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 .
DETAILED ACTION
Status of Claims
Claims 1-15 are rejected under35 USC §103 Rejection.
Remarks
Applicant’s arguments filed (03/11/2026) with respect to pending claims 1-15 have been fully considered and are directed to claims as amended. The arguments addressed to the 101 is persuasive.
Based on the changes introduced by amendment of (03/11/2026) the 101 Rejections of Claims 1-15 is withdrawn. See below the claim analysis– 35 USC § 101.
Regarding the 103 Rejection, Applicant's arguments with respect to the amendments to pending claims 1 and 7 have been considered but are moot because a new grounds of rejection is made with new prior art including Kim et al., (KR 20130129678A), hereinafter Kim and Li et al.,( CN113487145A), hereinafter Li, addressing the amended limitations in question. See below rejection for full detail.
Claim Analysis – 35 USC § 101
The new 2019 Revised Patent Subject Matter Eligibility Guidance published in the Federal register (Vol. 84 No.4, Jan 7, 2019 pp 50-57) has been applied and the claims are deemed as being patent eligible. In particular, in the prong 1 analysis claims 1 and 7 contain an abstract idea. However, they claim the practical application of improvement of arc risk management under prong 2 analysis. Thus claims 1 and 7 are deemed patent eligible under 35 USC 101.
Claims 1 and 7 comprise the additional elements/steps of obtaining from a current sensor… and shutting off a supply of electric power to the electric apparatus in response to the quantitative arc risk index satisfying a predetermined reference value, which is a real-world action and a specific sensor for measuring the data. Therefore, taking the claim as a whole, the claims is directed to a particular practical application of improvement arc risk management.
Claims 2-6 and 8-15 are dependent claims of claims 1 and 7 and they are directed to the practical application of the parent claim and are also patent eligible under 35 USC 101. Dependent claims 2-6 and 8-15 are eligible under 101.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1, 5, 7, and 8 are rejected under 35 U.S.C. 103 as being unpatentable over
Yong (KR10-20210103785A), [hereinafter Yong] and Geun (KR10-20220021973), [hereinafter Geun] and Kim et al., (KR 20130129678 A), hereinafter Kim and Li et al.,( CN113487145A), hereinafter Li.
Regarding Claim 1, Yong disclose an arc risk management method comprising:
obtaining, from a current sensor configured to measure currents flowing into an electric apparatus (Fig. 1, where first electric device 120 connect by conductor 120 with second electrical device 140, the current flow through electrical conductor into electrical apparatus), measurement values of the currents (page 3, lines 27-32, sensing unit 112 may obtain a measurement value of the current i flowing through the conductive wire 120 . The sensing unit 112 may generate time series data by measuring the current i flowing through the conductive wire 120 using a current sensor.The sensing unit 112 may periodically measure the current i, and store the measured value as time series data in a memory according to the order of time);
pre-processing the measurement values
indicating an arc risk of the electric apparatus in a quantitative way according to the level of arc energy (The ARC detection method comparing the first statistics and criteria range and increases the case where the first statistics deviates from the criteria range, and the indicator about the acr generation possibility or utilizes the first statistics as the mode setting the flag related to the arc (the short circuit page 5, para 8, Claim 1 and Fig. 6), e.g., criteria range of the acr generation corresponds to the level of arc energy).
Yong does not disclose estimating a level of arc energy in the electric apparatus by inputting the [pre-processed] measurement values into a single artificial intelligence network comprising a first layer including a dilated convolutional neural network and a second layer including a recurrent neural network;
generating a quantitative arc risk index and outputting the quantitative arc risk index for real-time display; and
shutting off a supply of electric power to the electric apparatus in response to the
quantitative arc risk index satisfying a predetermined reference value.
Geun discloses estimating a level of arc energy in the electric apparatus (invention can evaluate the possibility that the arc was generated in the first statistics in which the arc detection apparatus was calculated and the conducting wire which compared the criteria range (reference range) and in which the first statistics deviated from the criteria range the possibility is written in (the short page 5, para 3, and 6-8, Fig. 6)
by inputting the measurement values into a single artificial intelligence network comprising a first layer including a dilated convolutional neural network and a second layer including a recurrent neural network (the convolution neural network (dilated CNN) which is expanded it analyzes the correlation between time series data in which the bad battery estimator (630) is long is used as the basic structure and the front part learns using the recurrent neural network (Recurrent neural network) bay in which it repetitively accumulates the dilated CNN and arrangement normalization (Batch Normalization) technology and it can configure the whole artificial neural network the connection between (the short circuit page 8, para 2, and page 8, para 9, Fig.1), and the multilayer perceptron (Multi-Layer-Perceptron) building several perceptron neurons with a plurality of layers or 1D or 2D CNN or the unit have the recursive structure to CNN and the late learns to RNN and the artificial neural network is configured it is written as (the short circuit page 8, para 9, and Fig.1).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to input the measurement values into one artificial intelligence network comprising a first layer including a dilated convolutional neural network and a second layer including a recurrent neural network as taught by Geun using pre-processed measurement values of Yong in order to more accurately detect the location data for the particular area/region.
Kim discloses generating a quantitative arc risk index and outputting the quantitative arc risk index for real-time display (Page 5, lines 30-36, where display device 120 installed outside the enclosure of the switchboard 100 displays the risk level, the amount of arc-flash energy (accident energy) and further see page 5, line 35, where last saved risk class (minimum distance risk class) is displayed first; Page 6, lines 1-6, where risk level display control signal from the controller 150 in real time whenever a risk level is calculated, and displays the received risk level display control signal through a warning lamp. Here, the warning lamp means a warning lamp that lights in a color suitable for the risk class. For example, if the risk level is the maximum level (4th level based on NFPA 70E), it is lit in red color, if it is 3rd, it is lit in scarlet color, if it is 2nd, it is lit in pink color, if it is 1st, Lights up in blue color, and in case of 0 grade, lights up in green color, e.g., the risk level corresponds to the quantitative arc risk and quantitative arc index corresponds to the a color suitable for the risk class, maximum level of the risk correspond to the maximum level (red color)); and
[shutting off a supply of electric power to the electric apparatus]/ (stopping work) in response to the quantitative arc risk index (Page 5, lines 30-34, where display device 120 installed outside the enclosure of the switchboard 100 displays the risk level, the amount of arc-flash energy (accident energy) and the corresponding personal safety protective device system information, based on the displayed risk level related information. Do it. As a result, the operator can check the risk level in real time through the display device 120, change the personal safety protective device according to the risk level, wear or stop the live work);(page 4, para 10, intended to ensure safety by changing the system or recognizing the danger, stopping work and evacuating to a safe distance from the switchboard).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to shut off supply of electric power in order to protect the system and workers from high hazard use of device and protect users from harmful situations.
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to generating a quantitative arc risk index and outputting the quantitative arc risk index for real-time display as taught by Kim into Yong in order to visually and faster analyze the level of the hazard situation of the apparatus.
Li discloses quantitative arc risk index satisfying a predetermined reference value (Claim 3, where when the hidden danger level exceeds the preset level, sending the hidden trouble information to the safety management department to verify and confirm).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to determine when to shut off supply of electric power of Kim in response to satisfying a predetermined reference value as taught by Li in order to more accurately detect the hazard situation of electronic apparatus.
Regarding Claim 5, Yong and Geun and Kim and Li disclose the arc risk management method of claim 1, but Yong and Kim and Li do not disclose wherein the recurrent neural network comprises a transformer layer.
Geun discloses the recurrent neural network comprises a transformer layer
(Multi-Layer-Perceptron or 1D or 2D CNN with multiple perceptron neurons stacked in multiple layers It is okay to construct an artificial neural network by making a decoder structure and learning the speed with a Transformer architecture, or learning the first half with CNN and the second half with RNN. However, as in the present invention, it is time in the battery failure prediction field to which the present invention is applied to construct a neural network using a dilated Convolution Neural Network (CNN) in the first half and the middle part and a 1D-CNN (Convolution Neural Network) in the second half as in the present invention. It can be effective in several aspects, such as efficiency, Page 8, para 9, Page 9, para 1, and Fig.1).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide recurrent neural network comprises a transformer layer as taught by Geun in combination of Yong and Kim and Li in order to perform long-distance relationships efficiently and speeds up training and improves accuracy.
Regarding Claim 7, Yong and Geun and Kim and Li do disclose arc risk management system comprising:
a current sensor configured to measure currents flowing into an electric
apparatus,
a pre-processing module to pre-process the measurement values to generate pre-processed measurement values;
an artificial intelligence network module, comprising a single artificial intelligence neatwork having a first layer including a dilated convolutional neural network and a second layer including a recurrent neural network,
generating a quantitative arc risk index and outputting the quantitative arc risk index for real-time display, and to shut off a supply of electric power to the electric apparatus in response to the quantitative arc risk index satisfying a predetermined reference value, as recited in claim 1.
Further Yong discloses an arc risk management module to indicate an arc risk of the electric apparatus in a quantitative way according to the level of arc energy (The ARC detection method comparing the first statistics and criteria range and increases the case where the first statistics deviates from the criteria range, and the indicator about the arc generation possibility or utilizes the first statistics as the mode setting the flag related to the arc (the short circuit , see Page 5, para 8, Claim 1 and Fig. 6)).
to receive the pre-processed measurement values as an input ( Fig. 1, Page 3, lines 32-34, where the sensing unit 112 may generate time series data by storing the measured value of the current i as it is, or may generate time series data by filtering the measured value or scaling the measured value).
Geun discloses output a level of arc energy of the electric apparatus (Fig. 5, … an input unit 516 , an output unit 518 ,… As the output unit 518 , for example, a display device such as a cathode ray tube (CRT), a liquid crystal display (LCD), also see for “a level of arc energy”: invention can evaluate the possibility that the arc was generated in the first statistics in which the arc detection apparatus was calculated and the conducting wire which compared the criteria range(reference range) and in which the first statistics deviated from the criteria range the possibility is written in (the short page 5, para 3, and 6-8, fig. 6).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to input and output a level of arc energy of the electric apparatus, as taught by Geun in combination of Yong and Kim and Li in order to more to visually and faster analyze the level of the hazard situation of the apparatus.
Regarding Claim 8, Yong and Geun and Kim and Li disclose the arc risk management system of claim 7, further Young disclose wherein the pre-processing module normalizes the measurement values without causing dispersion information to vanish from the measurement values (Abstract, where calculating, in the time series data, a first statistical value indicating the degree of dispersion of the measurement value or the degree of dispersion of the variation of the measurement value over time).
Claims 2 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over
Yong in view of Geun, Kim and Li, as applied above and further in view of Xi-nan (CN201322957Y), [hereinafter Xi-nan].
Regarding Claim 2, Yong and Geun and Kim and Li disclose the arc risk management method of claim 1, Young does not discloses wherein in indicating an arc risk in a quantitative way, the arc risk is indicated in a form of a dial gauge in real time.
Geun disclose indicating an arc risk in a quantitative way, the arc risk is indicated (invention can evaluate the possibility that the arc was generated in the first statistics in which the arc detection apparatus was calculated and the conducting wire which compared the criteria range(reference range) and in which the first statistics deviated from the criteria range the possibility is written in (the short page 5, para 3, and 6-8, fig. 6) and (As the output unit 518, for example, a display device such as a cathode ray tube (CRT), a liquid crystal display (LCD), a plasma display panel (PDP), or an electro-luminescence display (ELD) may be used), e.g., the Geun discloses the output by the different type of display devises, but does not discloses dial gauge.
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide indicating an arc risk in a quantitative way, as taught by Geun in combination of Yong and Kim and Li in order to more accurately detect the potential risk energy status of the electric device.
Xi-nan disclose dial gauge in real time (the dial gauge, electric energy meter head by the high-level dial (2) and the low-level dial (4), see Fig. 1, gauge 18, Page 2, para 5) and (Page 3, para 3, where real-time clock (23), see page 2, para 5; and further …recording electric meter running real time information, see page 3, para 3).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide dial gauge in real time, as taught by Xi-nan into output unit of Geun and further into Yong and Kim and Li in order to more accurately detect the location data for the particular area/region.
Regarding Claim 14, Yong and Geun and Kim and Li disclose the arc risk management system of claim 7, further Young does not disclose comprising a display module to indicate quantitative indexes of the arc risk in a form of a gauge.
Geun discloses comprising a display module (the short page 5, para 3, and 6-8, fig. 6) and (As the output unit 518, for example, a display device such as a cathode ray tube (CRT), a liquid crystal display (LCD), a plasma display panel (PDP), or an electro-luminescence display (ELD) may be used) to indicate quantitative indexes of the arc risk (The ARC detection method comparing the first statistics and criteria range and increases the case where the first statistics deviates from the criteria range, and the indicator about the arc generation possibility or utilizes the first statistics as the mode setting the flag related to the arc (the short circuit Page 5, para 8, Claim 1 and fig. 6)).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide display, as taught by Xi-nan into output unit of Geun and further into Yong and Kim and Li in order to more accurately detect the location data for the particular area/region.
Yong and Geun and Kim and Li do not disclose gauge.
Xi-nan disclose dial gauge in real time (the dial gauge, electric energy meter head by the high-level dial (2) and the low-level dial (4)), see Fig. 1(displaying in form of gauge 18), Page 2, para 5 and (Page 3, para 3, where real-time clock(23), see page 2, para 5; and further…recording electric meter running real time information, see Page 3, para 3).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide display in a form of a gauge, information as taught by Xi-nan in combination of Yong and Geun and Kim and Li in order to more accurately detect the location data for the particular area/region.
Claims 3, 4, 6, 10, 11, 13 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Yong and Geun and Kim and Li, as applied above and further in view of Wook (KR 10-20220019376), [hereinafter Wook].
Regarding Claim 3, Yong and Geun and Kim and Li disclose the arc risk management method of claim 1, but do not disclose wherein the first layer is a layer transferred from another artificial intelligence network, which has conducted learning for classification of first factor values to influence the arc energy.
Wook disclose the first layer is a layer transferred from another artificial intelligence network, which has conducted learning for classification of first factor values to influence the arc energy (The learning unit 230 includes a spatial feature extracting unit 231, a temporal feature extracting unit 232 , and a predictive model generating unit 233 . An artificial intelligence model for power consumption prediction by receiving a training dataset output from the data preprocessor 220 and extracting spatial and temporal features from the training dataset (or “power consumption prediction AI model”) is generated and provided to the power consumption prediction unit 250, see page 4, para 12); (the spatial feature extraction unit 231 applies a convolutional neural network (CNN), extracts spatial features from the refined training dataset input from the data preprocessor 220, and extracts the spatial features from the temporal feature extraction unit 232) is output., see page 5, para 1, e.g., extract the space distinctive from the refined learning data set, which is corresponds to the factor values).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide the first layer, as taught by Wook in the combination of Geun and Yong and Kim and Li in order to enhance safety protocols and precise hazard assessment.
Regarding Claim 4, Yong and Geun and Kim and Li disclose the arc risk management method of claim 1, but do not disclose wherein the recurrent neural network comprises a long short-term memory (LSTM), which is advantageous for time series data analysis.
Wook disclose the recurrent neural network comprises a long short-term memory (LSTM), which is advantageous for time series data analysis (a power consumption prediction system combines a convolutional neural network (CNN) and a long short term memory (LSTM) model for power consumption prediction, and spatial feature extraction and LSTM through CNN A more accurate prediction is shown by performing modeling that reflects temporal information, see page 3, para 3).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide a long short-term memory (LSTM), as taught by Wook into Geun recurrent neural network and further into Yong and Kim and Li in order to more accurately model long-term dependencies in time-series data.
Regarding Claim 6, Yong and Geun and Wook and Kim and Li disclose the arc risk management method of claim 3, wherein the other artificial intelligence network distinguishes a normal state and an arc state of the electrical apparatus through the classification.
Young discloses the other artificial intelligence network distinguishes a normal state and an arc state of the electrical apparatus (arc detection device compares the first statistical value with the reference range (S604), and when the first statistical value is out of the reference range… the arc detection device may determine whether an arc is generated by comparing the first statistical value with the reference range, see Page 5, para 8).
Young and Wook and Kim and Li do not disclose artificial intelligence network distinguishes a normal state and arc state through the classification.
Geun disclose artificial intelligence network distinguishes a normal state and arc state through the classification (determination S240, and the predicted voltage and temperature data with the actual voltage and temperature data, respectively, and determines whether a difference between the predicted voltage and temperature data and the actual voltage and temperature data is within a predetermined range (S230)).
when the difference between the predicted voltage and temperature data and the actual voltage and temperature data is within a predetermined range, the battery cell failure diagnosis apparatus determines the battery determined as defective again as normal ( S240) when the difference between the predicted voltage and temperature data and the actual voltage and temperature data is out of a predetermined range, the battery cell failure diagnosis apparatus finally determines the battery determined to be defective as defective (S23), (the short circuit page, Fig. 2 and 3, Abstract, page 4, para 10-12) and the battery failure prediction unit 630 inputs the data stored in the storage unit 610 to the neural network, clusters the data collected through the neural network, and predicts a battery deviating from the main cluster as defective. Specifically, the battery failure prediction unit 630 infers the voltage and temperature data the case are written in (the short circuit page 7, para 7 and Fig.1).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide a artificial intelligence network distinguishes a state of the electrical apparatus through the classification, as taught by Geun in combination of Yong and Wook and Kim and Li in order to provide more safety, improve operational efficiency and reliability.
Regarding Claim 10, Yong and Geun and Kim and Li disclose the arc risk management system of claim 7, but do not disclose wherein the first layer is a layer transferred from another artificial intelligence network, which has conducted learning for classification of first factor values to influence the arc energy.
Wook disclose the first layer is a layer transferred from another artificial intelligence network, which has conducted learning for classification of first factor values to influence the arc energy (The learning unit 230 includes a spatial feature extracting unit 231, a temporal feature extracting unit 232 , and a predictive model generating unit 233 . An artificial intelligence model for power consumption prediction by receiving a training dataset output from the data preprocessor 220 and extracting spatial and temporal features from the training dataset (or “power consumption prediction AI model”) is generated and provided to the power consumption prediction unit 250, see page 4, para 12), (The spatial feature extraction unit 231 applies a convolutional neural network (CNN), extracts spatial features from the refined training dataset input from the data preprocessor 220, and extracts the spatial features from the temporal feature extraction unit 232) is output., see page 5, para 1, e.g., in the context of spatial analysis, factor values (or factor levels) are specific numerical or categorical data points that represent a particular characteristic or condition of a geographical location or spatial unit.).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide the first layer, as taught by Wook in the combination of Geun and Yong and Kim and Li in order to enhance safety protocols and precise hazard assessment.
Regarding Claim 11, Yong and Geun and Kim and Li disclose the arc risk management system of claim 7, wherein the recurrent neural network comprises a long short-term memory (LSTM), which is advantageous for time series data analysis.
Wook discloses the recurrent neural network comprises a long short-term memory (LSTM), which is advantageous for time series data analysis ( using a hybrid convolutional neural network that complexly applies a Neural Network (CNN) and a Long Short Term Memory AutoEncoder (LSTM-AE) that extracts temporal features, see page 2, para 1).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide a long short-term memory (LSTM),, as taught by Wook in the combination of Geun and Yong and Kim and Li in order to more accurately model long-term dependencies in time-series data.
Regarding Claim 13, Yong and Geun and Wook and Kim and Li disclose the arc risk management system of claim 10, wherein Young further disclose the other artificial intelligence network classifies the measurement values into two or more current levels (Page 3, para 13, where the statistical value may be, for example, a variance value of the current or a standard deviation value of the current. As another example, the statistical value may be a variance value or a standard deviation value of a value corresponding to an amount of change in current), ( Fig. 4, Page 4, para 5, where the arc detection device may calculate the amount of change of the instantaneous current (i) per predetermined time to calculate the amount of change (Δi) of the current, or may calculate the amount of change (Δi) of the current by differentiating the instantaneous current (i), e.g., the amount of change in current with respect to instantaneous current (i) and following amount of change current, corresponding for two currents levels).
Regarding Claim 15, Yong and Geun and Kim and Li disclose the arc risk management system of claim 10, wherein Young further disclose the first factor values are current levels (Fig. 9, see page 6, para 9, where a current measurement value to which the low-pass filter is applied, and a second waveform 920 is a waveform of a standard deviation value of a current measurement value to which the low-pass filter is applied, e.g, the peak of current waveform for low pass filter corresponds to the first factor value) and the artificial intelligence network which has learned to classify levels of currents flowing into the electrical apparatus in a state where an arc has occurred factor (FIG. 9 , Page 6, para 9, a first waveform 910 is a waveform of a current measurement value to which the low-pass filter is applied, and a second waveform 920 is a waveform of a standard deviation value of a current measurement value to which the low-pass filter is applied; In FIG. 10 , Page 6, para 11, where a first waveform 1010 is a standard deviation waveform for a current measurement value to which a low-pass filter is applied, a second waveform 1020 indicates an upper boundary of the reference range, and a third waveform 1030 is a reference range represents the lower boundary…).
Yong and Geun and Kim and Li do not disclose other artificial intelligence network is an artificial intelligence network which has learned to classify levels of currents.
Wook disclose the other artificial intelligence network is an artificial intelligence network which has learned to classify levels of power consumption prediction AI model/[currents](The learning unit 230 includes a spatial feature extracting unit 231, a temporal feature extracting unit 232 , and a predictive model generating unit 233 . An artificial intelligence model for power consumption prediction by receiving a training dataset output from the data preprocessor 220 and extracting spatial and temporal features from the training dataset (or “power consumption prediction AI model”) is generated and provided to the power consumption prediction unit 250, see page 4, para 12); (The spatial feature extraction unit 231 applies a convolutional neural network (CNN), extracts spatial features from the refined training dataset input from the data preprocessor 220, and extracts the spatial features from the temporal feature extraction unit 232) is output., see page 5, para 1, e.g., the CNN corresponds to the other artificial intelligence network and extracting the spatial features corresponding to the different power consumption AI model, corresponds to the different level of the power AI model/[currents], because current is a direct component of electric power).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide the other artificial intelligence network (CNN), as taught by Wook in the combination of Geun and Yong and Kim and Li in order to achieved the higher accuracy, better generalization, and less need for human intervention.
Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Yong and Geun, Kim and Li, as applied above and further in view of Dupont (WO2022020395), [hereinafter Dupont].
Regarding Claim 9, Yong and Geun and Kim and Li disclose the arc risk management system of claim 8, but do not disclose wherein the pre-processing module normalizes the measurement values by a mean subtraction normalization (MSN) method.
Dupont discloses the pre-processing module normalizes the measurement values by a mean subtraction normalization (MSN) method (para [0060], where a normalization can be performed, by “mean centering” (i.e., subtracting each data value from its measured mean to render its empirical average zero)).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide mean subtraction normalization (MSN) method, as taught by Dupont in combination of Yong and Geun and Kim and Li in order to improve algorithm performance and increase numerical stability and faster convergence.
Claim 12 is rejected under 35 U.S.C. 103 as being unpatentable over Yong and Geun, and Kim and Li, as applied above and further in view of Chao (CN114707471A), [hereinafter Chao].
Regarding Claim 12, Yong and Geun and Kim and Li disclose the arc risk management system of claim 7, but do not disclose wherein an artificial intelligence network has a structure of a residual neural network (ResNet).
Chao disclose an artificial intelligence network has a structure of a residual neural network (ResNet) (complete model form should include common classical model artificial intelligence such as Resnet, LSTM, VGG), which can realize a complete model structure through a network component, see page 7, para 2).
Therefore, it would have been obvious to one of ordinary skill in the art at the time the applicants' invention was made to provide a structure of a residual neural network (ResNet), as taught by Chao in combination of Geun and Yong and Kim and Li in order to provide training deeper models more stable and has led to state-of-the-art performance in areas like computer vision.
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 KALERIA KNOX whose telephone number is (571)270-5971. The examiner can normally be reached M-F 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, Andrew Schechter can be reached at (571)2722302. 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.
/KALERIA KNOX/
Examiner, Art Unit 2857
/ANDREW SCHECHTER/Supervisory Patent Examiner, Art Unit 2857
