DETAILED 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 .
Status of the Claims
Currently, claims 1-10 are pending in the application. Claims 1, 4-6 & 8-10 are amended.
Response to Arguments / Amendments
Applicant’s arguments have been fully considered but are rendered moot in view of the new ground of rejection necessitated by amendments initiated by the applicant.
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 of this title, 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 1-3 & 6-10 are rejected under 35 U.S.C. 103 as being unpatentable over Zappella et al. (US 20250013899, hereinafter Zappella) in view of Lambert et al. (US 20240037367, hereinafter Lambert) and Donderici (US 20230382407, hereinafter Donderici)
Regarding Claim 1, Zappella in view of Lambert discloses a method of integrally optimizing parameters ([0042] FIG. 2, Bayesian hyperparameters optimization), the method comprising:
performing training on a machine learning model by selecting ([0042] FIG. 1, FIG. 2, Step 200, determining an objective function to optimize and is validation error of a model 113 of a model training system 130 undergoing training using a Bayesian optimizer to optimize hyperparameters); and
determining, among the selected ([0042] FIG. 1, FIG. 2, the model training system 130 undergoing training using the Bayesian optimizer to optimize hyperparameters 140 of the model training system 130 in order to minimize validation error),
wherein the performing of the training on the machine learning model includes:
selecting the ([0044] FIG. 2, Step 210, initialize probabilistic models of the objective function and constraint functions including evaluation of the objective function at one or more points. Select one of more points may be selected through a random search and determine a set of operational metrics that are provided to the Bayesian optimizer); and
performing training on the machine learning model ([0041] FIG. 1, the Bayesian optimizer 150 employ probabilistic models 155, in combination with an analysis of Constraints 112 and Metrics 145 at Constraint evaluator 152, to determine hyperparameters 140 and direct the model training system 130 to use the determined set of hyperparameters to perform a training operation using the training data 111 to generate additional metrics 145).
Zappella does not explicitly disclose integrally optimizing sensor parameters
Lambert teaches optimizing sensor parameters ([0150], controller(s) 936 provide signals for controlling one or more components and/or systems of vehicle 900 in response to sensor data received from one or more sensors (e.g., sensor inputs)
Therefore, it would have been obvious to one ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of integrally optimizing sensor parameters as taught by Lambert ([0150]) into the machine learning system of Zappella in order to provide systems for improving control of systems with high accuracy, efficiency and safety (Lambert, [0003]).
Zappella & Lambert do not explicitly disclose training a model on newly generated sensor data.
Donderici teaches training a model on newly generated sensor data ([0039], the learning datastore 310 stores store training data sets, validation data sets, hyperparameters used to train control models, hyperparameters used to compress control models, internal parameters (e.g., weights of filters) of control models, or other data associated with AV control models - storing environmental sensor data collected by the AVs 110. For example, images obtained by exterior cameras of the AVs may be used to learn information about the AVs' environments. The sensor data may be processed to identify particular objects in the environment)
Therefore, it would have been obvious to one ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of integrally optimizing sensor parameters as taught by Donderici ([0039]) into the machine learning system of Zappella & Lambert in order to enable utilizing an autonomous vehicle (AV) system to use information from a Global Positioning System (GPS), navigation systems, vehicle-to-vehicle communication, vehicle to infrastructure technology, and/or drive-by-wire systems to navigate the vehicle efficiently (Donderici , [0016]).
Regarding Claim 2, Zappella in view of Lambert and Donderici discloses the method of claim 1, Lambert discloses wherein the sensor parameter includes at least one of a sampling frequency, a measurement range, and sensor sensitivity ([0148], a brake sensor system 946 may be used to operate vehicle brakes in response to receiving signals from brake actuator(s) 948 and/or brake sensors; [0149], one or more onboard (e.g., integrated) computing devices that process sensor signals, to enable autonomous driving and/or to assist a human driver in driving vehicle 900; [0155], cameras with an RCCC, an RCCB, and/or an RBGC color filter array to increase light sensitivity). The same reason or rational of obviousness motivation applied as used above in claim 1.
Regarding Claim 3, Zappella in view of Lambert and Donderici discloses the method of claim 1, Zappella discloses wherein the machine learning model hyperparameter includes at least one of an epoch, a batch size, and a learning rate ([0040] FIG. 1, optimization of hyperparameters for training of a machine learning system with constraints and submit various types of requests to the machine learning system 110 to train and/or execute machine learning models with constraints).
Regarding Claim 6, Analogous rejection as the rejection of Claim 1 applies.
Lambert further discloses preprocessing filter ([0018], A Gaussian process surrogate model may be created for y(x) and iteratively updated by evaluating the black-box function at new points. Points are selected by optimizing an acquisition function which trades off exploration and exploitation. For example, for a black-box function of the validation error of a deep neural network (DNN) as a function of hyperparameters x, the DNN may be trained using newly selected points to determine various operational metrics).
Regarding Claim 7, Zappella in view of Lambert and Donderici discloses the method of claim 6, Lambert discloses wherein the preprocessing filter includes at least one of an interval average filter, a Gaussian filter, a maximum value filter, and a minimum value filter ([0018], A Gaussian process surrogate model may be created for y(x) and iteratively updated by evaluating the black-box function at new points. Points are selected by optimizing an acquisition function which trades off exploration and exploitation. For example, for a black-box function of the validation error of a deep neural network (DNN) as a function of hyperparameters x, the DNN may be trained using newly selected points to determine various operational metrics).
Regarding Claim 8, Analogous rejection as the rejection of Claim 1 applies.
Zappella in view of Lambert further discloses indicating for example how many training iterations have been run, current status of determined metrics, such as quality of the objective function and operational constraint values, and/or the current sampling weights assigned to the different training examples (Zappella: [0065]) and inference and/or training logic 615 may be used in system for inferencing or predicting operations based, at least in part, on weight parameters calculated using neural network training operations, neural network functions and/or architectures, or neural network use cases described herein (Lambert: [0153]. FIG. 9A )
Regarding Claim 9, Zappella in view of Lambert and Donderici discloses the method of claim 8, Zappella in view of Lambert discloses wherein the performing of the training on the machine learning model includes generating training data by applying the weight to the sensor data, and performing training on the machine learning model using the training data (Zappella: [0065], indicating for example how many training iterations have been run, current status of determined metrics, such as quality of the objective function and operational constraint values, and/or the current sampling weights assigned to the different training examples; Lambert: [0153]. FIG. 9A inference and/or training logic 615 may be used in system for inferencing or predicting operations based, at least in part, on weight parameters calculated using neural network training operations, neural network functions and/or architectures, or neural network use cases described herein)
Regarding Claim 10, Zappella in view of Lambert and Donderici discloses the method of claim 8, Zappella in view of Lambert discloses wherein the performing of the training on the machine learning model includes performing training on the machine learning model by selecting a sensor providing sensor data used for training from a sensor candidate group, and the optimized weight is an optimized weight for the sensor data provided by the selected sensor wherein the selected sensor generates the sensor data in accordance with a selected sensor parameter applied to the selected sensor. (Zappella: [0065], indicating for example how many training iterations have been run, current status of determined metrics, such as quality of the objective function and operational constraint values, and/or the current sampling weights assigned to the different training examples; Lambert: [0153]. FIG. 9A inference and/or training logic 615 may be used in system for inferencing or predicting operations based, at least in part, on weight parameters calculated using neural network training operations, neural network functions and/or architectures, or neural network use cases described herein)
Claims 4-5 are rejected under 35 U.S.C. 103 as being unpatentable over Zappella et al. (US 20250013899, hereinafter Zappella) in view of Lambert et al. (US 20240037367, hereinafter Lambert), Donderici and of Gullikson et al. (US 20230110056, hereinafter Gullikson)
Regarding Claim 4, Zappella in view of Lambert and Donderici discloses the method of claim 1, but does not explicitly disclose wherein the performing of the training on the machine learning model includes performing training on the machine learning model by selecting preprocessing filters used for preprocessing the sensor data from a preprocessing filter candidate group until the termination condition is satisfied, and the determining the optimized sensor parameter and optimized machine learning model hyperparameter includes determining an optimized preprocessing filter that minimize the loss value among the selected preprocessing filters wherein the sensor data is generated by a sensor in accordance with the selected sensor parameters applied to the sensor.
Gullikson teaches wherein the performing of the training on the machine learning model includes performing training on the machine learning model by selecting preprocessing filters used for preprocessing the sensor data from a preprocessing filter candidate group until the termination condition is satisfied, and the determining the optimized sensor parameter and optimized machine learning model hyperparameter includes determining an optimized preprocessing filter that minimize the loss value among the selected preprocessing filters wherein the sensor data is generated by a sensor in accordance with the selected sensor parameters applied to the sensor. ([0050], FIG. 1, preprocessor 104 modifies and supplements the sensor data to generate preprocessed data for an anomaly detection model 106 such as filtering operations as filtering operations to remove outlying data samples, to reduce or limit bias (e.g., due to sensor drift or predictable variations), to remove sets of samples associated with particular events (such as data samples during a start-up period or during a known failure event), denoising, etc.)
PNG
media_image1.png
550
612
media_image1.png
Greyscale
Therefore, it would have been obvious to one ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of selecting preprocessing filters as taught by Gullikson ([0050]) into the machine learning system of Zappella & Lambert in order to provide systems for automatically generating and training the trained behavior model based on historic data, reducing time and expense involved in process with multiple normal operational states and downstream effects of errors introduced by imputation of residual values and allowing the risk score calculator to calculate risk scores based on residual data corresponding to imputed values (Gullikson, [0004]).
Regarding Claim 5, Zappella in view of Lambert and Donderici discloses the method of claim 1, but does not explicitly disclose wherein the performing of the training on the machine learning model includes performing training on the machine learning model by selecting a sensor providing sensor data used for training from a sensor candidate group until the termination condition is satisfied, and the optimized sensor parameter is an optimized sensor parameter for the selected sensor wherein the selected sensor generates the sensor data in accordance with the selected sensor parameters applied to the selected sensor.
Gullikson teaches wherein the performing of the training on the machine learning model includes performing training on the machine learning model by selecting a sensor providing sensor data used for training from a sensor candidate group until the termination condition is satisfied, and the optimized sensor parameter is an optimized sensor parameter for the selected sensor wherein the selected sensor generates the sensor data in accordance with the selected sensor parameters applied to the selected sensor. ([0050], FIG. 1, preprocessor 104 modifies and supplements the sensor data to generate preprocessed data for an anomaly detection model 106 such as filtering operations as filtering operations to remove outlying data samples, to reduce or limit bias (e.g., due to sensor drift or predictable variations), to remove sets of samples associated with particular events (such as data samples during a start-up period or during a known failure event), denoising, etc.)
Therefore, it would have been obvious to one ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of selecting preprocessing filters as taught by Gullikson ([0050]) into the machine learning system of Zappella & Lambert in order to provide systems for automatically generating and training the trained behavior model based on historic data, reducing time and expense involved in process with multiple normal operational states and downstream effects of errors introduced by imputation of residual values and allowing the risk score calculator to calculate risk scores based on residual data corresponding to imputed values (Gullikson, [0004]).
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action.
THIS ACTION IS MADE FINAL. 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 Samuel D Fereja whose telephone number is (469)295-9243. The examiner can normally be reached 8AM-5PM.
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, DAVID CZEKAJ can be reached at (571) 272-7327. 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.
/SAMUEL D FEREJA/Primary Examiner, Art Unit 2487
