DETAILED ACTION
This office action is issued in response to communication filed on 11/11/2025. Claims 1-17 are pending in this Office Action.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
The objection to the title and claim 6 has been withdrawn by the examiner.
Applicant’s amendments overcome the 101 and 112 rejection. Accordingly, the rejections have been withdrawn.
Response to Arguments
Applicant's arguments filed on 11/11/2025 with respect to rejection of claim 1 under 35 USC 103 have been considered but are moot in view of the new ground of rejection.
Applicant's arguments filed on 11/11/2025 with respect to rejection of claim 10 under 35 USC 103 have been considered and are not persuasive. The examiner respectfully traverses applicant’s arguments.
Applicant argues: “Therefore, in an information processing device recited in amended Applicant's claim 10, a "first learned model" and a "second learned model" are different models obtained through different processes.(Applicant argument at page 18)
The examiner respectfully disagrees. The In response to applicant's argument that the references fail to show certain features of applicant’s invention, it is noted that the features upon which applicant relies (i.e., first model and second model are different models) are not recited in the rejected claim(s). Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993). The second model in Petousis also different than the first model because the second model not only predicts the future values of vehicle sensor data, it also calculates error data and using the error data to correct the predicted future values of vehicle sensor data as shown in Petousis par [0053].
Applicant argues: “(2) In Petousis, the "2nd ML Unit" appears to merely predict future values of vehicle sensor data (S250). In contrast, amended Applicant's claim 10 recites "generate a second learned model by performing machine learning based on the training input data and difference data between output data generated by inputting the training input data to the first learned model, and the training correct data." Therefore, in an information processing device recited in proposed amended Applicant's claim 10, a second learned model predicts a difference data between an output data generated by inputting a training input data to a first learned model and the training correct data” (Applicant’s arguments at page 18)
The examiner respectfully disagrees because one cannot show non-obviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986). The examiner relies on Yan for the teaching of “approximation function generated based on training input data and training correct data corresponding to the training input data”, not Petousis alone as applicant asserts. Petousis [0053] teaches providing as input a subset of the vehicle sensor data into a second machine learning unit, calculating error data and using the error data to correct the predicted future values of vehicle sensor data. Accordingly, Petousis and Yan teach the claimed limitation of "generate a second learned model by performing machine learning based on the training input data and difference data between output data generated by inputting the training input data to the first learned model, and the training correct data" as recited in claim 10.
Applicant’s remaining arguments with respect to claims are substantially encompassed in the argument above, therefore examiner responds with the same rationale as stated above.
For at least the foregoing reasons, the examiner maintains prior art rejection of claims 10-13.
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.
Claims 1-4 ,7-9 and 14-16 are rejected under 35 U.S.C. 103 as being unpatentable over Petousis et al.(US Patent Application Publication 2020/0097841 A1, hereinafter “Petousis”) , in view of Yan et al.(US Patent Application Publication 2019/0367019 A1, hereinafter “Yan”) and further in view of Watanabe.(US Patent 5,095,443, hereinafter “Watanabe”)
As to claim 1, Petousis teaches an information processing device for controlling a machine whose characteristics change overtime comprising: processor circuitry configured to:
acquire sensor information from the machine as reference input data; (Petousis par [0053] teaches receiving vehicle sensor data S210)
generate first predicted output data by inputting the reference input data to a first approximation function [wherein the first approximate function is generated based on training input data and training correct data corresponding to the training input data]; (bold phasis added. Petousis par [0053] teaches receiving vehicle sensor data S210, providing as input a subset of the vehicle sensor data into a first machine learning unit implementing machine learning algorithm S220. Calculating error data between values of the predicted vehicle sensor data S260 and the values of the actual vehicle sensor data S260))
Petousis teaches approximate function but fails to expressly teach approximation function is generated based on training input data and training correct data corresponding to the training input data.
However, Yan teaches “. (Yan par [0046] teaches offline training phase for training and building a proximate vehicle intention prediction system. In training phase, the training data collection system 201 can be used to generate, train and configure the intention prediction model 173.Yan par [0047] teaches training data collection 201 can include a plurality of training data gathering mechanisms including obtaining training data or training images from a library or human driving database. Corresponding ground truth data can also be gathered by the training data collection system 201)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teaching of Petousis and Yan to achieve the claimed invention. One would have been motivated to make such combination to accurately train a machine learning model.(Yan par [0044])
Petousis and Yan further teach generate second predicted output data by inputting the reference input data to a second learned model, wherein the second learned model is generated by performing machine learning based on the training input data and difference data between output data generated by inputting the training input data to the first approximation function, and the training correct data (Petousis par [0053] teaches providing as input a subset of the vehicle sensor data into a second machine learning unit at a remote computing platform S240 and the second machine learning unit using the error data S270 (interpreted as difference data)) ;
[ generate final output data for controlling the machine by adding the first predicted output data and the second predicted output data];
acquire reference correct data (Petousis par [0053] teaches the values of the actual vehicle sensor data S260); and
update the second learned model by performing machine learning based on the reference input data and the difference data between the first predicted output data and the reference correct data, and the reference input data to adapt to the changes in the characteristics of the machine while a certain level of accuracy of the final output data is ensured. (Petousis par [0053] teaches correcting the future values of vehicle sensor data predicted by the second machine learning unit using the error data S270. The examiner interprets the language of “to adapt to the changes in the characteristics of the machine while a certain level of accuracy of the final output data is ensured” is non-functional descriptive material)
Petousis and Yan fail to expressly teach generate final output data for controlling the machine by adding the first predicted output data and the second predicted output data.
However, Watanabe teaches generate final output data for controlling the machine by adding the first predicted output data and the second predicted output data.( Watanabe col 3, lines 4-15 teaches letting the second neural network learn the correspondence so that the sum of the results of learning derived from the first and second neural networks becomes close to the teacher data)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teaching of Petousis, Yan and Watanabe to achieve the claimed invention. One would have been motivated to make such combination to improve neural network structure and neural network learning. (Watanabe col 2, lines 14-16)
As to claim 2, Petousis, Yan and Watanabe teach the information processing device according to Claim 1, wherein the first approximation function is a first learned model generated by performing machine learning based on the training input data and the training correct data. (Yan par [0046] teaches offline training phase for training and building a proximate vehicle intention prediction system . In training phase, the training data collection system 201 can be used to generate, train and configure the intention prediction model 173. Yan par [0047] teaches training data collection 201 can include a plurality of training data gathering mechanisms including obtaining training data or training images from a library or human driving database. Corresponding ground truth data can also be gathered by the training data collection system 201)
As to claim 3, Petousis, Yan and Watanabe teach the information processing device according to Claim 1, wherein the first approximation function is a function that formulates a relationship between the training input data and the training correct data. (Yan par [0049] teaches because the intention prediction model 173 is trained using real world, human behavior data, the predicted intention, behavior and trajectories of the vehicles or objects produced by the intention prediction model 173 are closely correlated to the actual intention , behavior, and trajectories of vehicles in real world environments with human drivers and based on a human driver behavior model implemented by the training data collection system 201)
As to claim 4, Petousis, Yan and Watanabe teach the information processing device according to Claim 1, wherein the processor circuitry is further configured to comprising output limiting the second predicted output data to a predetermined value range. ( Petousis par [0081] teaches error threshold)
Claims 7-9 merely recite a system and method and non-transitory computer readable medium with similar features of claim 1 and therefore being rejected for the same rationale as indicates in the above rejection of claim 1.
As to claim 14, Petousis, Yan and Watanabe teach the information processing device according to Claim 1, wherein the training input data and the training correct data are data obtained from the machine before operation of the machine. (Yan par [0047] teaches the training data collection system can include a plurality of training data gathering mechanism including obtaining training data or training images from a library or human driving database, and obtaining training data for images from plurality of sensors)
As to claim 15, Petousis, Yan and Watanabe teach the information processing device according to Claim 1, wherein the training input data comprises sensor information from the machine and the training correct data comprises data for controlling the machine. (Yan par [0047] teaches the training data collection system can include a plurality of training data gathering mechanism including obtaining training data or training images from a library or human driving database, and obtaining training data for images from plurality of sensors. Yan par [0049] teaches the image data and other perception data , ground truth data, context data, and other training data collected by the training data collection system reflects truly realistic, real-world traffic information related to the locations or routings, the scenarios and the driver actions, behaviors and intentions being monitored)
As to claim 16, Petousis, Yan and Watanabe teach the information processing device according to Claim 15, wherein the first approximate function is obtained through supervised learning using the training correct data as teacher data, and the second learned model is obtained through supervised learning using the difference data as teacher data. (Petousis par [0046] teaches supervised learning . Petousis par [0053] teaches correcting the future values of vehicle sensor data predicted by the second machine learning unit using the error data S270)
Claims 5-6 are rejected under 35 U.S.C. 103 as being unpatentable over Petousis, Yan, Wantanabe and further in view of Erlandson et al.(US Patent Application Publication 2019/0147357 A1, hereinafter “Erlandson”)
As to claim 5, Petousis, Yan and Watanabe teach the information processing device according to Claim 1 but fail to teach wherein the processor circuitry is further configured to perform judgement under a predetermined condition based on the second predicted output data and, if the predetermined condition is met, generates predetermined information to be presented to a user.
However, Erlandson teaches further configured to perform judgement under a predetermined condition based on the second predicted output data and, if the predetermined condition is met, generates predetermined information to be presented to a user.
. (Erlandson par [0040] teaches the predetermined criterion may be a probability threshold value such as 90% that identifies the particular threshold provability above which the alert should be generated and based on the alert, it is time to retrain the predictive learning model or take some other action)
Therefore, it would have been obvious to one of ordinary kill in the art before the effective filing date of the claimed invention was made to combine the teaching of Petousis, Yan , Watanabe and Erlandson to achieve the claimed invention. One would have been motivated to make such combination to provide the benefit of determining when a predictive learning model should be retrained on current operational data. (Erlandson par [0023])
As to claim 6, Petousis, Yan , Watanabe and Erlandson and teach the information processing device according to Claim 5, wherein the predetermined information is information on a timing of maintenance of a device from which the reference input data and the reference correct data have been acquired.( Erlandson par [0040] teaches based on the alert, it is time to retrain the predictive learning model or take some other action)
4. Claim 17 is rejected under 35 U.S.C. 103 as being unpatentable over Petousis, Yan and Watanabe and further in view of Kline et al.(US Patent Application Publication 2020/0130688 A1, hereinafter “Kline”)
As to claim 17, Petousis, Yan and Watanabe teach the information processing device according to Claim 15, wherein the machine is an automobile (Yan par [0019] teaches user’s vehicle),
Petousis, Yan and Watanabe fail to expressly teach the training input data comprises rotation speed of tires of the automobile, the training correct data comprises braking distance of the automobile, the reference input data comprises rotation speed of tires of the automobile, and reference correct data comprises braking distance of the automobile.
However, Kline teaches vehicle data includes rotation speed of tires of the automobile and brake distance . (Kline par [0014] teaches brake distances and par [0016] teaches tire rotation speed)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the training data of Petousis, Yan and Watanabe with the brake distance and tire rotation speed as taught by Kline to achieve the claimed invention. One would have been motivated to make such combination to provide additional safety level for instances of dangerous situations that are not always apparent to a driver of the vehicle.(Kline par [0008])
Claims 10-13 are rejected under 35 U.S.C. 103 as being unpatentable over Petousis and further in view of Yan .
As to claim 10, Petousis teaches an information processing device comprising: first learning processor circuitry configured to generate a first learned model by performing machine learning based on training input data and training correct data (Petousis par [0053] teaches receiving vehicle sensor data S210, providing as input a subset of the vehicle sensor data into a first machine learning unit implementing machine learning algorithm S220);
Petousis teaches first learning model but fails to expressly teach by performing machine learning based on training input data and training correct data.
However, Yan teaches performing machine learning based on training input data and training correct data. (Yan par [0046] teaches offline training phase for training and building a proximate vehicle intention prediction system. In training phase, the training data collection system 201 can be used to generate, train and configure the intention prediction model 173.Yan par [0047] teaches training data collection 201 can include a plurality of training data gathering mechanisms including obtaining training data or training images from a library or human driving database. Corresponding ground truth data can also be gathered by the training data collection system 201)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teaching of Petousis and Yan to achieve the claimed invention. One would have been motivated to make such combination to accurately train a machine learning model.(Yan par [0044])
Petousis and Yan further teach second learning processor circuitry configured to generate a second learned model by performing machine learning based on the training input data and difference data between output data generated by inputting the training input data to the first learned model, and the training correct data.( Petousis par [0053] teaches providing as input a subset of the vehicle sensor data into a second machine learning unit at a remote computing platform S240 and the second machine learning unit using the error data S270 (interpreted as difference data))
Claims 11-12 merely recite a system and method with similar features of claim 10 and therefore being rejected for the same rationale as indicates in the above rejection of claim 10.
As to claim 13, Petousis teaches learned model application method comprising: generating a first learned model by performing machine learning based on training input data and training correct data (Petousis par [0042] teaches first machine learning unit 114)
Petousis teaches first learning model but fails to expressly teach by performing machine learning based on training input data and training correct data.
However, Yan teaches performing machine learning based on training input data and training correct data. (Yan par [0046] teaches offline training phase for training and building a proximate vehicle intention prediction system. In training phase, the training data collection system 201 can be used to generate, train and configure the intention prediction model 173.Yan par [0047] teaches training data collection 201 can include a plurality of training data gathering mechanisms including obtaining training data or training images from a library or human driving database. Corresponding ground truth data can also be gathered by the training data collection system 201)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teaching of Petousis and Yan to achieve the claimed invention. One would have been motivated to make such combination to accurately train a machine learning model.(Yan par [0044])
Petousis and Yan further teach generating a second learned model by performing machine learning based on the training input data and difference data between output data generated by inputting the training input data to the first learned model, and the training correct data (Petousis par [0042] teaches second machine learning unit 124); and
installing a predetermined device with the first learned model and the second learned model so that the second learned model can be updated based on data acquired from the device. (Petousis par [0042] teaches first machine learning unit 114 at the vehicle and second machine 124 at remote computing platform .Petousis par [0047] teaches “but can alternatively be performed entirely by remote computer system, the vehicle system or any other suitable system. Bold underlined emphasis added)
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to HIEN DUONG whose telephone number is (571)270-7335. The examiner can normally be reached Monday-Friday 8:00AM-5:00PM.
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, Viker Lamardo can be reached at 571-270-5871. 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.
/HIEN L DUONG/Primary Examiner, Art Unit 2147