Detailed Action
This Office Action is in response to the remarks entered on 10/20/2025. Claims 1-20 are currently pending.
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 .
Claim Objections
Amended claims were received on 10/20/2025. The claim objection is withdrawn.
Information Disclosure Statement
The information disclosure statements (IDS) submitted on August 25, 2025 and October 22, 2025 are in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statements are being considered by the examiner.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding claim 1,
2A Prong 1:
generating a training data set utilizing at least the vibrational information and a sound perception score associated with the vibrational information in a lab environment, (a mental process of judgment – selecting and combining information and score can be performed in the human mind)
based on in an end of line (EOL) environment. (a mental process of evaluation – scoring the characteristics based on given information)
2A Prong 2:
A computer-implemented method, comprising: (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f))
receiving sound information from one or more microphones and vibrational data from one or more accelerometers associated with an electrical drive of a device, wherein the sound information is associated with the device; (insignificant extra-solution activity MPEP 2106.05(g)(iii) of gathering statistics)
wherein the training data set is sent to an un-trained machine learning model, wherein the sound perception score is automatically programmed and the training data set includes accelerometer data that includes x-axis, y-axis, and z-axis information; (insignificant extra-solution activity MPEP 2106.05(g)(iii) of gathering statistics)
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model; (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f) – generic machine learning training process to apply an exception)
receiving real-time vibrational information from the electrical drive associated with the device; and (insignificant extra-solution activity MPEP 2106.05(g)(iii) of gathering statistics)
based on the trained machine learning model, outputting a score (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f))
The additional elements as disclosed above alone or in combination do not integrate the judicial exception into practical application as they are mere insignificant extra solution activity, combination of generic computer functions that are restricted to field of use are implemented to perform the disclosed abstract idea above.
2B:
A computer-implemented method, comprising: (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f))
receiving sound information from one or more microphones and vibrational data from one or more accelerometers associated with an electrical drive of a device, wherein the sound information is associated with the device; (indicated as an insignificant extra-solution activity MPEP 2106.05(g)(iii) in Step 2A Prong 2. Thus, the limitation is re-evaluated as well understood, routine and conventional MPEP 2106.05(d)(II)(i) of transmitting and receiving data over a network Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362)
wherein the training data set is sent to an un-trained machine learning model, wherein the sound perception score is automatically programmed and the training data set includes accelerometer data that includes x-axis, y-axis, and z-axis information; (indicated as an insignificant extra-solution activity MPEP 2106.05(g)(iii) in Step 2A Prong 2. Thus, the limitation is re-evaluated as well understood, routine and conventional MPEP 2106.05(d)(II)(i) of transmitting data over a network Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362)
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model; (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f) – generic machine learning training process to apply an exception)
receiving real-time vibrational information from the electrical drive associated with the device; and (indicated as an insignificant extra-solution activity MPEP 2106.05(g)(iii) in Step 2A Prong 2. Thus, the limitation is re-evaluated as well understood, routine and conventional MPEP 2106.05(d)(II)(i) of transmitting and receiving data over a network Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362)
based on the trained machine learning model, outputting a score (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f))
The additional elements as disclosed above in combination of the abstract idea are not sufficient to amount to significantly more than the judicial exception as they are well, understood, routine and conventional activity as disclosed in combination of generic computer functions and usage of elements that are restricted to field of use that are implemented to perform the disclosed abstract idea above.
Regarding claim 2,
2A Prong 1: Incorporates the rejection of claim 1.
2A Prong 2: wherein the trained machine learning model is trained utilizing only the vibrational information and minimizing a score prediction error output by the un-trained machine learning model. (mere instructions to apply an exception using a generic computer MPEP 2106.05(f) – generic training process of a machine learning model)
2B: wherein the trained machine learning model is trained utilizing only the vibrational information and minimizing a score prediction error output by the un-trained machine learning model. (mere instructions to apply an exception using a generic computer MPEP 2106.05(f) – generic training process of a machine learning model)
Regarding claim 3,
2A Prong 1:
freezing weights associated with the score prediction network and training the weights of the sound prediction network to minimize a weighted sum of sound and score prediction errors. (a mental process of judgment and evaluation – anyone who knows the art (e.g., a programmer) can manually correct the weight value of a neural network model based on error values)
2A Prong 2:
The computer-implemented method of claim 1, wherein the trained machine learning model is trained via an in-direct method, wherein a first neural network of the machine learning model is trained utilizing the sound information and a second neural network is trained to predict measured sound utilizing the vibrational information and obtain a predicted sound; (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f) – generic training process of machine learning models based on the first data and the second data)
feeding the predicted sound into a score prediction network to generate a human-perception score; and (an insignificant extra-solution activity MPEP 2106.05(g)(iii) of gathering statistics)
2B:
The computer-implemented method of claim 1, wherein the trained machine learning model is trained via an in-direct method, wherein a first neural network of the machine learning model is trained utilizing the sound information and a second neural network is trained to predict measured sound utilizing the vibrational information and obtain a predicted sound; (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f) – generic training process of machine learning models based on the first data and the second data)
feeding the predicted sound into a score prediction network to generate a human-perception score; and (was indicated as an insignificant extra-solution activity MPEP 2106.05(g)(iii) in Step 2A Prong 2. Thus, the limitation is re-evaluated in Step 2B as well understood, routine and conventional activity MPEP 2106.05(d)(II)(i) of transmitting data over a network Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362)
Regarding claim 4,
2A Prong 1: Incorporates the rejection of claim 1.
2A Prong 2: wherein the training data utilizes sound information and accelerometer data obtained from a noise-free environment. (a field of use and technological environment MPEP 2106.05(h))
2B: wherein the training data utilizes sound information and accelerometer data obtained from a noise-free environment. (a field of use and technological environment MPEP 2106.05(h))
Regarding claim 5,
2A Prong 1: The computer-implemented method of claim 1, wherein the sound perception scored is generated manually in response to the sound information. (a mental process of evaluation – manually scoring the sound information can be performed in the human mind)
2A Prong 2: This judicial exception is not integrated into a practical application.
2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Regarding claim 6,
2A Prong 1: Incorporates the rejection of claim 1.
2A Prong 2: wherein the vibrational information is accelerometer data. (a field of use and technological environment MPEP 2106.05(h))
2B: wherein the vibrational information is accelerometer data. (a field of use and technological environment MPEP 2106.05(h))
Regarding claim 7,
2A Prong 1: Incorporates the rejection of claim 1.
2A Prong 2: wherein the machine learning model is a U-Net or Transformer network. (a field of use and technological environment MPEP 2106.05(h))
2B: wherein the machine learning model is a U-Net or Transformer network. (a field of use and technological environment MPEP 2106.05(h))
Regarding claim 8,
2A Prong 1: Incorporates the rejection of claim 1.
2A Prong 2: wherein the real-time sound perception score is generated utilizing only the real-time vibrational information. (a field of use and technological environment MPEP 2106.05(h))
2B: wherein the real-time sound perception score is generated utilizing only the real-time vibrational information. (a field of use and technological environment MPEP 2106.05(h))
Regarding claim 9,
2A Prong 1: Incorporates the rejection of claim 1.
2A Prong 2: wherein the machine learning model is a deep learning network. (a field of use and technological environment MPEP 2106.05(h))
2B: wherein the machine learning model is a deep learning network. (a field of use and technological environment MPEP 2106.05(h))
Regarding claim 10,
2A Prong 1:
generating a training data set utilizing at least the first set of vibrational information and an associated sound perception score,
based on in an end of line (EOL) environment. (a mental process of evaluation – scoring the characteristics based on given information)
2A Prong 2:
A computer-implemented method, comprising: (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f))
receiving a first set of sound information from one or more microphones and a first set of vibrational information from one or more accelerometers associated with an electrical drive of a device in a first environment; (insignificant extra-solution activity MPEP 2106.05(g)(iii) of gathering statistics)
wherein the training data set is sent to an un-trained machine learning model, wherein the sound perception score is automatically programmed and the training data set includes accelerometer data that includes x-axis, y-axis, and z-axis information; (insignificant extra-solution activity MPEP 2106.05(g)(iii) of gathering statistics)
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model; (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f) – generic machine learning model training process)
receiving real-time vibrational information from the electrical device associated with the device in a second environment; and (insignificant extra-solution activity MPEP 2106.05(g)(iii) of gathering statistics)
based on the trained machine learning model, outputting a score (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f))
The additional elements as disclosed above alone or in combination do not integrate the judicial exception into practical application as they are mere insignificant extra solution activity, combination of generic computer functions that are restricted to field of use are implemented to perform the disclosed abstract idea above.
2B:
A computer-implemented method, comprising: (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f))
receiving a first set of sound information from one or more microphones and a first set of vibrational information from one or more accelerometers associated with an electrical drive of a device in a first environment; (indicated as an insignificant extra-solution activity MPEP 2106.05(g)(iii) in Step 2A Prong 2. Thus, the limitation is re-evaluated as well understood, routine and conventional MPEP 2106.05(d)(II)(i) of transmitting and receiving data over a network Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362)
wherein the training data set is sent to an un-trained machine learning model, wherein the sound perception score is automatically programmed and the training data set includes accelerometer data that includes x-axis, y-axis, and z-axis information; (indicated as an insignificant extra-solution activity MPEP 2106.05(g)(iii) in Step 2A Prong 2. Thus, the limitation is re-evaluated as well understood, routine and conventional MPEP 2106.05(d)(II)(i) of transmitting and receiving data over a network Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362)
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model; (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f) – generic machine learning model training process)
receiving real-time vibrational information from the electrical device associated with the device in a second environment; and (indicated as an insignificant extra-solution activity MPEP 2106.05(g)(iii) in Step 2A Prong 2. Thus, the limitation is re-evaluated as well understood, routine and conventional MPEP 2106.05(d)(II)(i) of transmitting and receiving data over a network Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362)
based on the trained machine learning model, outputting a score (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f))
The additional elements as disclosed above in combination of the abstract idea are not sufficient to amount to significantly more than the judicial exception as they are well, understood, routine and conventional activity as disclosed in combination of generic computer functions and usage of elements that are restricted to field of use that are implemented to perform the disclosed abstract idea above.
Regarding claim 11,
2A Prong 1: Incorporates the rejection of claim 10.
2A Prong 2: wherein the vibrational data includes accelerometer data. (a field of use and technological environment MPEP 2106.05(h))
2B: wherein the vibrational data includes accelerometer data. (a field of use and technological environment MPEP 2106.05(h))
Regarding claim 12,
2A Prong 1: Incorporates the rejection of claim 10.
2A Prong 2: wherein the machine learning model is a U-Net or Transformer network. (a field of use and technological environment MPEP 2106.05(h))
2B: wherein the machine learning model is a U-Net or Transformer network. (a field of use and technological environment MPEP 2106.05(h))
Regarding claim 13,
2A Prong 1: Incorporates the rejection of claim 10.
2A Prong 2: wherein the real-time sound perception score is generated utilizing only the real-time vibrational data. (a field of use and technological environment MPEP 2106.05(h))
2B: wherein the real-time sound perception score is generated utilizing only the real-time vibrational data. (a field of use and technological environment MPEP 2106.05(h))
Regarding claim 14,
2A Prong 1: Incorporates the rejection of claim 10.
2A Prong 2: wherein the machine learning model is a deep learning network. (a field of use and technological environment MPEP 2106.05(h))
2B: wherein the machine learning model is a deep learning network. (a field of use and technological environment MPEP 2106.05(h))
Regarding claim 15,
2A Prong 1: Incorporates the rejection of claim 10.
2A Prong 2: wherein the first environment is a laboratory environment and the second environment is an end-of-line factory environment. (a field of use and technological environment MPEP 2106.05(h))
2B: wherein the first environment is a laboratory environment and the second environment is an end-of-line factory environment. (a field of use and technological environment MPEP 2106.05(h))
Regarding claim 16,
2A Prong 1:
generate a training data set utilizing at least the vibrational information and a sound perception score associated with the corresponding sound of the vibrational information,
based on the real-time vibrational information as an input to the trained machine learning model, output a real-time sound perception score indicating characteristics associated with sound emitted from the device in an end of line (EOL) environment. (a mental process of evaluation – scoring the characteristics based on given information)
2A Prong 2:
A system, comprising: a processor, wherein the processor is programmed to: (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f))
receive sound information from one or more microphones and vibrational information from one or more accelerometers associated with a device in a first environment; (insignificant extra-solution activity MPEP 2106.05(g)(iii) of gathering statistics)
wherein the training data set is sent to an un-trained machine learning model, wherein the sound perception score is automatically programmed and the training data set includes accelerometer data that includes x-axis, y-axis, and z-axis information; (insignificant extra-solution activity MPEP 2106.05(g)(iii) of gathering statistics)
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model; (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f) – generic machine learning model training process)
receive real-time vibrational information from the device in a second environment; (insignificant extra-solution activity MPEP 2106.05(g)(iii) of gathering statistics)
The additional elements as disclosed above alone or in combination do not integrate the judicial exception into practical application as they are mere insignificant extra solution activity, combination of generic computer functions that are restricted to field of use are implemented to perform the disclosed abstract idea above.
2B:
A system, comprising: a processor, wherein the processor is programmed to: (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f))
receive sound information from one or more microphones and vibrational information from one or more accelerometers associated with a device in a first environment; (indicated as an insignificant extra-solution activity MPEP 2106.05(g)(iii) in Step 2A Prong 2. Therefore, the limitation is re-evaluated in Step 2B as well understood, routine and conventional activity MPEP 2106.05(d)(II)(i) of transmitting and receiving data over a network Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362)
wherein the training data set is sent to an un-trained machine learning model, wherein the sound perception score is automatically programmed and the training data set includes accelerometer data that includes x-axis, y-axis, and z-axis information; (indicated as an insignificant extra-solution activity MPEP 2106.05(g)(iii) in Step 2A Prong 2. Thus, the limitation is re-evaluated as well understood, routine and conventional MPEP 2106.05(d)(II)(i) of transmitting and receiving data over a network Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362)
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model; (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f) – generic machine learning model training process)
receive real-time vibrational information from the device in a second environment; (indicated as an insignificant extra-solution activity MPEP 2106.05(g)(iii) in Step 2A Prong 2. Thus, the limitation is re-evaluated as well understood, routine and conventional MPEP 2106.05(d)(II)(i) of transmitting and receiving data over a network Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362)
The additional elements as disclosed above in combination of the abstract idea are not sufficient to amount to significantly more than the judicial exception as they are well, understood, routine and conventional activity as disclosed in combination of generic computer functions and usage of elements that are restricted to field of use that are implemented to perform the disclosed abstract idea above.
Regarding claim 17,
2A Prong 1: Incorporates the rejection of claim 16.
2A Prong 2: wherein the vibrational information includes three- dimensional information. (a field of use and technological environment MPEP 2106.05(h))
2B: wherein the vibrational information includes three- dimensional information. (a field of use and technological environment MPEP 2106.05(h))
Regarding claim 18,
2A Prong 1:
2A Prong 2: wherein the processor is further programmed to generate (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f))
2B: wherein the processor is further programmed to generate (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f))
Regarding claim 19,
2A Prong 1: Incorporates the rejection of claim 16.
2A Prong 2: wherein machine learning model includes two or more neural networks utilized to output a real-time sound perception score. (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f) – using a model to generate sound perception score)
2B: wherein machine learning model includes two or more neural networks utilized to output a real-time sound perception score. (mere instructions to apply an exception using a generic computer component MPEP 2106.05(f) – using a model to generate sound perception score)
Regarding claim 20,
2A Prong 1: Incorporates the rejection of claim 16.
2A Prong 2: wherein the first environment and the second environment are not a same environment. (a field of use and technological environment MPEP 2106.05(h))
2B: wherein the first environment and the second environment are not a same environment. (a field of use and technological environment MPEP 2106.05(h))
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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 1-2, 5-6, 8-11, 13-18 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Krishnaveni et al. (Krishnaveni et al. “Analysis and control of the motor vibration using arduino and machine learning model”, 2021, hereinafter ‘Krishnaveni’) in view of Filimonov et al. (US 20240196144 A1, hereinafter ‘Filimonov’) in view of Mikhailov & Vladimirovish (US 20200211544 A1, hereinafter ‘Mikhailov’) and further in view of Gogoana & Mehta (US 20210102925 A1, hereinafter ‘Gogoana’).
Regarding claim 1, Krishnaveni teaches:
A computer-implemented method, comprising: ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] The Arduino device is the computer)
receiving sound information and vibrational data from one or more accelerometers associated with an electrical drive of a device, wherein the sound information is associated with the device; ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal (sound information associated with the device) is collected and stored in cloud, which is connected to the Arduino device, collected data is used to train and test the model. [Krishnaveni, page 2553, left col, 2.4. MEMS accelerometer, line 1-12] and [Krishnaveni, page 2553, right col, 4. Results and discussion, line 1-12] discloses utilizing MEMS accelerometer to collect vibrational data and to train the machine learning model)
generating a training data set utilizing at least the vibrational information in a lab environment, wherein the training data set is sent to an un-trained machine learning model, ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal is collected and stored in cloud, which is connected to the Arduino device, collected data (generated data in a lab environment) is used to train and test the model. [Krishnaveni, page 2553, right col, 2.4. MEMS accelerometer, line 1-14] discloses collecting X, Y, and Z axis data using the 3-axis accelerometer)
… outputting a trained machine learning model; ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal is collected and stored in cloud, which is connected to the Arduino device, collected data (generated data in a lab environment) is used to train and test the model. Additionally, [Krishnaveni, page 2554, left col, 4.2. Tested motor were divided into two groups, line 1 – right col, line 29] discloses training a neural network model to diagnose an induction motor)
receiving real-time vibrational information from the electrical drive associated with the device; and ([Krishnaveni, ABSTRACT, line 13-16] discloses generating real-time vibration measurements and analysis using the tool. [Krishnaveni, page 2552, left col, line 14 – last line] discloses connecting sensors to the motor (electrical drive associated with the device) to collect motor signal (vibrational information) )
based on the trained machine learning model and the real-time vibrational information, outputting a real-time with sound emitted from the device . ([Krishnaveni, ABSTRACT, line 13-16] discloses generating real-time vibration measurements and analysis using the tool. [Krishnaveni, page 2554, left col, 4.1. Performance analysis using decision tree, line 1-8] discloses generating classification and classification accuracy score based on the trained DT algorithm. [Krishnaveni, page 2552, left col, 2. Proposed model, line 1-6 and line 11-17] discloses training a DT model to perform vibration analysis and diagnosis. The motor vibration is collected from motors. Additionally, [Krishnaveni, page 2554, left col, 4.2. Tested motor were divided into two groups, line 1 – right col, line 29] discloses training a neural network model to diagnose an induction motor)
However, Krishnaveni does not specifically disclose:
receiving sound information from one or more microphones and vibrational data
generating a training data set utilizing at least the vibrational information and a sound perception score associated with the vibrational information in a lab environment, wherein the training data set is sent to an un-trained machine learning model, wherein the sound perception score is automatically programmed and the training data set includes accelerometer data that includes x-axis, y-axis, and z-axis information;
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model;
based on the trained machine learning model and the real-time vibrational information, outputting a real-time sound perception score indicating characteristics associated with sound emitted from the device in an end of line (EOL) environment.
Filimonov teaches:
receiving sound information from one or more microphones and [Filimonov, 0033] The sound measuring is performed using a sound recording device, such as a microphone. [Filimonov, 0049] discloses a training dataset comprising the first frequency response and a reference scoring (sound perception score))
generating a training data set utilizing at least the , wherein the training data set is sent to an un-trained machine learning model, wherein the sound perception score is automatically programmed (Based on instant specification paragraph [0002], the perception score indicates the human perception of sounds from jury tests (sound pleasantness). [Filimonov, 0049] discloses a training dataset comprising the first frequency response (vibrational information) and a reference scoring (sound perception score) [Filimonov, 0050] discloses training the neural network on the training dataset to predict a scoring for the audio system. [Filimonov, 0069 and 0072] discloses storing the score as a training dataset and predicting the score using the machine learning model (automatically programmed) )
in response to meeting a Based on the specification [0081], the convergence threshold is met by reducing a sound prediction error. [Filimonov, 0031, 0065 and 0074] collectively disclose weights in the artificial neural network may be determined to locally minimize a loss function indicative of a discrepancy between the reference scoring and the predicted scoring. [Filimonov, 0066-0068] collectively disclose testing the trained neural network and using the neural network to predict real data)
based on the trained machine learning model and the real-time . ([Filimonov, 0069 and 0075] collectively disclose providing the frequency response as a production dataset to the neural network and predicting the sound quality by giving a score on the scale used by the evaluators. [Filimonov, 0066-0068] collectively disclose testing the trained neural network and using the neural network to predict real data)
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to apply the method of receiving and generating sound perception data by Filimonov to improve the performance of the machine learning method of the present invention. The suggestion and/or motivation for doing so is to improve the quality of the produced device by allowing the user/factory/computer to automatically test and adjust settings of the device in response to the generated sound perception data [Filimonov, 0042].
However, Krishnaveni in view of Filimonov does not specifically disclose:
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model;
Mikhailov teaches:
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model; ([Mikhailov, 0079] If the error is below the threshold value (step 1040, NO), process 1000 may proceed to step 1010 of acquiring a next training audio signal (outputting machine learning model). If the error is above the threshold value (step 1040, YES), process 1000 may proceed to a step 1050 of modifying model parameters and subsequently returning to step 1020. In various embodiments, the machine learning model may be rated based on the average error generated by the model)
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to apply the method of receiving and generating training vibrational data by Mikhailov to improve the performance of the machine learning method of the present invention. The suggestion and/or motivation for doing so is to improve the accuracy of the prediction model by leveraging more diverse data related to the received sounds and by optimizing the performance of the model by adjusting parameters [Mikhailov, 0057].
However, Krishnaveni in view of Filimonov and further in view of Mikhailov does not specifically disclose:
based on the trained machine learning model, outputting characteristics associated with (data) from the device in an end of line (EOL) environment.
Gogoana teaches:
based on the trained machine learning model, outputting characteristics associated with (data) from the device in an end of line (EOL) environment. ([Gogoana, 0030 and 0136] collectively disclose collecting training data from two different environments including a lab data and a factory data (an EOL environment) to properly classify (characteristics) odors)
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to apply the method of utilizing two environmental data including lab data and factory data by Gogoana to improve the performance of the machine learning method of the present invention. The suggestion and/or motivation for doing so is to improve the accuracy of the prediction model by leveraging more diverse data related to the received sounds [Gogoana, 0030].
Regarding claim 2, Krishnaveni in view of Filimonov teaches:
training utilizing the vibrational information ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal is collected and stored in cloud, which is connected to the Arduino device, collected data (generated data in a lab environment) is used to train and test the model. [Krishnaveni, page 2553, right col, 2.4. MEMS accelerometer, line 1-14] discloses collecting X, Y, and Z axis data using the 3-axis accelerometer)
Krishnaveni does not specifically disclose:
Filimonov teaches:
The computer-implemented method of claim 1, wherein the trained machine learning model is trained utilizing only the [Filimonov, 0031, 0065 and 0074] collectively disclose weights in the artificial neural network may be determined to locally minimize a loss function indicative of a discrepancy between the reference scoring and the predicted scoring. [Filimonov, 0066-0068] collectively disclose testing the trained neural network and using the neural network to predict real data)
Regarding claim 5, Krishnaveni in view of Filimonov teaches:
The computer-implemented method of claim 1, wherein the sound perception scored is generated manually in response to the sound information. ([Filimonov, 0063] discloses that the reference scoring related to the reference audio system is given by human audio system expert evaluators for a subjective quality of the audio system)
Regarding claim 6, Krishnaveni teaches:
The computer-implemented method of claim 1, wherein the vibrational information is accelerometer data. ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal is collected and stored in cloud, which is connected to the Arduino device, collected data (generated data in a lab environment) is used to train and test the model. [Krishnaveni, page 2553, right col, 2.4. MEMS accelerometer, line 1-14] discloses collecting X, Y, and Z axis data using the 3-axis accelerometer)
Regarding claim 8, Krishnaveni teaches:
The computer-implemented method of claim 1, wherein the real-time [Krishnaveni, ABSTRACT, line 13-16] discloses generating real-time vibration measurements and analysis using the tool. [Krishnaveni, page 2554, left col, 4.1. Performance analysis using decision tree, line 1-8] discloses generating classification and classification accuracy score based on the trained DT algorithm. [Krishnaveni, page 2552, left col, 2. Proposed model, line 1-6 and line 11-17] discloses training a DT model to perform vibration analysis and diagnosis. The motor vibration is collected from motors. Additionally, [Krishnaveni, page 2554, left col, 4.2. Tested motor were divided into two groups, line 1 – right col, line 29] discloses training a neural network model to diagnose an induction motor)
However, Krishnaveni does not specifically disclose:
The computer-implemented method of claim 1, wherein the real-time sound perception score is generated utilizing only the real-time vibrational information.
Filimonov teaches:
The computer-implemented method of claim 1, wherein the real-time sound perception score is generated utilizing only the real-time [Filimonov, 0075] The scoring is determined for a ‘new’ audio system, which is different from the reference audio system disclosed in [Filimonov, 0071])
Regarding claim 9, Krishnaveni in view of Filimonov teaches:
The computer-implemented method of claim 1, wherein the machine learning model is a deep learning network. ([Filimonov, 0066-0068] collectively disclose testing the trained neural network and using the neural network to predict real data)
Regarding claim 10, Krishnaveni teaches:
A computer-implemented method, comprising: ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] The Arduino device is the computer)
receiving a first set of sound information and a first set of vibrational information from one or more accelerometers associated with an electrical drive of a device in a first environment; ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal (sound information associated with the device) is collected and stored in cloud, which is connected to the Arduino device, collected data is used to train and test the model. [Krishnaveni, page 2553, left col, 2.4. MEMS accelerometer, line 1-12] and [Krishnaveni, page 2553, right col, 4. Results and discussion, line 1-12] discloses utilizing MEMS accelerometer to collect vibrational data and to train the machine learning model)
generating a training data set utilizing at least the first set of vibrational information , ; ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal is collected and stored in cloud, which is connected to the Arduino device, collected data (generated data in a lab environment) is used to train and test the model. [Krishnaveni, page 2553, right col, 2.4. MEMS accelerometer, line 1-14] discloses collecting X, Y, and Z axis data using the 3-axis accelerometer)
… outputting a trained machine learning model; ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal is collected and stored in cloud, which is connected to the Arduino device, collected data (generated data in a lab environment) is used to train and test the model. Additionally, [Krishnaveni, page 2554, left col, 4.2. Tested motor were divided into two groups, line 1 – right col, line 29] discloses training a neural network model to diagnose an induction motor)
receiving real-time vibrational information from the electrical device associated with the device in a second environment; and ([Krishnaveni, ABSTRACT, line 13-16] discloses generating real-time vibration measurements and analysis using the tool. [Krishnaveni, page 2552, left col, line 14 – last line] discloses connecting sensors to the motor (electrical drive associated with the device) to collect motor signal (vibrational information). [Krishnaveni, page 2554, left col, line 5-9] and [Krishnaveni, page 2554, right col, Fig. 5] collectively discloses the system analyzing both normal condition and peak condition. The peak condition is the second environment)
based on the trained machine learning model and the real-time vibrational information, outputting . ([Krishnaveni, page 2554, left col, 4.1. Performance analysis using decision tree, line 1-8] discloses generating classification and classification accuracy score based on the trained DT algorithm. [Krishnaveni, page 2552, left col, 2. Proposed model, line 1-6 and line 11-17] discloses training a DT model to perform vibration analysis and diagnosis. The motor vibration is collected from motors. Additionally, [Krishnaveni, page 2554, left col, 4.2. Tested motor were divided into two groups, line 1 – right col, line 29] discloses training a neural network model to diagnose an induction motor)
However, Krishnaveni does not specifically disclose:
receiving a first set of sound information from one or more microphones and a first set of vibrational information;
generating a training data set utilizing at least the first set of vibrational information and an associated sound perception score, wherein the training data set is sent to an un-trained machine learning model, wherein the sound perception score is automatically programmed and the training data set includes accelerometer data that includes x-axis, y-axis, and z-axis information;
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model;
based on the trained machine learning model, outputting characteristics associated with (data) from the device in an end of line (EOL) environment.
Filinomov teaches:
receiving a first set of sound information from one or more microphones and a first set of [Filimonov, 0033] The sound measuring is performed using a sound recording device, such as a microphone. [Filimonov, 0049] discloses a training dataset comprising the first frequency response and a reference scoring (sound perception score))
in response to meeting a [Filimonov, 0031, 0065 and 0074] collectively disclose weights in the artificial neural network may be determined to locally minimize a loss function indicative of a discrepancy between the reference scoring and the predicted scoring. [Filimonov, 0066-0068] collectively disclose testing the trained neural network and using the neural network to predict real data)
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to apply the method of receiving and generating sound perception data by Filimonov to improve the performance of the machine learning method of the present invention. The suggestion and/or motivation for doing so is to improve the quality of the produced device by allowing the user/factory/computer to automatically test and adjust settings of the device in response to the generated sound perception data [Filimonov, 0042].
However, Krishnaveni in view of Filinomov does not specifically disclose:
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model;
based on the trained machine learning model, outputting characteristics associated with (data) from the device in an end of line (EOL) environment.
Mikhailov teaches:
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model; ([Mikhailov, 0079] If the error is below the threshold value (step 1040, NO), process 1000 may proceed to step 1010 of acquiring a next training audio signal (outputting machine learning model). If the error is above the threshold value (step 1040, YES), process 1000 may proceed to a step 1050 of modifying model parameters and subsequently returning to step 1020. In various embodiments, the machine learning model may be rated based on the average error generated by the model)
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to apply the method of receiving and generating training vibrational data by Mikhailov to improve the performance of the machine learning method of the present invention. The suggestion and/or motivation for doing so is to improve the accuracy of the prediction model by leveraging more diverse data related to the received sounds and by optimizing the performance of the model by adjusting parameters [Mikhailov, 0057].
However, Krishnaveni in view of Filimonov and further in view of Mikhailov does not specifically disclose:
based on the trained machine learning model, outputting characteristics associated with (data) from the device in an end of line (EOL) environment.
Gogoana teaches:
based on the trained machine learning model, outputting characteristics associated with (data) from the device in an end of line (EOL) environment. ([Gogoana, 0030 and 0136] collectively disclose collecting training data from two different environments including a lab data and a factory data (an EOL environment) to properly classify (characteristics) odors)
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to apply the method of utilizing two environmental data including lab data and factory data by Gogoana to improve the performance of the machine learning method of the present invention. The suggestion and/or motivation for doing so is to improve the accuracy of the prediction model by leveraging more diverse data related to the received sounds [Gogoana, 0030].
Regarding claim 11, Krishnaveni teaches:
The computer-implemented method of claim 10, wherein the vibrational data includes accelerometer data. ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal is collected and stored in cloud, which is connected to the Arduino device, collected data (generated data in a lab environment) is used to train and test the model. [Krishnaveni, page 2553, right col, 2.4. MEMS accelerometer, line 1-14] discloses collecting X, Y, and Z axis data using the 3-axis accelerometer)
Regarding claim 13,
Krishnaveni teaches:
wherein the real-time [Krishnaveni, ABSTRACT, line 13-16] discloses generating real-time vibration measurements and analysis using the tool. [Krishnaveni, page 2554, left col, 4.1. Performance analysis using decision tree, line 1-8] discloses generating classification and classification accuracy score based on the trained DT algorithm. [Krishnaveni, page 2552, left col, 2. Proposed model, line 1-6 and line 11-17] discloses training a DT model to perform vibration analysis and diagnosis. The motor vibration is collected from motors. Additionally, [Krishnaveni, page 2554, left col, 4.2. Tested motor were divided into two groups, line 1 – right col, line 29] discloses training a neural network model to diagnose an induction motor)
Krishnaveni does not specifically disclose:
The computer-implemented method of claim 10, wherein the real-time sound perception score is generated utilizing only the real-time vibrational data.
Filimonov teaches:
The computer-implemented method of claim 10, wherein the real-time sound perception score is generated utilizing only the real-time [Filimonov, 0069 and 0075] collectively disclose providing the production dataset for new audio system (second environment) to the neural network and predicting the sound quality by giving a score on the scale used by the evaluators. [Filimonov, 0066-0068] collectively disclose testing the trained neural network and using the neural network to predict real data)
Regarding claim 14, Krishnaveni in view of Filimonov teaches:
the computer-implemented method of claim 10, wherein the machine learning model is a deep learning network. ([Filimonov, 0066-0068] collectively disclose testing the trained neural network and using the neural network to predict real data)
Regarding claim 15, Krishnaveni in view of Filimonov teaches:
The computer-implemented method of claim 10, wherein the first environment is a laboratory environment and the second environment [Filimonov, 0028 and 0034] collectively disclose that the first environment is a reference audio system and the second environment as production audio system. The production audio system that is being tested by the system may be, for example, a prototype of a new audio system)
Krishnaveni in view of Filimonov in view of Mikhailov does not specifically disclose:
the second environment is an end-of-line factory environment.
Gogoana teaches the second environment is an end-of-line factory environment. ([Gogoana, 0030 and 0136] collectively disclose collecting training data from two different environments including a lab data and a factory data)
Regarding claim 16, Krishnaveni teaches:
A system, comprising: ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] The Arduino device is the computer)
a processor, wherein the processor is programmed to: ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] The Arduino device is the computer that contains a programmed processor)
receive sound information and vibrational information from one or more accelerometers associated with a device in a first environment; ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal (sound information associated with the device) is collected and stored in cloud, which is connected to the Arduino device, collected data is used to train and test the model. [Krishnaveni, page 2553, left col, 2.4. MEMS accelerometer, line 1-12] and [Krishnaveni, page 2553, right col, 4. Results and discussion, line 1-12] discloses utilizing MEMS accelerometer to collect vibrational data and to train the machine learning model)
generate a training data set utilizing at least the vibrational information , ; ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal is collected and stored in cloud, which is connected to the Arduino device, collected data (generated data in a lab environment) is used to train and test the model. [Krishnaveni, page 2553, right col, 2.4. MEMS accelerometer, line 1-14] discloses collecting X, Y, and Z axis data using the 3-axis accelerometer)
… outputting a trained machine learning model; ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal is collected and stored in cloud, which is connected to the Arduino device, collected data (generated data in a lab environment) is used to train and test the model. Additionally, [Krishnaveni, page 2554, left col, 4.2. Tested motor were divided into two groups, line 1 – right col, line 29] discloses training a neural network model to diagnose an induction motor)
receive real-time vibrational information from the device in a second environment; and ([Krishnaveni, ABSTRACT, line 13-16] discloses generating real-time vibration measurements and analysis using the tool. [Krishnaveni, page 2552, left col, line 14 – last line] discloses connecting sensors to the motor (electrical drive associated with the device) to collect motor signal (vibrational information). [Krishnaveni, page 2554, left col, line 5-9] and [Krishnaveni, page 2554, right col, Fig. 5] collectively discloses the system analyzing both normal condition and peak condition. The peak condition is the second environment)
based on the real-time vibrational information as an input to the trained machine learning model, output a real-time with sound emitted from the device . ([Krishnaveni, page 2554, left col, 4.1. Performance analysis using decision tree, line 1-8] discloses generating classification and classification accuracy score based on the trained DT algorithm. [Krishnaveni, page 2552, left col, 2. Proposed model, line 1-6 and line 11-17] discloses training a DT model to perform vibration analysis and diagnosis. The motor vibration is collected from motors. Additionally, [Krishnaveni, page 2554, left col, 4.2. Tested motor were divided into two groups, line 1 – right col, line 29] discloses training a neural network model to diagnose an induction motor)
However, Krishnaveni does not specifically disclose:
receiving sound information and vibrational data from one or more sensors associated with a device;
generate a training data set utilizing at least the vibrational information and a sound perception score associated with the corresponding sound of the vibrational information, wherein the training data set is sent to an un-trained machine learning model, wherein the sound perception score is automatically programmed and the training data set includes accelerometer data that includes x-axis, y-axis, and z-axis information;
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model;
based on the real-time vibrational information as an input to the trained machine learning model, output a real-time sound perception score indicating characteristics associated with sound emitted from the device in an end of line (EOL) environment.
Filimonov teaches:
receive sound information from one or more microphones and vibrational information [Filimonov, 0046-0047] discloses the frequency response detector measuring a first and a second frequency response of a production audio system, and the input unit configured to receive reference scoring indicative of a sound quality of the audio system. [Filimonov, 0071] discloses that the impulse response is measured by the sound recording device 310, e.g. a microphone)
generate a training data set utilizing at least the vibrational information and a sound perception score associated with the corresponding sound of the vibrational information, wherein the training data set is sent to an un-trained machine learning model, wherein the sound perception score is automatically programmed; (Based on instant specification paragraph [0002], the perception score indicates the human perception of sounds from jury tests (sound pleasantness). [Filimonov, 0049] discloses a training dataset comprising the first frequency response (vibrational information) and a reference scoring (sound perception score) [Filimonov, 0050] discloses training the neural network on the training dataset to predict a scoring for the audio system. [Filimonov, 0069 and 0072] discloses storing the score as a training dataset and predicting the score using the machine learning model (automatically programmed) )
in response to meeting a [Filimonov, 0031, 0065 and 0074] collectively disclose weights in the artificial neural network may be determined to locally minimize a loss function indicative of a discrepancy between the reference scoring and the predicted scoring. [Filimonov, 0066-0068] collectively disclose testing the trained neural network and using the neural network to predict real data)
based on the real-time vibrational information as an input to the trained machine learning model, output a real-time sound perception score indicating characteristics associated with sound emitted from the device . ([Filimonov, 0069 and 0075] collectively disclose providing the production dataset to the neural network and predicting the sound quality by giving a score on the scale used by the evaluators. [Filimonov, 0066-0068] collectively disclose testing the trained neural network and using the neural network to predict real data)
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to apply the method of receiving and generating sound perception data by Filimonov to improve the performance of the machine learning method of the present invention. The suggestion and/or motivation for doing so is to improve the quality of the produced device by allowing the user/factory/computer to automatically test and adjust settings of the device in response to the generated sound perception data [Filimonov, 0042].
However, Krishnaveni in view of Filimonov does not specifically disclose:
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model;
based on the trained machine learning model, outputting characteristics associated with (data) from the device in an end of line (EOL) environment.
Mikhailov teaches:
in response to meeting a convergence threshold of the un-trained machine learning model, outputting a trained machine learning model; ([Mikhailov, 0079] If the error is below the threshold value (step 1040, NO), process 1000 may proceed to step 1010 of acquiring a next training audio signal (outputting machine learning model). If the error is above the threshold value (step 1040, YES), process 1000 may proceed to a step 1050 of modifying model parameters and subsequently returning to step 1020. In various embodiments, the machine learning model may be rated based on the average error generated by the model)
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to apply the method of receiving and generating training vibrational data by Mikhailov to improve the performance of the machine learning method of the present invention. The suggestion and/or motivation for doing so is to improve the accuracy of the prediction model by leveraging more diverse data related to the received sounds and by optimizing the performance of the model by adjusting parameters [Mikhailov, 0057].
However, Krishnaveni in view of Filimonov and further in view of Mikhailov does not specifically disclose:
based on the trained machine learning model, outputting characteristics associated with (data) from the device in an end of line (EOL) environment.
Gogoana teaches:
based on the trained machine learning model, outputting characteristics associated with (data) from the device in an end of line (EOL) environment. ([Gogoana, 0030 and 0136] collectively disclose collecting training data from two different environments including a lab data and a factory data (an EOL environment) to properly classify (characteristics) odors)
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to apply the method of utilizing two environmental data including lab data and factory data by Gogoana to improve the performance of the machine learning method of the present invention. The suggestion and/or motivation for doing so is to improve the accuracy of the prediction model by leveraging more diverse data related to the received sounds [Gogoana, 0030].
Regarding claim 17, Krishnaveni teaches:
The system of claim 16, wherein the vibrational information includes three-dimensional information. ([Krishnaveni, page 2552, left col, 2. Proposed model, line 1-17] discloses that in training of machine learning model, motor signal is collected and stored in cloud, which is connected to the Arduino device, collected data (generated data in a lab environment) is used to train and test the model. [Krishnaveni, page 2553, right col, 2.4. MEMS accelerometer, line 1-14] discloses collecting X, Y, and Z axis data using the 3-axis accelerometer)
Regarding claim 18, Filimonov teaches:
the system of claim 16, wherein the processor is further programmed to generate the training data set utilizing both the [Filimonov, Claim 1 and Claim 12] is a computer-implemented method which includes at least one processor. [Filimonov, 0046-0047] discloses the frequency response detector measuring a first and a second frequency response of a production audio system, and the input unit configured to receive reference scoring indicative of a sound quality of the audio system)
Filimonov does not specifically disclose generate the training data set utilizing both the vibrational information.
Mikhailov teaches:
generate the training data set utilizing both the vibrational information. ([Mikhailov, 0073] discloses data related to vibrations in devices may be transmitted to a computing device associated with VA 130 (e.g., computing device 171), and this data may be correlated with audio signal 120. The vibration data may include various frequencies that may be correlated with various words that may be pronounced by the lead speaker)
Regarding claim 20, Krishnaveni in view of Filimonov teaches:
the system of claim 16, wherein the first environment and the second environment are not a same environment. ([Filimonov, 0075] The scoring is determined for a ‘new’ audio system, which is different from the reference audio system disclosed in [Filimonov, 0071])
Claims 3, 7, 12 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Krishnaveni in view of Filimonov in view of Mikhailov in view of Gogoana in view of Effinger et al (US 10841666 B1, hereinafter ‘Effinger’) and further in view of Mansour & Seifert (US 20190278990 A1, hereinafter ‘Mansour’).
Regarding claim 3, Krishnaveni in view of Filimonov teaches:
The computer-implemented method of claim 1, wherein the trained machine learning model is trained, wherein a first neural network of the machine learning model is trained utilizing the sound information; ([Filimonov, 0031, 0065 and 0074] collectively disclose weights in the artificial neural network may be determined to locally minimize a loss function indicative of a discrepancy between the reference scoring and the predicted scoring. [Filimonov, 0066-0068] collectively disclose testing the trained neural network and using the neural network to predict real data)
feeding the [Filimonov, 0069 and 0075] collectively disclose providing the production dataset to the neural network and predicting the sound quality by giving a score on the scale used by the evaluators. [Filimonov, 0066-0068] collectively disclose testing the trained neural network and using the neural network to predict real data)
freezing weights associated with the score prediction network and training the weights of the [Filimonov, 0031, 0065 and 0074] collectively disclose weights in the artificial neural network may be determined to locally minimize a loss function indicative of a discrepancy between the reference scoring and the predicted scoring using mean squared error (average of many error values). [Filimonov, 0069 and 0075] collectively disclose providing the production dataset to the neural network and predicting the sound quality by giving a score on the scale used by the evaluators. The weights of the neural network are frozen as it will not be trained again. [Filimonov, 0066-0068] collectively disclose testing the trained neural network and using the neural network to predict real data)
However, Krishnaveni in view of Filimonov in view of Mikhailov and further in view of Gogoana does not specifically disclose:
wherein the trained machine learning model is trained via an in-direct method, wherein a first neural network of the machine learning model is trained utilizing the sound information and a second neural network is trained to predict measured sound utilizing the vibrational information and obtain a predicted sound;
feeding the predicted sound into a score prediction network to generate a human-perception score; and
freezing weights associated with the score prediction network and training the weights of the sound prediction network to minimize a weighted sum of sound and score prediction errors.
Effinger teaches:
wherein the trained machine learning model is trained via an in-direct method, wherein a first neural network of the machine learning model is trained utilizing the sound information and a second neural network is trained to predict measured sound utilizing the vibrational information and obtain a predicted sound; ([Effinger, col 5, line 59 – col 6, line 3] discloses processing features from the feature generator unit 122 using the scoring subsystem 120. [Effinger, col 6, line 61-67; Fig. 1] discloses applying the pre-trained aural transformer model included in the feature generator unit 122 to generate aural features (second neural network). [Effinger, col 7, line 41-62; col 8, line 2-16; Fig. 1] discloses the scoring subsystem 120 including a grading unit 124 that can determine scores 125 for respective candidate content using aural features associated with the content insertion points. The grading unit 124 can apply a machine-learning model 125 (first neural network) to the received aural features such as a CNN. [Effinger, col 8, line 23-26] discloses training the scoring model on audio data)
feeding the predicted sound into a score prediction network to generate a human-perception score; and ([Effinger, col 5, line 59 – col 6, line 3] discloses processing features from the feature generator unit 122 using the scoring subsystem 120. [Effinger, col 6, line 61-65; Fig. 1] discloses applying the aural transformer model included in the feature generator unit 122 to generate aural features (second neural network) )
freezing weights associated with the score prediction network [Effinger, col 7, line 41-62; col 8, line 2-16; Fig. 1] discloses the scoring subsystem 120 including a grading unit 124 that can determine scores 125 for respective candidate content using aural features associated with the content insertion points. The grading unit 124 can apply a machine-learning model 125 (first neural network) to the received aural features such as a CNN. [Effinger, col 8, line 23-26] discloses training the scoring model on audio data. Since the scoring model will not be trained again, the model is frozen)
It would have been obvious before the effective filing date of the claimed invention to a person
having ordinary skill in the art to apply the method of utilizing both a sound prediction neural network and a score prediction neural network (indirect prediction) by Effinger to improve the performance of the machine learning method of the present invention. The suggestion and/or motivation for doing so is to improve the accuracy of the prediction model by generating a prediction based on different variables.
However, Filimonov in view of Mikhailov and further in view of Effinger does not specifically disclose:
freezing weights associated with the
Mansour teaches:
freezing weights associated with the [Mansour, 0033] discloses training the first neural network while freezing the weights of the second neural network. [Mansour, 0034] discloses that the training is performed based on a model loss total L.sub.Total is a weighted sum of the individual loss functions L.sub.1, L.sub.2, L.sub.n. The objective is to minimize the weighted sum of the individual loss functions)
It would have been obvious before the effective filing date of the claimed invention to a person
having ordinary skill in the art to apply the method of freezing weights of a first prediction network while training a second prediction network by Mansour to improve the performance of the machine learning method of the present invention. The suggestion and/or motivation for doing so is to improve the efficiency of the machine learning model by not re-training pre-trained neural networks with sufficient performance.
Regarding claim 7, Krishnaveni in view of Filimonov in view of Mikhailov in view of Gogoana and further in view of Effinger teaches:
the computer-implemented method of claim 1, wherein the machine learning model is a U-Net or Transformer network. ([Effinger, col 6, line 61-65; Fig. 1] discloses applying the aural transformer model included in the feature generator unit 122 to generate aural features (second neural network) )
Regarding claim 12, Krishnaveni in view of Filimonov in view of Mikhailov in view of Gogoana and further in view of Effinger teaches:
the computer-implemented method of claim 10, wherein the machine learning model is a U-Net or Transformer network. ([Effinger, col 6, line 61-65; Fig. 1] discloses applying the aural transformer model included in the feature generator unit 122 to generate aural features (second neural network) )
Regarding claim 19, Krishnaveni in view of Filimonov in view of Mikhailov in view of Gogoana and further in view of Effinger teaches:
the system of claim 16, wherein machine learning model includes two or more neural networks utilized to output a real-time sound perception score. ([Effinger, col 5, line 59 – col 6, line 3] discloses processing features from the feature generator unit 122 using the scoring subsystem 120. [Effinger, col 6, line 61-65; Fig. 1] discloses applying the aural transformer model included in the feature generator unit 122 to generate aural features (second neural network). [Effinger, col 7, line 41-62; col 8, line 2-16; Fig. 1] discloses the scoring subsystem 120 including a grading unit 124 that can determine scores 125 for respective candidate content using aural features associated with the content insertion points. The grading unit 124 can apply a machine-learning model 125 (first neural network) to the received aural features such as a CNN)
Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Krishnaveni in view of Filimonov in view of Mikhailov in view of Gogoana and further in view of Grigoryan et al (Grigoryan et al, “Subject Filtering for Passive Biometric Monitoring”, 2004, hereinafter ‘Grigoryan’)
Regarding claim 4, Krishnaveni in view of Filimonov in view of Mikhailov and further in view of Gogoana teaches:
The computer-implemented method of claim 1.
However, Krishnaveni in view of Filimonov in view of Mikhailov and further in view of Gogoana does not specifically disclose:
wherein the training data utilizes sound information and accelerometer data obtained from a noise-free environment.
Grigoryan teaches:
wherein the training data utilizes sound information and accelerometer data obtained from a noise-free environment. ([Grigoryan, page 488, Gaussian Mixture Model, line 12-13] discloses using the training data for the target speaker to estimate the mean vectors, weights, and component densities for his/her model. [Grigoryan, page 486, line 18-20] discloses using two microphones and an accelerometer to collect sound data. [Grigoryan, page 491, line 3-4] The speech was recorded in a relatively noise-free environment)
It would have been obvious before the effective filing date of the claimed invention to a person
having ordinary skill in the art to apply the method of obtaining training data using an accelerometer from a noise-free environment by Grigoryan to improve the performance of the machine learning method of the present invention. The suggestion and/or motivation for doing so is to improve the accuracy of the machine learning model by using clean training data collected from multiple sources.
Double Patenting
Amended claims were received on 10/20/2025. The double patenting rejection has been withdrawn.
Response to Arguments
Response to Arguments under 35 U.S.C. 112
Applicant’s arguments regarding the rejection under 35 U.S.C. 112(b) are persuasive. The rejection has been withdrawn.
Response to Arguments under 35 U.S.C. 101
Arguments: Applicant asserts (a) that the claimed invention improves the functioning of evaluating a machine learning model utilizing a specific type of technical process (utilizing a shallow repressor model and residuals to generate a modified effect of an optimized model), a specific technological area, and (b) that USPTO Example 37 and 42 (Training a Neural Network for Facial Detection), and USPTO Director Squires’ statement support Applicant’s argument. [Remarks, page 6-7]
Examiner’s Response: Examiner respectfully disagrees. Regarding (a), MPEP 2106.05(a) indicates that, when considering whether additional elements integrate a judicial exception into a practical application, the examiner must evaluate if the disclosure provides sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement. If the specification explicitly sets forth an improvement but in a conclusory manner (i.e., a bare assertion of an improvement without the detail necessary to be apparent to a person of ordinary skill in the art), the examiner should not determine the claim improves technology. Examiner performed this evaluation and concluded that the specification [0022]-[0023] set forth the improvement but in a conclusory manner. The specification merely discloses the conclusion that the sound prediction model and EOL testing setups with different accelerometer positioning improves the technology without sufficient details necessary to be apparent to a person of ordinary skill in the art.
Regarding (b), Examiner would first note that, the USPTO Example 37 and USPTO Director Squires’ statement cited by Applicant are nonprecedential and do not bind Examiner in the instant case.
Second, the USPTO Example 37 and 42 (Training a Neural Network for Facial Detection) is distinguishable from the instant application. The USPTO Example is eligible because the claim does not recite any mathematical relationships, formulas, or calculations. While some of the limitations may be based on mathematical concepts, the mathematical concepts are not recited in the claims. Further, the claim does not recite a mental process because the steps are not practically performed in the human mind. However, the instant application recites mental processes of evaluation. The limitations of “generating a training data set utilizing at least the vibrational information and a sound perception score associated with the vibrational information in a lab environment” is directed to a mental process of judgment as selecting and combining information and score can practically be performed in the human mind. The limitations of “based on the real-time vibrational information, outputting a real-time sound perception score indicating characteristics associated with sound emitted from the device in an end of line (EOL) environment” is directed to a mental process of evaluation as the limitation recites scoring process which can be performed with the aid of pen and paper. The limitations of ‘receiving sound information from one or more microphones and vibrational data from one or more accelerometers …’ are directed to an insignificant extra-solution activity MPEP 2106.05(g) in Step 2A Prong 2 analysis and well understood, routine and conventional activity MPEP 2106.05(d) in Step 2B, and ‘outputting a trained machine learning model’ is directed to mere instructions to perform the abstract idea of selecting and combining information and scoring the data records using a generic computer component MPEP 2106.05(f).
Lastly, Desjardins cited in Squires’ statement is distinguishable from the instant application. As an initial matter, it is worth noting that Desjardins does not stand for the proposition that all claims directed to methods of training neural networks are per se eligible. Furthermore, the claims at issue in Desjardins explicitly linked the steps recited to an improvement in training by reciting “optimizing performance of the machine learning model on the second machine learning task while protecting performance of the machine learning model on the first machine learning task.” However, there is no such nexus in the instant claims.
Accordingly, the arguments regarding claim 1 are not persuasive. Similarly, the arguments regarding claims 10 and 16 are not persuasive. Therefore, the arguments to claims 2-9, 11-15, and 17-20 depend from independent claims 1, 10 and 16 are not persuasive.
Response to Arguments under 35 U.S.C. 103
Argument: Applicant asserts that neither Filimonov nor Mikhailov teaches the vibrational information obtained from one or more accelerometers. [Remarks, page 8]
Examiner’s response: Applicant’s arguments with respect to claim 1 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Argument: Applicant asserts that neither Filimonov nor Mikhailov teaches “in response to meeting a convergence threshold of the untrained machine learning model”. [Remarks, page 8]
Examiner’s response: Examiner respectfully disagrees. Even though Filimonov does not clearly specify that the trained machine learning model performance has to be higher than a specific threshold, [Filimonov, 0066] clearly indicates that the trained artificial neural network is deployed and used to predict a sound quality of a new or modified prototype of a vehicle audio system. Mikhailov is introduced to teach the thresholding operation that completes the training process after the threshold is met.
Accordingly, the arguments regarding claim 1 are not persuasive. Similarly, the arguments regarding claims 10 and 16 are not persuasive. Therefore, the arguments to claims 2-9, 11-15, and 17-20 depend from independent claims 1, 10 and 16 are not persuasive.
Double Patenting
Amended claims were received on 10/20/2025. The double patenting rejection has been withdrawn.
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 JUN KWON whose telephone number is (571)272-2072. The examiner can normally be reached Monday – Friday 7:30AM – 4:30PM.
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, Abdullah Kawsar can be reached at (571)270-3169. 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.
/JUN KWON/Examiner, Art Unit 2127
/RYAN C VAUGHN/Primary Examiner, Art Unit 2125