michael; Christine Perakslis, Contemporary Computing and Informatics (IC3I), 2014 International Conference on Year: 2014 pp. 1381-1386, DOI: 10.11… [cited by applicant]

Inferring User Relationship from Hidden Information in WLAN s Ningning Cheng; Prasant Mohapatra; Mathieu Cunche; Mohamed Ali Kaafar; Roksana Boreli; Srikanth Krishnamurthy MILCOM 2012-2012 IEEE Military Communications C… [cited by applicant]

Financial Time Series Prediction Using Least Squares Support Vector Machines Within the Evidence Framework T. Van Gestel; J. A. K. Suykens; D.E. Baestaens; A. Lambrechts; G. Lanckriet; B. Vandaele; B. De Moor; J. Vandew… [cited by applicant]

Opinion Formation in Ising Networks M.H. Afrasiabi; R. Guerin; S.S.Venkatesh Information Theory and Applications Workshop (ITA), 2013 Year: 2013 pp. 1-10, Doi: 10.1109/ITA.2013.6502982 IEEE Conference Publications. [cited by applicant]

MsLRR: A Unified Multiscale Low-Rank Representation for Image Segmentation Xiaobai Liu; Qian Xu; Jiayi Ma; Hai iin; Yanduo Zhang IEEE Transactions on Image Processing Year: 2014, vol. 23, Issue 5 pp. 2159-2167, DOI: 10.… [cited by applicant]

A Two-Layer Text Clustering Approach for Retrospective News Event Detection Xianying Dai; Yancheng He; Yunlian Sun Artificial Intelligence and Computational Intelligence (AICI), 2010 International Conference on Year: 20… [cited by applicant]

Synergizing people, process, and technology to motivate knowledge sharing and collaboration industry case study Pickering, C.K. Collaboration Technologies and Systems (CTS), 2013 International Conference on Year: 2013 p… [cited by applicant]

Temporally adaptive motion estimation using temporal subband analysis Zung Kon Yim; Jae Ho Chung Circuits and Systems, 1996. ISCAS '96., Connecting the World., 1996 IEEE International Symposium on Year: 1996, vol. 2 pp.… [cited by applicant]

Leveraging temporal, contextual and ordering constraints or recognizing complex activities in video Laxton, B; Jongwoo Lim; Krieg·man, D Computer vision and Pattern Recognition, 2007. CVPR '07 IEEE Conference on Year: 2… [cited by applicant]

An implicit spatiotemporal shape model for human activity localization and recognition Oikonomopoulos, A; Patras, I; Pantie, M. Computer Vision and Pattern Recognition Workshops, 2009, CVPR Workshops 2009. IEEE Computer… [cited by applicant]

Relationships Among Distinct Models and Notions of Equivalence for Stochastic Finite-State Systems De Renna E, C.; Leake, R.J. Computers, IEEE Transactions on Year: 1969, vol. C-18, Issue: 7 pp. 633-641, DOI: 10.1109/T-… [cited by applicant]

Study on relationship between optical signals and charge quantity of partial discharge under four typical insulation defects Zhuo Ran; Tang Ju; Fan Lei; Liu Yonggng; Zhang Xiaoxing. Electrical Insulation and Dielectric … [cited by applicant]

Analysis of spatial-temporal regularization methods for linear inverse problems from a common statistical framework Yiheng Zhang; Ghodrati, Alizera; Brooks, D.H. Biomedical Imaging: Nano to Macro, 2004. IEEE Internation… [cited by applicant]

On the Spectral Characterization and Scalable Mining of Network Communities Bo Yang; Jiming Liu; Jianfeng Feng Knowledge and Data Engineering, IEEE Transactions on Year: 2012, vol. 24, Issue: 2 pp. 326-337, DOI: 10.1109… [cited by applicant]

Automated Building of Domain Ontologies from Lecture Notes in Courseware Gantayat, N.; lyer, S. Technology for Education (T4E), 2011 IEEE International Conference on Year: 2011 pp. 89-95, DOI: 10.1109/T4E.2011.22 Refere… [cited by applicant]

Use of usability evaluation methods in France: The reality in professional practices Roche, A.; Lespinet-Njib, V.; Andre, J.-M. User Science and Engineering (i-USEr), 2014 3rd International Conference on Year: 2014 pp. … [cited by applicant]

Modeling the Risk & Utility of Information Sharing in Social Networks Fouad, M.R.; Elbassioni, K.; Bertino, E. Privacy, Security, Risk and Trust (PAS SAT), 2012 International Conference on Social Computing (SocialCom) Y… [cited by applicant]

Droid Analytics: A Signature Based Analytic System to Collect, Extract, Analyze and Associate Android Malware Min Zheng; Mingshen Sun; Lui, J.C.S. Trust, Security, and Privacy in Computing and Communications (Trustcom),… [cited by applicant]

Exploiting Similarities among Languages for Machine Translation. Mikolov, T.; Quoc Le; Sutskever, I. arXiv: 1309.4168vl [cs.CL] Sep. 17, 2013 http://arxiv.org/abs/1309.4168. [cited by applicant]

Your Brain on Metaphors. Chorost, M. The Chronicle of Higher Education. Sep. 1, 2014. http://chronicle.com/article/ Your-Brain·on-Metaphors/148495/. [cited by applicant]

Building Watson: An Overview of the DeepQA Project. Ferrucci, D. et al. AI Magazine. vol. 31 No. 3 Fall 2010 pp. 59-79. Association for the Advancement of Artificial Intelligence http://www.aaai.org/ oj s/index.php/ air… [cited by applicant]

Semantic Extraction from Texts. Jusoh, S.; AI Fawareh, H. 2009 International Conference on Computer Engineering and Applications IPCSIT vol. 2 (2011) pp. 520-525 IACSIT Press, Singapore. [cited by applicant]

A semantic approach for text clustering using WordNet and lexical chains. Tingting Wei;, Yonghe Lu; Huiyou Chang; Qiang Zhou; Xianyu Bao. Expert Systems with Applications vol. 42, Issue 4, Mar. 2015, pp. 2264-2275 DOI: … [cited by applicant]

A Neural Conversational Model. Vinyals, O.; Quoc Le. Proceedings of the 31st International Conference on Machine Learning, Lille, France, 2015. JMLR: W&CP vol. 37 arXiv: 1506.05869v2 [cs.CL] Jun. 23, 2015 http://andv.or… [cited by applicant]

Human-level concept learning through probabilistic program induction. Lake, B.; Salakhutdinov, R.; Tenenbaum, J. Science Dec. 11, 2015: vol. 350, Issue 6266, pp. 1332-1338, DOI: 10.1126/science. aab3050. [cited by applicant]

Torreano et al., When Dogs Can Fly: Level of Abstraction as a Cue to Metaphorical Use of Verbs, Metaphor and Symbol, 20(4), 259-274 (2005). [cited by applicant]

Garrell et al., Proactive Behavior of an Autonomous Mobile Robot for Human-Assisted Learning, 2013 IEEE RO-MAN: The 22nd IEEE International Symposium on Robot and Human Interactive Communication, Aug. 26-29, 2013. [cited by applicant]

Pelachaud et al., Expressive Gestures Displayed by a Humanoid Robot during a Storytelling Application, Jan. 2010. [cited by applicant]

Gandy et al., Automatic Identification of Conceptual Metaphors with Limited Knowledge, Proceedings of the Twenty-Seventh AAAI Conference on Artificial Intelligence, 2013. [cited by applicant]

Bouayad-Agha et al., Natural Language Generation and Semantic Web Technologies, IOS Press, 2012. [cited by applicant]

Karpathy, Andrej, and Li Fei-Fei. “Deep visual-semantic alignments for generating image descriptions.” Proceedings of the IEEE conference on computer vision and pattern recognition. 2015. (Year: 2015).* [cited by applicant]

Mao, Junhua, et al. “Deep captioning with multimodal recurrent neural networks (m-rnn).” arXiv preprint arXiv: 1412.6632 (2014). (Year: 2014).* [cited by applicant]

Dong, Li, and Mirella Lapata. “Language to logical form with neural attention.” arXiv preprint arXiv: 1601.01280v1 (2016). (Year: 2016). [cited by applicant]