DETAILED ACTION
This action is in response to the initial filing of application no. 18/951,121 on 11/18/2024.
Claims 21- 40 are still pending in this application, with claims 21 and 31 being independent.
Allowable Subject Matter
Aside from the non-prior art rejection, it has been determined that that prior art fails to teach or suggest in reasonable combination the limitations recited by claims 24 and 34. Claims 24 and 34 recite the following.
receiving input data captured by the apparatus; processing the input data, using the at least one neural network of the control system, to obtain an output, the at least one neural network of the control system having at least one hidden layer; determining, using a second neural network having at least one hidden layer, if a confidence level associated with the output is lower than a predetermined level; controlling the apparatus using the output if the confidence level is not lower than the predetermined level; and re-training the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network, and wherein the second neural network is prevented from being re-trained.
Specifically, the prior art fails to teach or suggest: re-training the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network, and wherein the second neural network is prevented from being re-trained. For example, Hu discloses training both the first neural network ([0050] [0055 – 0058]) and the second neural network ([0051] [0052] [0055 – 0060]). Yet, Hu fails to teach or suggest retraining the first neural network while preventing the second neural network from being retraining.
Furthermore, Paquet discloses training a neural network based on a confidence level associated with output generated by the neural network (as discussed below). Yet, Paquet fails to teach or suggest retraining a first neural network while preventing a second neural network from being retraining.
Aside from the non-prior art rejection, it has been determined that that prior art fails to teach or suggest in reasonable combination the limitations recited by claims 25 and 35. Claims 25 and 35 recite the following.
receiving input data captured by the apparatus; processing the input data, using the at least one neural network of the control system, to obtain an output, the at least one neural network of the control system having at least one hidden layer; determining, using a second neural network having at least one hidden layer, if a confidence level associated with the output is lower than a predetermined level; controlling the apparatus using the output if the confidence level is not lower than the predetermined level; and re-training the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network, and wherein the re-training utilizes reinforcement training.
As discussed below, the combination of Hu and Paquet teach or suggest supervised re-training instead of retraining utilizing reinforcement training.
Aside from the non-prior art rejection, it has been determined that that prior art fails to teach or suggest in reasonable combination the limitations recited by claims 27 and 37. Claims 27 and 37 recite the following.
receiving input data captured by the apparatus; processing the input data, using the at least one neural network of the control system, to obtain an output, the at least one neural network of the control system having at least one hidden layer; determining, using a second neural network having at least one hidden layer, if a confidence level associated with the output is lower than a predetermined level; controlling the apparatus using the output if the confidence level is not lower than the predetermined level; and re-training the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network, and wherein the re-training utilizes unsupervised learning.
As discussed below, the combination of Hu and Paquet teach or suggest supervised re-training instead of retraining utilizing reinforcement training unsupervised learning.
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP § 2146 et seq. for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The filing of a terminal disclaimer by itself is not a complete reply to a nonstatutory double patenting (NSDP) rejection. A complete reply requires that the terminal disclaimer be accompanied by a reply requesting reconsideration of the prior Office action. Even where the NSDP rejection is provisional the reply must be complete. See MPEP § 804, subsection I.B.1. For a reply to a non-final Office action, see 37 CFR 1.111(a). For a reply to final Office action, see 37 CFR 1.113(c). A request for reconsideration while not provided for in 37 CFR 1.113(c) may be filed after final for consideration. See MPEP §§ 706.07(e) and 714.13.
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The actual filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/apply/applying-online/eterminal-disclaimer.
The claim mapping is as follows regarding parent application US 12,174,631.
Current Application
21. (New) A method of operating an apparatus using a control system that includes at least one neural network, comprising the steps of: receiving input data captured by the apparatus; processing the input data, using the at least one neural network of the control system, to obtain an output, the at least one neural network of the control system having at least one hidden layer; determining, using a second neural network having at least one hidden layer, if a confidence level associated with the output is lower than a predetermined level; and controlling the apparatus using the output if the confidence level is not lower than the predetermined level.
23. (New) The method of claim 21, further comprising the step of: re-training the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
24. (New) The method of claim 23, wherein the second neural network is prevented from being re-trained.
25. (New) The method of claim 23, wherein the re-training utilizes reinforcement training.
26. (New) The method of claim 23, wherein the re-training utilizes supervised learning.
27. (New) The method of claim 23, wherein the re-training utilizes unsupervised learning.
31. (New) An apparatus being operated in part by a controller, comprising: an input device constructed to generate input data; at least one neural network coupled to the controller and constructed to receive the input data and to generate an output, the at least one neural network having at least one hidden layer; and a second neural network with at least one hidden layer constructed to use the output from the at least one neural network to determine whether a confidence level associated with the output is below a predetermined level, wherein the controller is further constructed to operate the apparatus using the output unless the confidence level of the output from the at least one neural network is below the predetermined level.
33. (New) The apparatus of claim 31, wherein the controller is further structured to re- train the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
34. (New) The apparatus of claim 33, wherein the second neural network is prevented from being re-trained.
35. (New) The apparatus of claim 33, wherein the re-training utilizes reinforcement training.
36. (New) The apparatus of claim 33, wherein the re-training utilizes supervised learning.
37. (New) The apparatus of claim 33, wherein the re-training utilizes unsupervised learning.
US 12,174,631
1. A method of operating an apparatus using a control system that includes at least one neural network, comprising the steps of: receiving input data captured by the apparatus; processing the input data using the at least one neural network of the control system, the at least one neural network having at least one hidden layer; obtaining an output from the at least one neural network resulting from processing the input data; and receiving, using a second neural network having at least one hidden layer, the output from the at least one neural network; generating, using the second neural network, a confidence level of the output; and controlling the apparatus using the output if the confidence level is not lower than the predetermined level.
11. The method of claim 1, further comprising re-training the at least one neural network if the confidence level of the output is lower than the predetermined level.
12. A system for controlling a plurality of machines, comprising: a controller; an apparatus in communication with the controller and being operated in part by the controller, the apparatus comprising: an input device constructed to generate input data, and at least one neural network constructed to receive the input data and to generate an output, the at least one neural network having at least one hidden layer; and a second neural network with at least one hidden layer constructed to use the output from the at least one neural network to generate a confidence level of the output, wherein the controller is constructed to operate the apparatus using the output unless the confidence level of the output is below a predetermined level.
20. The system of claim 10, wherein the at least one neural network was retrained using locally collected training data from at least one autonomous machine.
Claims 21 and 31 are rejected on the ground of nonstatutory double patenting as being unpatentable over claims 1 and 12 of U.S. Patent No.12,174,631, respectively. Although the claims at issue are not identical, they are not patentably distinct from each other.
Regarding claim 21 of the current application, claim 1 of US 12,164,631 recites the limitations of claim 21 except for: determining if a confidence level associated with the output is lower than a predetermined level. However, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 1 of US 12,164,631.
Thus, the limitations of claim 21 of the current application are obvious variants of the limitations of claim 1 of 12,164, 631. And, claim 21 of the current application and claim 1 of 12,164,631 are not patentably distinct.
Regarding claim 31 of the current application, claim 12 of US 12,164,631 recites the limitations of claim 31 except for: determining if a confidence level associated with the output is lower than a predetermined level. However, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 1 of US 12,164,631.
Thus, the limitations of claim 31 and the current application are obvious variants of the limitations of claim 12 of 12,164, 631. And, claim 31 of the current application and claim 12 of 12,164,631 are not patentably distinct.
Claims 23 and 24 are rejected on the ground of nonstatutory double patenting as being unpatentable over claim 11 of U.S. Patent No. 12,174,631 in view of Paquet et al. (US 2012/0158620) (“Paquet”). Although the claims at issue are not identical, they are not patentably distinct from each other.
Regarding claim 23 of the current application, claim 11 of US 12,164,631 recites the limitations of claim 23 except for: determining if a confidence level associated with the output is lower than a predetermined level; and re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
However, Paquet discloses a system and method for training automated classifiers (artificial neural networks) (Abstract), comprising executing the following on a processor (Fig.5, 64): re-training an artificial neural network (automated classifier) when a confidence level (score) is below a pre-determined level (Fig.4, 12, 14, 16, 18, 32, 54 and Fig.5, 70, 72, 24, 76; [0030] [0031]), wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (image data, [0021]) that is different from what was originally used to train the at least one neural network (The training data used to retrain the artificial neural network comprises content set (content item A and content item C) which is different than the content set (content item1, content item 2, content item 3) used to train the artificial network., Fig. 4, initial training set and supplemental training set; [0030]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 11 of 12,174,631 in the same way that Paquet’s invention has been improved to achieve the following, predictable results for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006]): re-training the at least one neural network further comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 11 of US 12,164,631.
Thus, the limitations of claim 23 of the current application are obvious variants of the limitations of claim 11 of 12,164, 631. And, claim 23 of the current application and claim 11 of 12,164,631 are not patentably distinct.
Regarding claim 24 of the current application, claim 11 of US 12,164,631 recites the limitations of claim 24 except for: determining if a confidence level associated with the output is lower than a predetermined level; and re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network., wherein the second neural network is prevented from being re-trained.
However, Paquet discloses a system and method for training automated classifiers (artificial neural networks) (Abstract), comprising executing the following on a processor (Fig.5, 64): re-training an artificial neural network (automated classifier) when a confidence level (score) is below a pre-determined level (Fig.4, 12, 14, 16, 18, 32, 54 and Fig.5, 70, 72, 24, 76; [0030] [0031]), wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (image data, [0021]) that is different from what was originally used to train the at least one neural network (The training data used to retrain the artificial neural network comprises content set (content item A and content item C) which is different than the content set (content item1, content item 2, content item 3) used to train the artificial network., Fig. 4, initial training set and supplemental training set; [0030]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 11 of 12,174,631 in the same way that Paquet’s invention has been improved to achieve the following, predictable results for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006]): re-training the at least one neural network further comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 11 of US 12,164,631. Furthermore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing that by only re-training the at least one neural network as recited by claim 11 of US 12,164,631, the second neural network is prevented from being re-trained.
Thus, the limitations of claim 24 of the current application are obvious variants of the limitations of claim 11 of 12,164, 631. And, claim 24 of the current application and claim 11 of 12,164,631 are not patentably distinct.
Claims 25 - 27 are rejected on the ground of nonstatutory double patenting as being unpatentable over claim 11 of U.S. Patent No. 12,174,631 in view of Paquet et al. (US 2012/0158620) (“Paquet”) and further in view of Cardinaux et al. (US 2015/0278686) (“Cardinaux”) . Although the claims at issue are not identical, they are not patentably distinct from each other.
Regarding claim 25 of the current application, claim 11 of US 12,164,631 recites the limitations of claim 25 except for: determining if a confidence level associated with the output is lower than a predetermined level; and re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network, wherein the re-training utilizes reinforcement training.
However, Paquet discloses a system and method for training automated classifiers (artificial neural networks) (Abstract), comprising executing the following on a processor (Fig.5, 64): re-training an artificial neural network (automated classifier) when a confidence level (score) is below a pre-determined level (Fig.4, 12, 14, 16, 18, 32, 54 and Fig.5, 70, 72, 24, 76; [0030] [0031]), wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (image data, [0021]) that is different from what was originally used to train the at least one neural network (The training data used to retrain the artificial neural network comprises content set (content item A and content item C) which is different than the content set (content item1, content item 2, content item 3) used to train the artificial network., Fig. 4, initial training set and supplemental training set; [0030]).
Additionally, Cardinaux discloses a system, method and artificial neural network (Abstract), wherein the artificial neural network is trained using reinforcement training ([0021]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 11 of 12,174,631 in the same way that Paquet’s invention has been improved to achieve the following, predictable results for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006]): re-training the at least one neural network further comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to modify the teachings disclosed by the combination of the limitations recited by claim 11 of 12,174,631 and Paquet’s invention with Cardinaux’s teachings so that the re-training (training) comprises reinforcement training for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006])
Moreover, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 11 of US 12,164,631.
Thus, the limitations of claim 25 of the current application are obvious variants of the limitations of claim 11 of 12,164, 631. And, claim 25 of the current application and claim 11 of 12,164,631 are not patentably distinct.
Regarding claim 26 of the current application, claim 11 of US 12,164,631 recites the limitations of claim 26 except for: determining if a confidence level associated with the output is lower than a predetermined level; and re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network, wherein the re-training utilizes supervised learning.
However, Paquet discloses a system and method for training automated classifiers (artificial neural networks) (Abstract), comprising executing the following on a processor (Fig.5, 64): re-training an artificial neural network (automated classifier) when a confidence level (score) is below a pre-determined level (Fig.4, 12, 14, 16, 18, 32, 54 and Fig.5, 70, 72, 24, 76; [0030] [0031]), wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (image data, [0021]) that is different from what was originally used to train the at least one neural network (The training data used to retrain the artificial neural network comprises content set (content item A and content item C) which is different than the content set (content item1, content item 2, content item 3) used to train the artificial network., Fig. 4, initial training set and supplemental training set; [0030]).
Additionally, Cardinaux discloses a system, method and artificial neural network (Abstract), wherein the artificial neural network is trained using supervised learning ([0021]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 11 of 12,174,631 in the same way that Paquet’s invention has been improved to achieve the following, predictable results for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006]): re-training the at least one neural network further comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to modify the teachings disclosed by the combination of the limitations recited by claim 11 of 12,174,631 and Paquet’s invention with Cardinaux’s teachings so that the re-training (training) comprises supervised learning for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006])
Moreover, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 11 of US 12,164,631.
Thus, the limitations of claim 26 of the current application are obvious variants of the limitations of claim 11 of 12,164, 631. And, claim 26 of the current application and claim 11 of 12,164,631 are not patentably distinct.
Regarding claim 27 of the current application, claim 11 of US 12,164,631 recites the limitations of claim 27 except for: determining if a confidence level associated with the output is lower than a predetermined level; and re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network, wherein the re-training utilizes unsupervised learning.
However, Paquet discloses a system and method for training automated classifiers (artificial neural networks) (Abstract), comprising executing the following on a processor (Fig.5, 64): re-training an artificial neural network (automated classifier) when a confidence level (score) is below a pre-determined level (Fig.4, 12, 14, 16, 18, 32, 54 and Fig.5, 70, 72, 24, 76; [0030] [0031]), wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (image data, [0021]) that is different from what was originally used to train the at least one neural network (The training data used to retrain the artificial neural network comprises content set (content item A and content item C) which is different than the content set (content item1, content item 2, content item 3) used to train the artificial network., Fig. 4, initial training set and supplemental training set; [0030]).
Additionally, Cardinaux discloses a system, method and artificial neural network (Abstract), wherein the artificial neural network is trained using unsupervised learning ([0021]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 11 of 12,174,631 in the same way that Paquet’s invention has been improved to achieve the following, predictable results for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006]): re-training the at least one neural network further comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to modify the teachings disclosed by the combination of the limitations recited by claim 11 of 12,174,631 and Paquet’s invention with Cardinaux’s teachings so that the re-training (training) comprises unsupervised learning for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006])
Moreover, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 11 of US 12,164,631.
Thus, the limitations of claim 27 of the current application are obvious variants of the limitations of claim 11 of 12,164, 631. And, claim 27 of the current application and claim 11 of 12,164,631 are not patentably distinct.
Claims 33 and 34 are rejected on the ground of nonstatutory double patenting as being unpatentable over claim 20 of U.S. Patent No. 12,174,631 in view of Paquet et al. (US 2012/0158620) (“Paquet”). Although the claims at issue are not identical, they are not patentably distinct from each other.
Regarding claim 33 of the current application, claim 20 of US 12,164,631 recites the limitations of claim 33 except for: determine if a confidence level associated with the output is lower than a predetermined level; and re- train the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
However, Paquet discloses a system and method for training automated classifiers (artificial neural networks) (Abstract), comprising executing the following on a processor (Fig.5, 64): re-training an artificial neural network (automated classifier) when a confidence level (score) is below a pre-determined level (Fig.4, 12, 14, 16, 18, 32, 54 and Fig.5, 70, 72, 24, 76; [0030] [0031]), wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (image data, [0021]) that is different from what was originally used to train the at least one neural network (The training data used to retrain the artificial neural network comprises content set (content item A and content item C) which is different than the content set (content item1, content item 2, content item 3) used to train the artificial network., Fig. 4, initial training set and supplemental training set; [0030]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 11 of 12,174,631 in the same way that Paquet’s invention has been improved to achieve the following, predictable results for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006]): retraining the at least one neural network further comprises retraining when the confidence level is below the predetermined level using a set of training data that is different from what was originally used to train the at least one neural network.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 20 of US 12,164,631.
Thus, the limitations of claim 33 of the current application are obvious variants of the limitations of claim 20 of 12,164, 631. And, claim 33 of the current application and claim 11 of 12,164,631 are not patentably distinct.
Regarding claim 34 of the current application, claim 20 of US 12,164,631 recites the limitations of claim 34 except for: determine if a confidence level associated with the output is lower than a predetermined level; and re- train the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network, and wherein the second neural network is prevented from being re-trained.
However, Paquet discloses a system and method for training automated classifiers (artificial neural networks) (Abstract), comprising executing the following on a processor (Fig.5, 64): re-training an artificial neural network (automated classifier) when a confidence level (score) is below a pre-determined level (Fig.4, 12, 14, 16, 18, 32, 54 and Fig.5, 70, 72, 24, 76; [0030] [0031]), wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (image data, [0021]) that is different from what was originally used to train the at least one neural network (The training data used to retrain the artificial neural network comprises content set (content item A and content item C) which is different than the content set (content item1, content item 2, content item 3) used to train the artificial network., Fig. 4, initial training set and supplemental training set; [0030]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 11 of 12,174,631 in the same way that Paquet’s invention has been improved to achieve the following, predictable results for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006]): retraining the at least one neural network further comprises retraining when the confidence level is below the predetermined level using a set of training data that is different from what was originally used to train the at least one neural network.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 20 of US 12,164,631. Furthermore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing that by only re-training the at least one neural network as recited by claim 20 of US 12,164,631, the second neural network is prevented from being re-trained.
Thus, the limitations of claim 34 of the current application are obvious variants of the limitations of claim 20 of 12,164, 631. And, claim 34 of the current application and claim 20 of 12,164,631 are not patentably distinct.
Claims 35 - 37 are rejected on the ground of nonstatutory double patenting as being unpatentable over claim 20 of U.S. Patent No. 12,174,631 in view of Paquet et al. (US 2012/0158620) (“Paquet”) and further in view of Cardinaux et al. (US 2015/0278686) (“Cardinaux”) . Although the claims at issue are not identical, they are not patentably distinct from each other.
Regarding claim 35 of the current application, claim 20 of US 12,164,631 recites the limitations of claim 35 except for: determining if a confidence level associated with the output is lower than a predetermined level; and re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network, wherein the re-training utilizes reinforcement training.
However, Paquet discloses a system and method for training automated classifiers (artificial neural networks) (Abstract), comprising executing the following on a processor (Fig.5, 64): re-training an artificial neural network (automated classifier) when a confidence level (score) is below a pre-determined level (Fig.4, 12, 14, 16, 18, 32, 54 and Fig.5, 70, 72, 24, 76; [0030] [0031]), wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (image data, [0021]) that is different from what was originally used to train the at least one neural network (The training data used to retrain the artificial neural network comprises content set (content item A and content item C) which is different than the content set (content item1, content item 2, content item 3) used to train the artificial network., Fig. 4, initial training set and supplemental training set; [0030]).
Additionally, Cardinaux discloses a system, method and artificial neural network (Abstract), wherein the artificial neural network is trained using reinforcement training ([0021]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 11 of 12,174,631 in the same way that Paquet’s invention has been improved to achieve the following, predictable results for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006]): retraining the at least one neural network further comprises retraining when the confidence level is below the predetermined level using a set of training data that is different from what was originally used to train the at least one neural network.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to modify the teachings disclosed by the combination of the limitations recited by claim 20 of 12,174,631 and Paquet’s invention with Cardinaux’s teachings so that the re-training (training) comprises reinforcement training for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006])
Moreover, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 20 of US 12,164,631.
Thus, the limitations of claim 35 of the current application are obvious variants of the limitations of claim 20 of 12,164, 631. And, claim 35 of the current application and claim 20 of 12,164,631 are not patentably distinct.
Regarding claim 36 of the current application, claim 20 of US 12,164,631 recites the limitations of claim 36 except for: determining if a confidence level associated with the output is lower than a predetermined level; and re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network, wherein the re-training utilizes supervised learning.
However, Paquet discloses a system and method for training automated classifiers (artificial neural networks) (Abstract), comprising executing the following on a processor (Fig.5, 64): re-training an artificial neural network (automated classifier) when a confidence level (score) is below a pre-determined level (Fig.4, 12, 14, 16, 18, 32, 54 and Fig.5, 70, 72, 24, 76; [0030] [0031]), wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (image data, [0021]) that is different from what was originally used to train the at least one neural network (The training data used to retrain the artificial neural network comprises content set (content item A and content item C) which is different than the content set (content item1, content item 2, content item 3) used to train the artificial network., Fig. 4, initial training set and supplemental training set; [0030]).
Additionally, Cardinaux discloses a system, method and artificial neural network (Abstract), wherein the artificial neural network is trained using supervised learning ([0021]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 20 of 12,174,631 in the same way that Paquet’s invention has been improved to achieve the following, predictable results for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006]): retraining the at least one neural network further comprises retraining when the confidence level is below the predetermined level using a set of training data that is different from what was originally used to train the at least one neural network.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to modify the teachings disclosed by the combination of the limitations recited by claim 20 of 12,174,631 and Paquet’s invention with Cardinaux’s teachings so that the re-training (training) comprises supervised learning for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006])
Moreover, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 20 of US 12,164,631.
Thus, the limitations of claim 36 of the current application are obvious variants of the limitations of claim 20 of 12,164, 631. And, claim 36 of the current application and claim 20 of 12,164,631 are not patentably distinct.
Regarding claim 37 of the current application, claim 20 of US 12,164,631 recites the limitations of claim 37 except for: determining if a confidence level associated with the output is lower than a predetermined level; and re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network, wherein the re-training utilizes unsupervised learning.
However, Paquet discloses a system and method for training automated classifiers (artificial neural networks) (Abstract), comprising executing the following on a processor (Fig.5, 64): re-training an artificial neural network (automated classifier) when a confidence level (score) is below a pre-determined level (Fig.4, 12, 14, 16, 18, 32, 54 and Fig.5, 70, 72, 24, 76; [0030] [0031]), wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (image data, [0021]) that is different from what was originally used to train the at least one neural network (The training data used to retrain the artificial neural network comprises content set (content item A and content item C) which is different than the content set (content item1, content item 2, content item 3) used to train the artificial network., Fig. 4, initial training set and supplemental training set; [0030]).
Additionally, Cardinaux discloses a system, method and artificial neural network (Abstract), wherein the artificial neural network is trained using unsupervised learning ([0021]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 20 of 12,174,631 in the same way that Paquet’s invention has been improved to achieve the following, predictable results for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006]): retraining the at least one neural network further comprises retraining when the confidence level is below the predetermined level using a set of training data that is different from what was originally used to train the at least one neural network.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to modify the teachings disclosed by the combination of the limitations recited by claim 20 of 12,174,631 and Paquet’s invention with Cardinaux’s teachings so that the re-training (training) comprises unsupervised learning for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006])
Moreover, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 20 of US 12,164,631.
Thus, the limitations of claim 37 of the current application are obvious variants of the limitations of claim 20 of 12,164, 631. And, claim 20 of the current application and claim 11 of 12,164,631 are not patentably distinct.
The claim mapping regarding parent application US 11,815,893 is as follows.
Current Application
21. (New) A method of operating an apparatus using a control system that includes at least one neural network, comprising the steps of: receiving input data captured by the apparatus; processing the input data, using the at least one neural network of the control system, to obtain an output, the at least one neural network of the control system having at least one hidden layer; determining, using a second neural network having at least one hidden layer, if a confidence level associated with the output is lower than a predetermined level; and controlling the apparatus using the output if the confidence level is not lower than the predetermined level.
23. (New) The method of claim 21, further comprising the step of: re-training the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
24. (New) The method of claim 23, wherein the second neural network is prevented from being re-trained.
31. (New) An apparatus being operated in part by a controller, comprising: an input device constructed to generate input data ;at least one neural network coupled to the controller and constructed to receive the input data and to generate an output, the at least one neural network having at least one hidden layer; and a second neural network with at least one hidden layer constructed to use the output from the at least one neural network to determine whether a confidence level associated with the output is below a predetermined level, wherein the controller is further constructed to operate the apparatus using the output unless the confidence level of the output from the at least one neural network is below the predetermined level.
33. (New) The apparatus of claim 31, wherein the controller is further structured to re- train the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
34. (New) The apparatus of claim 33, wherein the second neural network is prevented from being re-trained.
US 11,815,893
1. A method of operating an apparatus using a control system that includes at least one neural network, comprising: receiving input data captured by the apparatus; processing the input data using the at least one neural network, the at least one neural network having at least one hidden layer; obtaining an output from the at least one neural network resulting from processing the input data; using the output from the at least one neural network to control the apparatus unless a confidence level of the output is below a predetermined level; and re-training the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
2. The method of claim 1, wherein using the obtained output from the at least one neural network to control the apparatus includes processing the output from the at least one neural network with a second neural network with at least one hidden layer to determine whether the confidence level is below the predetermined level.
3. The method of claim 2, wherein the second neural network is prevented from being retrained.
7. An apparatus being operated in part by a controller, comprising: an input device constructed to generate input data; at least one neural network coupled to the controller and constructed to receive the input data and to generate an output, the at least one neural network having at least one hidden layer; and a second neural network with at least one hidden layer constructed to use the output from the at least one neural network to determine whether a confidence level of the output is below a predetermined level, wherein the controller is further constructed to operate the apparatus using the output unless the confidence level of the output from the at least one neural network is below a predetermined level, and wherein the controller is further constructed to re-train the at least one neural network when the confidence level of the output from the at least one neural network is below the predetermined level, wherein re-training the at least one neural network comprises using a set of training data that is different from what was originally used to train the at least one neural network.
8. The apparatus of claim 7, wherein the second neural network is prevented from being retrained.
Claims 23, 24, 33 and 34 are rejected on the ground of nonstatutory double patenting as being unpatentable over claims 2, 3, 7 and 8 of U.S. Patent No. 11,815,893, respectively. Although the claims at issue are not identical, they are not patentably distinct from each other.
Regarding claim 23 of the current application, claim 2 of US 11,815,893 recites the limitations of claim 23 except for: the at least one neural network is of the control system. However, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing the at least one neural network which processes the input is the at least one neural network included in the control system.
Thus, the limitations of claim 23 of the current application are obvious variants of the limitations of claim 2 of 11,815,893. And, claim 23 of the current application and claim 2 of 11,815,893 are not patentably distinct.
Regarding claim 24 of the current application, claim 3 of US 11,815,893 recites the limitations of claim 24 except for: the at least one neural network is of the control system. However, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing the at least one neural network which processes the input is the at least one neural network included in the control system.
Thus, the limitations of claim 24 of the current application are obvious variants of the limitations of claim 3 of 11,815,893. And, claim 24 of the current application and claim 3 of 11,815,893 are not patentably distinct.
Regarding claim 33 of the current application, claim 7 of US 11,815,893 recites the limitations of claim 33 except for: determining if a confidence level associated with the output is lower than a predetermined level. However, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 7 of US 11,815,893.
Thus, the limitations of claim 33 of the current application are obvious variants of the limitations of claim 7 of US 11,815,89. And, claim 33 of the current application and claim 7 of US 11,815,89 are not patentably distinct.
Regarding claim 34 of the current application, claim 8 of US 11,815,893 recites the limitations of claim 34 except for: determining if a confidence level associated with the output is lower than a predetermined level. However, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 8 of US 11,815,893.
Thus, the limitations of claim 34 of the current application are obvious variants of the limitations of claim 8 of US 11,815,893. And, claim 34 of the current application and claim 8 of US 11,815,893 are not patentably distinct.
The claim mapping regarding parent application US 11,366,472 is as follows.
Current Application
21. (New) A method of operating an apparatus using a control system that includes at least one neural network, comprising the steps of: receiving input data captured by the apparatus; processing the input data, using the at least one neural network of the control system, to obtain an output, the at least one neural network of the control system having at least one hidden layer; determining, using a second neural network having at least one hidden layer, if a confidence level associated with the output is lower than a predetermined level; and controlling the apparatus using the output if the confidence level is not lower than the predetermined level.
31. (New) An apparatus being operated in part by a controller, comprising: an input device constructed to generate input data ;at least one neural network coupled to the controller and constructed to receive the input data and to generate an output, the at least one neural network having at least one hidden layer; and a second neural network with at least one hidden layer constructed to use the output from the at least one neural network to determine whether a confidence level associated with the output is below a predetermined level, wherein the controller is further constructed to operate the apparatus using the output unless the confidence level of the output from the at least one neural network is below the predetermined level.
US 11,366,472
1. A method of operating an apparatus using a control system, comprising: processing, using at least one first neural network of the control system implemented on one or more solid-state chips of the control system, an input value from a sensor coupled to the apparatus; obtaining an output from the at least one first neural network resulting from processing the input value; processing the output of the at least one first neural network with a second neural network of the control system implemented on a second one or more solid-state chips to obtain an output of the second neural network, the output of the second neural network indicative of whether the output of the at least one first neural network breaches a predetermined operating condition for operating the apparatus, the predetermined operating condition comprising a confidence level of the output of the first neural network; using the output from the at least one first neural network to control the apparatus unless the output of the second neural network indicates the output of the at least one first neural network breaches the predetermined operating condition; and re-training the at least one first neural network when the output of the at least one first neural network is determined to breach the predetermined condition.
8. An apparatus, comprising: a controller configured to receive an input value generated by a sensor coupled to the controller, the controller including at least one first neural network implemented on a first one or more solid-state chips coupled to the controller, the at least one neural network constructed to receive the input value and to generate an output; and a second neural network implemented on a second one or more solid-state chips constructed to receive the output from the at least one first neural network and determine whether the output breaches a predetermined condition, the predetermined condition comprising a confidence level of the output of the first neural network; wherein the controller is constructed to operate the apparatus using the output from the at least one neural network unless the confidence level of the output from the at least one first neural network is determined to fall below a predetermined level, wherein the controller is further constructed to re-train the at least one first neural network when the confidence level of the output from the at least one first neural network falls below the predetermined level.
Claims 21 and 31 are rejected on the ground of nonstatutory double patenting as being unpatentable over claims 1 and 8 of U.S. Patent No. 11,366,472, respectively. Although the claims at issue are not identical, they are not patentably distinct from each other.
Regarding claim 21 of the current application, claim 1 of US 11,366,472 recites the limitations of claim 21 except for: each of the first and second neural networks comprising at least one hidden layer. However, it would have been well known and obvious to one of ordinary skill in the art at the time of applicant’s filing that neural networks may comprise hidden layers.
Thus, the limitations of claim 21 of the current application are obvious variants of the limitations of claim 1 of 11,366,472. And, claim 21 of the current application and claim 1 of 11,366,472 are not patentably distinct.
Regarding claim 31 of the current application, claim 8 of US 11,366,472 recites the limitations of claim 31 except for: each of the first and second neural networks comprising at least one hidden layer. However, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to determine if a confidence level associated with the output is lower than a predetermined level to control the apparatus using the output if the confidence level is not lower than the predetermined level as recited by claim 8 of US 11,366,472.
Thus, the limitations of claim 31 of the current application are obvious variants of the limitations of claim 8 of US 11,366,472. And, claim 31 of the current application and claim 8 of US 11,366,472 are not patentably distinct.
The claim mapping regarding parent application US 10,802,489 is as follows.
Current Application
21. (New) A method of operating an apparatus using a control system that includes at least one neural network, comprising the steps of: receiving input data captured by the apparatus; processing the input data, using the at least one neural network of the control system, to obtain an output, the at least one neural network of the control system having at least one hidden layer; determining, using a second neural network having at least one hidden layer, if a confidence level associated with the output is lower than a predetermined level; and controlling the apparatus using the output if the confidence level is not lower than the predetermined level.
31. (New) An apparatus being operated in part by a controller, comprising: an input device constructed to generate input data ;at least one neural network coupled to the controller and constructed to receive the input data and to generate an output, the at least one neural network having at least one hidden layer; and a second neural network with at least one hidden layer constructed to use the output from the at least one neural network to determine whether a confidence level associated with the output is below a predetermined level, wherein the controller is further constructed to operate the apparatus using the output unless the confidence level of the output from the at least one neural network is below the predetermined level.
US 10,802,489
1. A method of operating an apparatus using a control system, the method comprising: receiving an input value from an apparatus, the input value based on an image captured by a device coupled to the apparatus; processing the input value using at least one neural network of the control system, the at least one neural network implemented on a first one or more solid-state chips, and obtaining an output from the at least one neural network resulting from processing the input value; processing the output with a second neural network of the control system implemented on a second one or more solid-state chips to determine whether the output breaches a predetermined condition that is unchangeable after an initial installation onto the control system, wherein the second neural network is prevented from being retrained; re-training the at least one neural network with a second training data set of images that is different than a first training set of images used to initially train the at least one neural network, when the output is determined to breach the predetermined condition; and using the output from the at least one neural network to control the apparatus unless the output breaches the predetermined condition.
6. An apparatus operated in part by a controller, the apparatus comprising: an input device coupled to the apparatus, the input device constructed to generate an input value based on an image captured by the input device; a controller; at least one neural network implemented on a first one or more solid-state chips coupled to the controller, the at least one neural network constructed to receive the input value and to generate an output; and wherein the controller comprises a second neural network implemented on a second one or more solid-state chips constructed to receive the output from the at least one neural network and determine whether the output breaches a predetermined condition unchangeable after an initial installation onto the controller, wherein the second neural network is prevented from being retrained, wherein the controller is constructed to operate the apparatus using the output from the at least one neural network unless the output from the at least one neural network is determined to breach the predetermined condition, and wherein the controller is further constructed to re-train the at least one neural network, with an image training data set that is different than an initial image training data set used to train the at least one neural, when the output is determined to breach the predetermined condition.
Claims 21 and 31 are rejected on the ground of nonstatutory double patenting as being unpatentable over claim 1 and 6 of U.S. Patent No. 10,802,489, respectively in view of Hu et al. (US 2018/0157934) (“Hu”). Although the claims at issue are not identical, they are not patentably distinct from each other.
Regarding claim 21 of the current application, claim 1 of US 10,802,489 recites the limitations of claim 21 except for: each of the first and second neural networks comprising at least one hidden layer; and determining, using the second neural network, if a confidence level associated with the output is lower than a predetermined level .
However, Hu discloses a method of operating an apparatus using a control system that includes at least one neural network (Abstract), comprising: determining, using a second neural network (inspection neural network in an analyzer inspector, Fig.1, 120 and 122; [0027] [0028]) if a confidence level associated with the output is lower than a predetermined level (The inspection neural network receives the input, intermediate values generated by the primary neural network and output generated by the primary neural network. The inspection neural network generates a reliability indicator which is a value that indicates reliability/confidence of the output of the primary neural network. The reliability indicator is compared to a threshold., Fig.3, 306 and 308; [0027] [0030] [0038] [0040] [0045 – 0047]); and controlling the apparatus using the output if the confidence level is not lower than the predetermined level (“The controller 104 may monitor the output 114 and reliability indicator 126 and alter the operation conditions of a dynamic system under its control … The navigator 220 may use the output received from the data analyzers 240A and 240B and the reliability indicators received from the analyzer inspectors 250A and 250B to make decisions about vehicle movements. For example, the output from data analyzers 240A and 240B may indicate a confidence map of drivable regions in an image that represents the frontal view of the vehicle 200. The navigator may receive repeated confidence maps from the data analyzers 240A and 240B and use the confidence maps to determine the direction of the road as the vehicle 200 moves forward. The reliability indicators generated by the analyzer inspectors 250A and 250B may be used in a number of ways. In one example, if the indicator falls below a threshold, the navigator 220 may simply ignore the accompanying output. Where the output comprises confidence maps of drivable regions, the navigator 220 may ignore the particular confidence map and wait for the next confidence map. In some cases, the navigator 220 may cause the vehicle to slow down until it begins to receive confidence maps with better reliability indicators.”, [0031] [0039] [0040] [0047]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 1 of 10,802,489 in the same way that Hu’s invention has been improved to achieve the following, predictable results for the purpose of automating processes with a reliable use of neural networks (Hu, [0003]): the predetermined condition that is unchangeable after an initial installation further comprises a predetermined level, wherein breaching the predetermined condition further comprises a confidence level associated with the output being lower than the predetermined level .
Additionally, it would have been well known and obvious to one of ordinary skill in the art at the time of applicant’s filing that neural networks may comprise hidden layers.
Thus, the limitations of claim 21of the current application are obvious variants of the limitations of claim 1 of 10,802,489. And, claim 21 of the current application and claim 1 of 10,802,489 are not patentably distinct.
Regarding claim 31 of the current application, claim 6 of US 10,802,489 recites the limitations of claim 31 except for: each of the first and second neural networks comprising at least one hidden layer; and determining, using the second neural network, if a confidence level associated with the output is lower than a predetermined level .
However, Hu discloses a method of operating an apparatus using a control system that includes at least one neural network (Abstract), comprising: determining, using a second neural network (inspection neural network in an analyzer inspector, Fig.1, 120 and 122; [0027] [0028]) if a confidence level associated with the output is lower than a predetermined level (The inspection neural network receives the input, intermediate values generated by the primary neural network and output generated by the primary neural network. The inspection neural network generates a reliability indicator which is a value that indicates reliability/confidence of the output of the primary neural network. The reliability indicator is compared to a threshold., Fig.3, 306 and 308; [0027] [0030] [0038] [0040] [0045 – 0047]); and controlling the apparatus using the output if the confidence level is not lower than the predetermined level (“The controller 104 may monitor the output 114 and reliability indicator 126 and alter the operation conditions of a dynamic system under its control … The navigator 220 may use the output received from the data analyzers 240A and 240B and the reliability indicators received from the analyzer inspectors 250A and 250B to make decisions about vehicle movements. For example, the output from data analyzers 240A and 240B may indicate a confidence map of drivable regions in an image that represents the frontal view of the vehicle 200. The navigator may receive repeated confidence maps from the data analyzers 240A and 240B and use the confidence maps to determine the direction of the road as the vehicle 200 moves forward. The reliability indicators generated by the analyzer inspectors 250A and 250B may be used in a number of ways. In one example, if the indicator falls below a threshold, the navigator 220 may simply ignore the accompanying output. Where the output comprises confidence maps of drivable regions, the navigator 220 may ignore the particular confidence map and wait for the next confidence map. In some cases, the navigator 220 may cause the vehicle to slow down until it begins to receive confidence maps with better reliability indicators.”, [0031] [0039] [0040] [0047]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 6 of 10,802,489 in the same way that Hu’s invention has been improved to achieve the following, predictable results for the purpose of automating processes with a reliable use of neural networks (Hu, [0003]): the predetermined condition that is unchangeable after an initial installation further comprises a predetermined level, wherein breaching the predetermined condition further comprises a confidence level associated with the output being lower than the predetermined level
Additionally, it would have been well known and obvious to one of ordinary skill in the art at the time of applicant’s filing that neural networks may comprise hidden layers.
Thus, the limitations of claim 31 of the current application are obvious variants of the limitations of claim 6 of 10,802,489. And, claim 31 of the current application and c claim 6 of 10,802,489 are not patentably distinct.
The claim mapping regarding parent application US 10,672,389 is as follows.
Current Application
21. (New) A method of operating an apparatus using a control system that includes at least one neural network, comprising the steps of: receiving input data captured by the apparatus; processing the input data, using the at least one neural network of the control system, to obtain an output, the at least one neural network of the control system having at least one hidden layer; determining, using a second neural network having at least one hidden layer, if a confidence level associated with the output is lower than a predetermined level; and controlling the apparatus using the output if the confidence level is not lower than the predetermined level.
31. (New) An apparatus being operated in part by a controller, comprising: an input device constructed to generate input data ;at least one neural network coupled to the controller and constructed to receive the input data and to generate an output, the at least one neural network having at least one hidden layer; and a second neural network with at least one hidden layer constructed to use the output from the at least one neural network to determine whether a confidence level associated with the output is below a predetermined level, wherein the controller is further constructed to operate the apparatus using the output unless the confidence level of the output from the at least one neural network is below the predetermined level.
US 10,672,389
1. A method of operating an apparatus using a control system that includes at least one neural network, comprising the steps of: receiving input data captured by the apparatus; processing the input data using the at least one neural network of the control system, the at least one neural network having at least one hidden layer; obtaining an output from the at least one neural network resulting from processing the input data; and using the obtained output from the at least one neural network to control the apparatus unless the obtained output from the at least one neural network is determined to breach a predetermined condition that is unchangeable after an initial installation onto the control system, wherein using the obtained output from the at least one neural network to control the apparatus includes processing the output from the at least one neural network with a second neural network with at least one hidden layer to determine whether the output breaches the predetermined condition, and re-training the at least one neural network when the output is determined to breach the predetermined condition, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
8. An apparatus being operated in part by a controller, comprising: an input device constructed to generate input data; at least one neural network coupled to the controller and constructed to receive the input data and to generate an output, the at least one neural network having at least one hidden layer; and a comparator comprising a second neural network with at least one hidden layer constructed to compare the output from the at least one neural network with a predetermined condition to determine whether the output breaches the predetermined condition, the predetermined condition being unchangeable after an initial installation onto the control system, wherein the controller is further constructed to operate the apparatus using the output unless the obtained output from the at least one neural network is determined to breach the predetermined condition, and to retrain the at least one neural network when the output from the at least one neural network is determined to breach the predetermined condition, wherein re-training the at least one neural network comprises using a set of training data that is different from what was originally used to train the at least one neural network
Claims 21 and 31 are rejected on the ground of nonstatutory double patenting as being unpatentable over claim 1 and 8 of U.S. Patent No. 10,672,389, respectively in view of Hu et al. (US 2018/0157934) (“Hu”). Although the claims at issue are not identical, they are not patentably distinct from each other.
Regarding claim 21 of the current application, claim 1 of US 10,672,389 recites the limitations of claim 21 except for: determining, using the second neural network, if a confidence level associated with the output is lower than a predetermined level.
However, Hu discloses a method of operating an apparatus using a control system that includes at least one neural network (Abstract), comprising: determining, using a second neural network (inspection neural network in an analyzer inspector, Fig.1, 120 and 122; [0027] [0028]) if a confidence level associated with the output is lower than a predetermined level (The inspection neural network receives the input, intermediate values generated by the primary neural network and output generated by the primary neural network. The inspection neural network generates a reliability indicator which is a value that indicates reliability/confidence of the output of the primary neural network. The reliability indicator is compared to a threshold., Fig.3, 306 and 308; [0027] [0030] [0038] [0040] [0045 – 0047]); and controlling the apparatus using the output if the confidence level is not lower than the predetermined level (“The controller 104 may monitor the output 114 and reliability indicator 126 and alter the operation conditions of a dynamic system under its control … The navigator 220 may use the output received from the data analyzers 240A and 240B and the reliability indicators received from the analyzer inspectors 250A and 250B to make decisions about vehicle movements. For example, the output from data analyzers 240A and 240B may indicate a confidence map of drivable regions in an image that represents the frontal view of the vehicle 200. The navigator may receive repeated confidence maps from the data analyzers 240A and 240B and use the confidence maps to determine the direction of the road as the vehicle 200 moves forward. The reliability indicators generated by the analyzer inspectors 250A and 250B may be used in a number of ways. In one example, if the indicator falls below a threshold, the navigator 220 may simply ignore the accompanying output. Where the output comprises confidence maps of drivable regions, the navigator 220 may ignore the particular confidence map and wait for the next confidence map. In some cases, the navigator 220 may cause the vehicle to slow down until it begins to receive confidence maps with better reliability indicators.”, [0031] [0039] [0040] [0047]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 1 of 10,802,489 in the same way that Hu’s invention has been improved to achieve the following, predictable results for the purpose of automating processes with a reliable use of neural networks (Hu, [0003]): the predetermined condition that is unchangeable after an initial installation further comprises a predetermined level, wherein breaching the predetermined condition further comprises a confidence level associated with the output being lower than the predetermined level .
Thus, the limitations of claim 21 of the current application are obvious variants of the limitations of claim 1 of 10,672,389. And, claim 21 of the current application and claim 1 of 10,672,389 are not patentably distinct.
Regarding claim 31 of the current application, claim 8 of US 10,672,389 recites the limitations of claim 31 except for: each of the first and second neural networks comprising at least one hidden layer; and determining, using the second neural network, if a confidence level associated with the output is lower than a predetermined level .
However, Hu discloses a method of operating an apparatus using a control system that includes at least one neural network (Abstract), comprising: determining, using a second neural network (inspection neural network in an analyzer inspector, Fig.1, 120 and 122; [0027] [0028]) if a confidence level associated with the output is lower than a predetermined level (The inspection neural network receives the input, intermediate values generated by the primary neural network and output generated by the primary neural network. The inspection neural network generates a reliability indicator which is a value that indicates reliability/confidence of the output of the primary neural network. The reliability indicator is compared to a threshold., Fig.3, 306 and 308; [0027] [0030] [0038] [0040] [0045 – 0047]); and controlling the apparatus using the output if the confidence level is not lower than the predetermined level (“The controller 104 may monitor the output 114 and reliability indicator 126 and alter the operation conditions of a dynamic system under its control … The navigator 220 may use the output received from the data analyzers 240A and 240B and the reliability indicators received from the analyzer inspectors 250A and 250B to make decisions about vehicle movements. For example, the output from data analyzers 240A and 240B may indicate a confidence map of drivable regions in an image that represents the frontal view of the vehicle 200. The navigator may receive repeated confidence maps from the data analyzers 240A and 240B and use the confidence maps to determine the direction of the road as the vehicle 200 moves forward. The reliability indicators generated by the analyzer inspectors 250A and 250B may be used in a number of ways. In one example, if the indicator falls below a threshold, the navigator 220 may simply ignore the accompanying output. Where the output comprises confidence maps of drivable regions, the navigator 220 may ignore the particular confidence map and wait for the next confidence map. In some cases, the navigator 220 may cause the vehicle to slow down until it begins to receive confidence maps with better reliability indicators.”, [0031] [0039] [0040] [0047]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 8 of US 10,672,389 in the same way that Hu’s invention has been improved to achieve the following, predictable results for the purpose of automating processes with a reliable use of neural networks (Hu, [0003]): the predetermined condition that is unchangeable after an initial installation further comprises a predetermined level, wherein breaching the predetermined condition further comprises a confidence level associated with the output being lower than the predetermined level .
Thus, the limitations of claim 31 of the current application are obvious variants of the limitations of claim 8 of US 10,672,389. And, claim 31 of the current application and claim 8 of US 10,672,389 are not patentably distinct.
The claim mapping regarding parent application is as follows.
Current Application
21. (New) A method of operating an apparatus using a control system that includes at least one neural network, comprising the steps of: receiving input data captured by the apparatus; processing the input data, using the at least one neural network of the control system, to obtain an output, the at least one neural network of the control system having at least one hidden layer; determining, using a second neural network having at least one hidden layer, if a confidence level associated with the output is lower than a predetermined level; and controlling the apparatus using the output if the confidence level is not lower than the predetermined level.
31. (New) An apparatus being operated in part by a controller, comprising: an input device constructed to generate input data ;at least one neural network coupled to the controller and constructed to receive the input data and to generate an output, the at least one neural network having at least one hidden layer; and a second neural network with at least one hidden layer constructed to use the output from the at least one neural network to determine whether a confidence level associated with the output is below a predetermined level, wherein the controller is further constructed to operate the apparatus using the output unless the confidence level of the output from the at least one neural network is below the predetermined level.
US 10,242,665
1. A method of operating an apparatus using a control system that includes at least one neural network, comprising: receiving an input vector captured by the apparatus; processing the input vector using the at least one neural network of the control system; obtaining an output from the at least one neural network resulting from processing the input vector; processing the output with a second neural network to determine whether the output breaches a predetermined condition that is unchangeable after an initial installation onto the control system, wherein the second neural network is prevented from being retrained; and using the output from the at least one neural network to control the apparatus unless the output breaches the predetermined condition.
8. An apparatus being operated in part by a controller, the apparatus comprising: an input device coupled to the apparatus, the input device constructed to generate an input vector; a controller; at least one neural network coupled to the controller, the at least one neural network constructed to receive the input vector and to generate an output; and wherein the controller comprises a second neural network constructed to receive the output from the at least one neural network and determine whether the output breaches a predetermined condition unchangeable after an initial installation onto the controller, wherein the second neural network is prevented from being retrained, and wherein the controller is constructed to operate the apparatus using the output from the at least one neural network unless the output from the at least one neural network is determined to breach the predetermined condition.
Claims 21 and 31 are rejected on the ground of nonstatutory double patenting as being unpatentable over claim 1 and 8 of US 10,242,665, respectively in view of Hu et al. (US 2018/0157934) (“Hu”). Although the claims at issue are not identical, they are not patentably distinct from each other.
Regarding claim 21 of the current application, claim 1 of US 10,242,665 recites the limitations of claim 21 except for: each of the first and second neural networks comprising at least one hidden layer; and determining, using the second neural network, if a confidence level associated with the output is lower than a predetermined level .
However, Hu discloses a method of operating an apparatus using a control system that includes at least one neural network (Abstract), comprising: determining, using a second neural network (inspection neural network in an analyzer inspector, Fig.1, 120 and 122; [0027] [0028]) if a confidence level associated with the output is lower than a predetermined level (The inspection neural network receives the input, intermediate values generated by the primary neural network and output generated by the primary neural network. The inspection neural network generates a reliability indicator which is a value that indicates reliability/confidence of the output of the primary neural network. The reliability indicator is compared to a threshold., Fig.3, 306 and 308; [0027] [0030] [0038] [0040] [0045 – 0047]); and controlling the apparatus using the output if the confidence level is not lower than the predetermined level (“The controller 104 may monitor the output 114 and reliability indicator 126 and alter the operation conditions of a dynamic system under its control … The navigator 220 may use the output received from the data analyzers 240A and 240B and the reliability indicators received from the analyzer inspectors 250A and 250B to make decisions about vehicle movements. For example, the output from data analyzers 240A and 240B may indicate a confidence map of drivable regions in an image that represents the frontal view of the vehicle 200. The navigator may receive repeated confidence maps from the data analyzers 240A and 240B and use the confidence maps to determine the direction of the road as the vehicle 200 moves forward. The reliability indicators generated by the analyzer inspectors 250A and 250B may be used in a number of ways. In one example, if the indicator falls below a threshold, the navigator 220 may simply ignore the accompanying output. Where the output comprises confidence maps of drivable regions, the navigator 220 may ignore the particular confidence map and wait for the next confidence map. In some cases, the navigator 220 may cause the vehicle to slow down until it begins to receive confidence maps with better reliability indicators.”, [0031] [0039] [0040] [0047]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 1 of US 10,242,665 in the same way that Hu’s invention has been improved to achieve the following, predictable results for the purpose of automating processes with a reliable use of neural networks (Hu, [0003]): the predetermined condition that is unchangeable after an initial installation further comprises a predetermined level, wherein breaching the predetermined condition further comprises a confidence level associated with the output being lower than the predetermined level .
Additionally, it would have been well known and obvious to one of ordinary skill in the art at the time of applicant’s filing that neural networks may comprise hidden layers.
Thus, the limitations of claim 21 of the current application are obvious variants of the limitations of claim 1 of US 10,242,665. And, claim 21 of the current application and claim 1 of US 10,242,665 are not patentably distinct.
Regarding claim 31 of the current application, claim 8 of US 10,242,665 recites the limitations of claim 31 except for: each of the first and second neural networks comprising at least one hidden layer; and determining, using the second neural network, if a confidence level associated with the output is lower than a predetermined level .
However, Hu discloses a method of operating an apparatus using a control system that includes at least one neural network (Abstract), comprising: determining, using a second neural network (inspection neural network in an analyzer inspector, Fig.1, 120 and 122; [0027] [0028]) if a confidence level associated with the output is lower than a predetermined level (The inspection neural network receives the input, intermediate values generated by the primary neural network and output generated by the primary neural network. The inspection neural network generates a reliability indicator which is a value that indicates reliability/confidence of the output of the primary neural network. The reliability indicator is compared to a threshold., Fig.3, 306 and 308; [0027] [0030] [0038] [0040] [0045 – 0047]); and controlling the apparatus using the output if the confidence level is not lower than the predetermined level (“The controller 104 may monitor the output 114 and reliability indicator 126 and alter the operation conditions of a dynamic system under its control … The navigator 220 may use the output received from the data analyzers 240A and 240B and the reliability indicators received from the analyzer inspectors 250A and 250B to make decisions about vehicle movements. For example, the output from data analyzers 240A and 240B may indicate a confidence map of drivable regions in an image that represents the frontal view of the vehicle 200. The navigator may receive repeated confidence maps from the data analyzers 240A and 240B and use the confidence maps to determine the direction of the road as the vehicle 200 moves forward. The reliability indicators generated by the analyzer inspectors 250A and 250B may be used in a number of ways. In one example, if the indicator falls below a threshold, the navigator 220 may simply ignore the accompanying output. Where the output comprises confidence maps of drivable regions, the navigator 220 may ignore the particular confidence map and wait for the next confidence map. In some cases, the navigator 220 may cause the vehicle to slow down until it begins to receive confidence maps with better reliability indicators.”, [0031] [0039] [0040] [0047]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 8 of US 10,242,665 in the same way that Hu’s invention has been improved to achieve the following, predictable results for the purpose of automating processes with a reliable use of neural networks (Hu, [0003]): the predetermined condition that is unchangeable after an initial installation further comprises a predetermined level, wherein breaching the predetermined condition further comprises a confidence level associated with the output being lower than the predetermined level .
Additionally, it would have been well known and obvious to one of ordinary skill in the art at the time of applicant’s filing that neural networks may comprise hidden layers.
Thus, the limitations of claim 31 of the current application are obvious variants of the limitations of claim 8 of US 10,242,665. And, claim 31 of the current application and claim 8 of US 10,242,665 are not patentably distinct.
The claim mapping regarding parent application US 10,324,467 is as follows.
Current Application
21. (New) A method of operating an apparatus using a control system that includes at least one neural network, comprising the steps of: receiving input data captured by the apparatus; processing the input data, using the at least one neural network of the control system, to obtain an output, the at least one neural network of the control system having at least one hidden layer; determining, using a second neural network having at least one hidden layer, if a confidence level associated with the output is lower than a predetermined level; and controlling the apparatus using the output if the confidence level is not lower than the predetermined level.
31. (New) An apparatus being operated in part by a controller, comprising: an input device constructed to generate input data ;at least one neural network coupled to the controller and constructed to receive the input data and to generate an output, the at least one neural network having at least one hidden layer; and a second neural network with at least one hidden layer constructed to use the output from the at least one neural network to determine whether a confidence level associated with the output is below a predetermined level, wherein the controller is further constructed to operate the apparatus using the output unless the confidence level of the output from the at least one neural network is below the predetermined level.
US 10,324,467
1. An apparatus having a plurality of components, comprising: an input device constructed to generate an input data vector; a first neural network trained to generate an output to control the apparatus by processing the input data vector; a second neural network configured to receive the output from the first neural network, the second neural network trained to determine whether the output of the first neural network corresponds to a predetermined condition and generate a control output from the output of the first neural network; and a processor configured to: receive the control output from the second neural network, and in response to the control output indicating the output of the first neural network corresponds to a predetermined condition, control an operation of the first neural network, and not using the output from the first neural network to control the apparatus.
9. A computer-implemented method of controlling an apparatus, the method comprising: processing an input data vector using a first neural network trained to generate an output by inferencing on the input data vector; controlling an operation of an aspect of the apparatus using the output from the first neural network; generating a control output using a second neural network trained to generate the control output by inferencing on the output generated by the first neural network, the control output indicating the output generated by the first neural network corresponds to a predetermined condition; and in response to the first neural network output corresponding to the predetermined condition, controlling an operation of the first neural network using the control output from the second neural network and not controlling an operation of the apparatus using the output from the first neural network, wherein the method is performed by one or more computer hardware processors configured to execute computer-executable instructions on a non-transitory computer storage medium.
Claims 21 and 31 are rejected on the ground of nonstatutory double patenting as being unpatentable over claim 1 and 9 of US 10,324,467, respectively in view of Hu et al. (US 2018/0157934) (“Hu”). Although the claims at issue are not identical, they are not patentably distinct from each other.
Regarding claim 21 of the current application, claim 9 of US 10,324,467 recites the limitations of claim 21 except for: each of the first and second neural networks comprising at least one hidden layer; and determining, using the second neural network, if a confidence level associated with the output is lower than a predetermined level .
However, Hu discloses a method of operating an apparatus using a control system that includes at least one neural network (Abstract), comprising: determining, using a second neural network (inspection neural network in an analyzer inspector, Fig.1, 120 and 122; [0027] [0028]) if a confidence level associated with the output is lower than a predetermined level (The inspection neural network receives the input, intermediate values generated by the primary neural network and output generated by the primary neural network. The inspection neural network generates a reliability indicator which is a value that indicates reliability/confidence of the output of the primary neural network. The reliability indicator is compared to a threshold., Fig.3, 306 and 308; [0027] [0030] [0038] [0040] [0045 – 0047]); and controlling the apparatus using the output if the confidence level is not lower than the predetermined level (“The controller 104 may monitor the output 114 and reliability indicator 126 and alter the operation conditions of a dynamic system under its control … The navigator 220 may use the output received from the data analyzers 240A and 240B and the reliability indicators received from the analyzer inspectors 250A and 250B to make decisions about vehicle movements. For example, the output from data analyzers 240A and 240B may indicate a confidence map of drivable regions in an image that represents the frontal view of the vehicle 200. The navigator may receive repeated confidence maps from the data analyzers 240A and 240B and use the confidence maps to determine the direction of the road as the vehicle 200 moves forward. The reliability indicators generated by the analyzer inspectors 250A and 250B may be used in a number of ways. In one example, if the indicator falls below a threshold, the navigator 220 may simply ignore the accompanying output. Where the output comprises confidence maps of drivable regions, the navigator 220 may ignore the particular confidence map and wait for the next confidence map. In some cases, the navigator 220 may cause the vehicle to slow down until it begins to receive confidence maps with better reliability indicators.”, [0031] [0039] [0040] [0047]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 9 of US 10,324,46 in the same way that Hu’s invention has been improved to achieve the following, predictable results for the purpose of automating processes with a reliable use of neural networks (Hu, [0003]): the predetermined condition further comprises a predetermined level, wherein corresponding to the predetermined condition further comprises a confidence level associated with the output being lower than the predetermined level.
Additionally, it would have been well known and obvious to one of ordinary skill in the art at the time of applicant’s filing that neural networks may comprise hidden layers.
Thus, the limitations of claim 21 of the current application are obvious variants of the limitations of claim 9 of US 10,324,46. And, claim 21 of the current application and claim 9 of US 10,324,46 are not patentably distinct.
Regarding claim 31 of the current application, claim 1 of US 10,324,467 recites the limitations of claim 31 except for: each of the first and second neural networks comprising at least one hidden layer; and determining, using the second neural network, if a confidence level associated with the output is lower than a predetermined level .
However, Hu discloses a method of operating an apparatus using a control system that includes at least one neural network (Abstract), comprising: determining, using a second neural network (inspection neural network in an analyzer inspector, Fig.1, 120 and 122; [0027] [0028]) if a confidence level associated with the output is lower than a predetermined level (The inspection neural network receives the input, intermediate values generated by the primary neural network and output generated by the primary neural network. The inspection neural network generates a reliability indicator which is a value that indicates reliability/confidence of the output of the primary neural network. The reliability indicator is compared to a threshold., Fig.3, 306 and 308; [0027] [0030] [0038] [0040] [0045 – 0047]); and controlling the apparatus using the output if the confidence level is not lower than the predetermined level (“The controller 104 may monitor the output 114 and reliability indicator 126 and alter the operation conditions of a dynamic system under its control … The navigator 220 may use the output received from the data analyzers 240A and 240B and the reliability indicators received from the analyzer inspectors 250A and 250B to make decisions about vehicle movements. For example, the output from data analyzers 240A and 240B may indicate a confidence map of drivable regions in an image that represents the frontal view of the vehicle 200. The navigator may receive repeated confidence maps from the data analyzers 240A and 240B and use the confidence maps to determine the direction of the road as the vehicle 200 moves forward. The reliability indicators generated by the analyzer inspectors 250A and 250B may be used in a number of ways. In one example, if the indicator falls below a threshold, the navigator 220 may simply ignore the accompanying output. Where the output comprises confidence maps of drivable regions, the navigator 220 may ignore the particular confidence map and wait for the next confidence map. In some cases, the navigator 220 may cause the vehicle to slow down until it begins to receive confidence maps with better reliability indicators.”, [0031] [0039] [0040] [0047]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the limitations recited by claim 1 of US 10,324,467 in the same way that Hu’s invention has been improved to achieve the following, predictable results for the purpose of automating processes with a reliable use of neural networks (Hu, [0003]): the predetermined condition further comprises a predetermined level, wherein corresponding to the predetermined condition further comprises a confidence level associated with the output being lower than the predetermined level.
Additionally, it would have been well known and obvious to one of ordinary skill in the art at the time of applicant’s filing that neural networks may comprise hidden layers.
Thus, the limitations of claim 31 of the current application are obvious variants of the limitations of claim 1 of US 10,324,467. And, claim 31 of the current application and claim 1 of US 10,324,467 are not patentably distinct.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 21, 22, 29, 31, 32 and 39 is/are rejected under 35 U.S.C. 103 as being unpatentable over Hu et al. (US 2018/0157934) (“Hu”).
For claim 21, Hu discloses a method of operating an apparatus using a control system that includes at least one neural network (Abstract), comprising the steps of: receiving input data captured by the apparatus (Fig.1, 106, Fig.2, 230A, 230B,Fig.3, 302; [0021] [0022] [0034 – 0036] [0043]); processing the input data, using the at least one neural network (primary neural net, Fig.1, 112; [0023 – 0025]) of the control system (Fig.1, 104 and Fig.2, 220; [0020] [0031] [0039]) to obtain an output (A primary neural network in a data analyzer processes the input data to generate an output., Fig.3, 304; [0023] [0036] [0037] [0044]), the at least one neural network of the control system having at least one hidden layer (The primary neural network comprises a set of hidden neurons., [0029]); determining, using a second neural network (inspection neural network in an analyzer inspector, Fig.1, 120 and 122; [0027] [0028]) if a confidence level associated with the output is lower than a predetermined level (The inspection neural network receives the input, intermediate values generated by the primary neural network and output generated by the primary neural network. The inspection neural network generates a reliability indicator which is a value that indicates reliability/confidence of the output of the primary neural network. The reliability indicator is compared to a threshold., Fig.3, 306 and 308; [0027] [0030] [0038] [0040] [0045 – 0047]); and controlling the apparatus using the output if the confidence level is not lower than the predetermined level (“The controller 104 may monitor the output 114 and reliability indicator 126 and alter the operation conditions of a dynamic system under its control … The navigator 220 may use the output received from the data analyzers 240A and 240B and the reliability indicators received from the analyzer inspectors 250A and 250B to make decisions about vehicle movements. For example, the output from data analyzers 240A and 240B may indicate a confidence map of drivable regions in an image that represents the frontal view of the vehicle 200. The navigator may receive repeated confidence maps from the data analyzers 240A and 240B and use the confidence maps to determine the direction of the road as the vehicle 200 moves forward. The reliability indicators generated by the analyzer inspectors 250A and 250B may be used in a number of ways. In one example, if the indicator falls below a threshold, the navigator 220 may simply ignore the accompanying output. Where the output comprises confidence maps of drivable regions, the navigator 220 may ignore the particular confidence map and wait for the next confidence map. In some cases, the navigator 220 may cause the vehicle to slow down until it begins to receive confidence maps with better reliability indicators.”, [0031] [0039] [0040] [0047]).
Yet, Hu fails to teach that the second neural network has at least one hidden layer.
However, Hu further discloses that a neural network may have at least one hidden layer (A primary neural network comprises a set of hidden neurons., [0029]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to apply the neural network architecture disclosed by Hu’s first neural network to the second neural network so that the second neural has at least one hidden layer for the purpose of increasing the accuracy of the classification/prediction performed by the second neural network.
For claim 31, Hu discloses An apparatus being operated in part by a controller, comprising: an input device (sensors, Fig.1, 102 and Fig.2, 230A, 230B) constructed to generate input data ([0021] [0022] [0034 – 0036] [0043]); at least one neural network (primary neural net, Fig.1, 112; [0023 – 0025]) coupled to the controller (Fig.1, 104 and Fig.2, 220; [0020] [0031] [0039]) and constructed to receive the input data and to generate an output (A primary neural network in a data analyzer processes the input data to generate an output., Fig.3, 304; [0023] [0036] [0037] [0044]), the at least one neural network having at least one hidden layer (The primary neural network comprises a set of hidden neurons., [0029]); and a second neural network (inspection neural network in an analyzer inspector, Fig.1, 120 and 122; [0027] [0028]) constructed to use the output from the at least one neural network to determine whether a confidence level associated with the output is below a predetermined level (The inspection neural network receives the input, intermediate values generated by the primary neural network and output generated by the primary neural network. The inspection neural network generates a reliability indicator which is a value that indicates reliability/confidence of the output of the primary neural network. The reliability indicator is compared to a threshold., Fig.3, 306 and 308; [0027] [0030] [0038] [0040] [0045 – 0047]), wherein the controller is further constructed to operate the apparatus using the output unless the confidence level of the output from the at least one neural network is below the predetermined level (“The controller 104 may monitor the output 114 and reliability indicator 126 and alter the operation conditions of a dynamic system under its control … The navigator 220 may use the output received from the data analyzers 240A and 240B and the reliability indicators received from the analyzer inspectors 250A and 250B to make decisions about vehicle movements. For example, the output from data analyzers 240A and 240B may indicate a confidence map of drivable regions in an image that represents the frontal view of the vehicle 200. The navigator may receive repeated confidence maps from the data analyzers 240A and 240B and use the confidence maps to determine the direction of the road as the vehicle 200 moves forward. The reliability indicators generated by the analyzer inspectors 250A and 250B may be used in a number of ways. In one example, if the indicator falls below a threshold, the navigator 220 may simply ignore the accompanying output. Where the output comprises confidence maps of drivable regions, the navigator 220 may ignore the particular confidence map and wait for the next confidence map. In some cases, the navigator 220 may cause the vehicle to slow down until it begins to receive confidence maps with better reliability indicators.”, [0031] [0039] [0040] [0047]).
Yet, Hu fails to teach that the second neural network has at least one hidden layer.
However, Hu further discloses that a neural network may have at least one hidden layer (A primary neural network comprises a set of hidden neurons., [0029]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to apply the neural network architecture disclosed by Hu’s first neural network to the second neural network so that the second neural has at least one hidden layer for the purpose of increasing the accuracy of the classification/prediction performed by the second neural network.
For claims 22 and 32, Hu further discloses determining the confidence level associated with the output ([0029] [0030] [0036 – 0038]).
For claims 29 and 39, Hu further discloses, wherein the apparatus includes an autonomous land vehicle (Fig.2, 200; [0034]) and the step of using the obtained output further includes the step of generating a signal to control the autonomous land vehicle ([0039] [0040]).
Claim(s) 23, 26, 33 and 36 is/are rejected under 35 U.S.C. 103 as being unpatentable over Hu et al. (US 2018/0157934) (“Hu”) in view of Paquet et al. (US 2012/0158620) (“Paquet”).
For claim 23, Hu fails to teach the following: re-training the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
However, Paquet discloses a system and method for training automated classifiers (artificial neural networks) (Abstract), comprising the following: re-training an artificial neural network (automated classifier) when a confidence level (score) is below a pre-determined level (Fig.4, 12, 14, 16, 18, 32, 54 and Fig.5, 70, 72, 24, 76; [0030] [0031]), wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (image data, [0021]) that is different from what was originally used to train the at least one neural network (The training data used to retrain the artificial neural network comprises content set (content item A and content item C) which is different than the content set (content item1, content item 2, content item 3) used to train the artificial network., Fig. 4, initial training set and supplemental training set; [0030]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve Hu’s invention in the same way that Paquet’s invention has been improved to achieve the following, predictable results for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006]): further re-training the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (comprising data, e.g. image data) (Hu, [0035] [0036] [0061]) that is different from what was originally used to train the at least one neural network.
For claim 33, Hu fails to teach the following: wherein the controller is further structured to re- train the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data that is different from what was originally used to train the at least one neural network.
However, Paquet discloses a system and method for training automated classifiers (artificial neural networks) (Abstract), comprising executing the following on a processor (Fig.5, 64): re-training an artificial neural network (automated classifier) when a confidence level (score) is below a pre-determined level (Fig.4, 12, 14, 16, 18, 32, 54 and Fig.5, 70, 72, 24, 76; [0030] [0031]), wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (image data, [0021]) that is different from what was originally used to train the at least one neural network (The training data used to retrain the artificial neural network comprises content set (content item A and content item C) which is different than the content set (content item1, content item 2, content item 3) used to train the artificial network., Fig. 4, initial training set and supplemental training set; [0030]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve Hu’s invention in the same way that Paquet’s invention has been improved to achieve the following, predictable results for the purpose of improving performance of the neural network when results generated by the neural network are unacceptably low (Paquet, [0004] [0006]): further re-training, using a the controller (processor), the at least one neural network when the confidence level is below the predetermined level, wherein re-training the at least one neural network comprises retraining the at least one neural network using a set of training data (comprising data, e.g. image data) (Hu, [0035] [0036] [0061]) that is different from what was originally used to train the at least one neural network.
For claims 26 and 36, Hu and Paquet further disclose, wherein the re-training utilizes supervised learning (Hu, Training the at least one neural network, i.e. PNN, using ground truth labels as supervised learning, [0048] [0050]) (Paquet, Adjusting the weights of the interconnections among the neurons of the neural network until a known correct output is generated as re-training using supervised learning., [0001] [0023] [0030] [0031] [0038]).
Claim(s) 30 and 40 is/are rejected under 35 U.S.C. 103 as being unpatentable over Hu et al. (US 2018/0157934) (“Hu”) in view of Griesmeyer (US 2016/0323281) and further in view of Chen et al. (US 2017/0024660) (”Chen”).
For claims 30 and 40, Hu fails to teach, wherein the input data comprises an image, and the confidence level lower than the predetermined level indicates the image is inappropriate.
However, Griesmeyer discloses a system and method for filtering content in an online system (Abstract), comprising the following: a classification model (Fig.5, 540; [0026] [0027]) receives an image input (Fig.5, 520, 530); and output a confidence level, wherein the confidence level (value) exceeding a predetermined level (threshold) indicates an image is inappropriate (not safe for work) ([0013 – 0015] [0026] [0027]).
Additionally, Chen discloses a system and method for the purpose of generating, training, or refining machine learning classifier models (Abstract), wherein a confidence level (value) exceeds a threshold by being either greater than or less than the threshold ([0026]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve Hu’s invention in the same way that Griesmeyer’s invention has been improved to achieve the following, predictable results for the purpose of providing and ensuring compliance of content with a content policy in an online content system (Griesmeyer, [0013 – 0015]) by automating processes with a reliable use of neural networks (Hu, [0003]): the at least one first neural network is a classification model in an online content system (Hu, The PNN/INN may be implemented in a system without a controller. A person of ordinary skill would understand that the data analyzer 110, PNN 112, analyzer inspector 120, and INN 122 are components that may be implemented in a variety of data analysis systems. These systems may be implemented using numerous combinations of components to provide a variety of features, [0033]); the input data further comprises an image; and the confidence level indicates the image is inappropriate by exceeding a predetermined level.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the invention disclosed by the combination of Hu and Griesmeyer in the same way that Chen’s invention has been improved to achieve the following, predictable results for the purpose of providing and ensuring compliance of content with a content policy in an online content system (Griesmeyer, [0013 – 0015]) by automating processes with a reliable use of neural networks (Hu, [0003]): the confidence level exceeding the predetermined level further comprises being less than the predetermined level.
Claim(s) 28 and 38 is/are rejected under 35 U.S.C. 103 as being unpatentable over Hu et al. (US 2018/0157934) (“Hu”) in view of Tamrakar et al. (US 2021/0129748) (“Tamrakar”)
For claims 28 and 38, Hu fails to teach, wherein the apparatus includes a human speech generator and the step of using the output further includes the step of generating human speech.
However, Tamrakar discloses an evaluation engine has two or more modules to assist a driver of a vehicle (Abstract), comprising the following: receiving input data captured by the apparatus ([0021] [0025] [0040] ); processing the input data, using the at least one neural network ([0030] [0031] [0037] [0040]) of a control system, to obtain an output ([0040]); and generating human speech using the output and a human speech generator ([0041] [0042]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve Hu’s invention in the same way that Tamrakar’s invention has been improved to achieve the following, predictable results for the purpose of providing a driver monitoring and response system (Tamrakar, [0003] [0003]) by automating processes with a reliable use of neural networks (Hu, [0003]): the at least one first neural network is a module in the driver monitoring system that receives input data and generates an output. Furthermore, the output is used to generate human speech using a human speech generator which is coupled to the first neural network.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Scharfenberger et al. (US 2020/0247433) (Inventive concept of testing the output of a neural network that detects of classifies objects)
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SONIA L GAY whose telephone number is (571)270-1951. The examiner can normally be reached Monday-Friday 9-5 ET.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Daniel Washburn can be reached at 571-272-5551. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/SONIA L GAY/Primary Examiner, Art Unit 2657
Prosecution Insights