DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This action is made final.
Claims 1-20 are pending. Claims 1, 9 and 17 are independent claims.
Drawings
The objection to the drawings has been withdrawn.
Response to Arguments
With respect to the 35 U.S.C. 112(a) rejections of the previous office action, the applicant argues that the binomial distribution is well-known. The examiner agrees that the binomial distribution is well-known; however, the claims fail to satisfy the enablement requirement because it is not explained how the distribution will be used to determine subset sizes.
With respect to the 35 U.S.C. 101 rejections of the previous office action, the applicant argues that the claims are not directed to a judicial exception because the claims as a whole integrate the recited judicial exception into a practical application. Due to the claim amendments, the scope of the claims has changed and new grounds of rejection are applied – see the updated rejection below. Applicant's arguments have been fully considered but they are not persuasive. More specifically, the applicant argues, that the invention provides technical benefits as described in the specification ¶11-13. However, the claim limitations do not reflect this improvement because the additional elements in claim 1 that refer to satisfying an average calibration, “by training a neural network to perform an analysis that satisfies average calibration” and the newly added “to enable the performance of the analysis that satisfies the average calibration”, specify a field of use without significantly more, and generally recite an effect of applying the judicial exception with no description as to how the abstract idea steps are to be executed. The applicant further argues that Claim 4 provides further evidence of integration into a practical application. The examiner argues that claim 4 is specifying a field of use (i.e., equipment maintenance) without significantly more, and therefore cannot integrate the judicial exception into a practical application.
With respect to the 35 U.S.C. 103 rejections of the previous office action, the applicant argues that Lafond in view of Buttner, Du, and Bongaarts fails to teach the limitations of claim 1. Applicant's arguments have been fully considered but they are not persuasive. More specifically, applicant argues that Buttner does not disclose or suggest “updating weights of the neural network based on the Bregman divergence to enable performance of the analysis that satisfies the average calibration” (amended portion emphasized). Due to the claim amendments, the scope of the claims has changed and new grounds of rejection are applied – see the updated rejection below. Applicant further argues that Du does not disclose or suggest a “distribution vector”, and more specifically does not disclose or suggest “generating a distribution vector for the subset of the outcomes vector that corresponds to the subset of the set of feature vectors”. The examiner argues that the vector disclosed by Du is a distribution vector, as it reflects the possible outcomes (3 possible outcomes, i.e., an outcomes vector) and their probabilities (3 probabilities that add up to 1, i.e., a probability distribution) given an event sequence (i.e., a feature vector). The applicant further argues that Lafond does not disclose or suggest “calculating a Bregman divergence between the distribution vector and a scoring distribution vector of the prediction vector”. The examiner argues that because the Bregman divergence is used to calculate a disagreement score in Lafond, either of the distribution vectors used can be broadly interpreted as a “scoring distribution vector”.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 7 and 15 are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the enablement requirement. The claim(s) contain subject matter which was not described in the specification in such a way as to enable one skilled in the art to which it pertains, or with which it is most nearly connected, to make and/or use the invention. Specifically, both claims recite ".
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1 rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 17 recites the limitation "the set of feature vectors" in lines. There is insufficient antecedent basis for this limitation in the claim.
Claim 18 recites the limitation "the set of feature vectors" in line 8. There is insufficient antecedent basis for this limitation in the claim.
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:
Step 1: This part of the eligibility analysis evaluates whether the claim falls within any statutory
category. See MPEP 2106.03. Claim 1 is directed to a process (Step 1: YES).
Step 2A prong 1: Does the claim recite a judicial exception? Claim 1 recites: A computer-implemented method for improving the performance of a machine learning system… the method comprising… repeatedly… selecting a subset of the set of feature vectors (selecting a subset from a set is a mental process); generating a distribution vector for a subset of the outcomes vector that corresponds to the subset of the set of feature vectors (generating a distribution vector is a mathematical calculation); producing a prediction vector by running the neural network on the subset of the set of feature vectors (running the neural network on an input is a series of mathematical calculations); calculating a Bregman divergence between the distribution vector and a scoring distribution vector of the prediction vector (calculating a Bregman divergence between two distribution vectors is a mathematical calculation). These steps can be performed mentally or are mathematical calculations (Step 2A prong 1: YES).
Step 2A prong 2: Does the claim recite additional elements? Do those additional elements,
considered individually and in combination, integrate the judicial exception into a practical application? Claim 1 recites: by training a neural network to perform an analysis that satisfies average calibration… obtaining, using at least one hardware processor, an outcomes vector and a set of feature vectors, each of which corresponds to one of the outcomes in the outcomes vector; and… using at least one hardware processor… and updating weights of the neural network based on the Bregman divergence to enable the performance of the analysis that satisfies the average calibration. Specifying that the improvement of the system is achieved by training the neural network to satisfy average calibration and using a hardware processor are additional elements specifying a field of use without significantly more. Obtaining the outcomes vector and feature vectors is extra-solution activity of data gathering that does not add a meaningful limitation to the process of improving the machine learning system. Updating weights is an additional element of data outputting that does not add a meaningful limitation to the process of improving the machine learning system. Specifying that the updating enables performance of analysis that satisfies the average calibration is an attempt to use the neural network model by merely applying the abstract idea without placing any limits on how the neural network model operates. Further, the claim omits any details as to how the neural network model solves a technical problem and instead recites only the idea of a solution or outcome, equivalent to adding the words “apply it” to the judicial exception – see MPEP 2106.05(f) (Step 2A prong 2: NO).
Step 2B: These elements are recited at such a high level of generality that they fail to integrate the abstract idea into a practical application, since they only amount to data gathering or outputting without significantly more (MPEP 2106.05(g)), limit the field of use without significantly more (MPEP 2106.05(h)), or provide nothing more than mere instructions to implement an abstract idea on a generic computer (MPEP 2106.05(f)). These limitations, taken either alone or in combination, fails to provide an inventive concept (Step 2B: NO). Thus, the claim is not patent eligible.
Regarding claims 2-8, they recite limitations which further narrow the abstract idea by specifying more details of the mental and mathematical process that occurs (Claim 2, splitting a vector into buckets and creating a distribution vector of a specified size are mental processes, and assigning a value equal to the result of subtraction to an extra dimension is a mathematical calculation; Claim 3, using a Kaplan-Meier estimator is a mathematical formula; Claim 4, specifying an equipment outcomes vector and performing maintenance is specifying a field of use without significantly more; Claim 5, using squared loss is a mathematical formula; Claim 6, using a Kullback-Leibler divergence is a mathematical formula; Claim 7, drawing a subset from other subsets is a mental process, and the use of a binomial distribution is a mathematical formula; Claim 8, drawing a subset randomly from a collection of subsets is a mental process).
Regarding claim 9:
Step 1: This part of the eligibility analysis evaluates whether the claim falls within any statutory
category. See MPEP 2106.03. Claim 9 is directed to a process (Step 1: YES).
Step 2A prong 1: Does the claim recite a judicial exception? Claim 9 recites: A computer-implemented method for improving the performance of a machine learning system... the method comprising: selecting… a subset of a set of feature vectors; selecting… a subset of an outcomes vector, wherein each outcome in the subset of the outcomes vector corresponds to one of the feature vectors in the subset of the set of feature vectors (selecting a subset of vectors from a set of vectors and selecting a subset of a vector are mental processes); splitting… the subset of the outcomes vector into a number of buckets (splitting a vector into buckets is a mental process); initializing… a distribution vector that has a number of dimensions corresponding to the number of buckets plus an extra dimension (creating a vector with certain dimensions is a mental process); assigning to each dimension of the distribution vector… a value equal to a number of data points in a corresponding bucket (assigning values in a vector is a mental process); and assigning to the extra dimension of the distribution vector… a value equal to a total population of the subset of the set of feature vectors minus a total number of data points in the subset of the outcomes vector (assigning a result of subtraction to a vector is a mental or mathematical process). These steps can be performed mentally or are mathematical calculations (Step 2A prong 1: YES).
Step 2A prong 2: Does the claim recite additional elements? Do those additional elements, considered individually and in combination, integrate the judicial exception into a practical application? Claim 1 recites: by training a neural network to perform an analysis that satisfies average calibration… using at least one hardware processor… using the at least one hardware processor… using the at least one hardware processor… using the at least one hardware processor… using the at least one hardware processor… using the at least one hardware processor. Specifying that the improvement of the system is achieved by training the neural network to satisfy average calibration and using a hardware processor are additional elements specifying a field of use without significantly more (Step 2A prong 2: NO).
Step 2B: These elements are recited at such a high level of generality that they fail to integrate the abstract idea into a practical application, since they amount to limiting the field of use without significantly more (MPEP 2106.05(h)). These limitations, taken either alone or in combination, fails to provide an inventive concept (Step 2B: NO). Thus, the claim is not patent eligible.
Regarding claims 10-16, they recite limitations which further narrow the abstract idea by specifying more details of the mental and mathematical process that occurs (Claim 10, calculating a Bregman divergence is a mathematical calculation, updating the neural network weights is extra-solution activity of data outputting that does not add a meaningful limitation to the process of improving the machine learning system, and selecting a different subset of a set of feature vectors and the outcomes vector is a mental process; Claim 11, using squared loss is a mathematical formula; Claim 12, using a Kullback-Leibler divergence is a mathematical formula; Claim 13, using a Kaplan-Meier estimator is a mathematical formula; Claim 14, using a Nelson-Aalen estimator is a mathematical formula; Claim 15, drawing a subset from other subsets is a mental process, and the use of a binomial distribution is a mathematical formula; Claim 16, drawing a subset randomly from a collection of subsets is a mental process and having a specific number of subsets that is derived from the result of a division operation is a mathematical calculation).
Regarding claim 17, it is an apparatus implementing the method of claim 1 and is rejected on the same grounds – see above.
Regarding claims 18-20, they recite limitations which further narrow the abstract idea by specifying more details of the mental and mathematical process that occur within the apparatus (Claim 18, splitting a vector into buckets and creating a distribution vector of a specified size are mental processes, and assigning a value equal to the result of subtraction to an extra dimension is a mathematical calculation; Claim 19, using a Kullback-Leibler divergence is a mathematical formula; Claim 20, receiving a feature vector and producing an outcomes vector are extra-solution activities of date inputting and data outputting that do not add a meaningful limitation to the machine learning apparatus).
Allowable Subject Matter
Claims 2, 3, 7 and 18 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Claims 9-16 allowed.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1, 6, 17, 19 and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lafond et al. (US 20220027764 A1), herein Lafond, in view of Büttner et al. (US 20210110253 A1), herein Büttner, Du et al. (US 20200342305 A1), herein Du, and Bongaarts et al. (US 20190132414 A1), herein Bongaarts.
Regarding claim 1, Lafond teaches A computer-implemented method for improving the performance of a machine learning system (¶17, enable capturing human decision policies with minimal labeled examples without reducing prediction accuracy of models) by training a neural network (¶23, training, by the set of supervised machine learning algorithms, the set of machine learning models)… the method comprising: obtaining, using at least one hardware processor, an outcomes vector and a set of feature vectors, each of which corresponds to one of the outcomes in the outcomes vector (¶69, obtaining, for each training feature vector of the set of training feature vectors, a respective label – a label can be an outcome); and repeatedly, using at least one hardware processor: selecting a subset of the set of feature vectors (¶389, It will be appreciated that processing steps 904 to 910 may be repeated every time a feature vector is obtained by the processing device)… producing a prediction vector by running the neural network on the subset of the set of feature vectors (¶250, processes the feature vector representing an object, and outputs a respective prediction to obtain a set of predictions for the set of ML models 260. Each respective prediction is associated with a respective accuracy metric); calculating a Bregman divergence between the distribution vector and a scoring distribution vector of the prediction vector (¶38, In one or more embodiments of the method, said determining the respective disagreement score comprising determining a respective Kullback-Leibler (KL) divergence for the respective set of predictions – a Kullback-Leibler divergence is an example of a Bregman divergence); and updating weights of the neural network (¶209, During training, internal parameters of each of the set of MLAs 250 are updated by performing predictions using the feature vectors, and by comparing the predictions with the labels).
Lafond fails to explicitly teach a method to perform an analysis that satisfies average calibration… to enable the performance of the analysis that satisfies the average calibration..
However, in the same field of endeavor, Büttner teaches a method to perform an analysis that satisfies average calibration (¶6, A new training strategy is provided by the present invention combining conventional loss with… an adversarial calibration loss term – this term, as described in ¶8, is computed as the Euclidian norm (L2 norm) of an expected calibration error ECE)… updating weights to enable the performance of the analysis that satisfies the average calibration (¶166, weights of the NN are updated second time based on the adversarial calibration loss – updating the NN weights based on a loss term based on the ECE would result in an analysis that satisfies the ECE, which is an averaged calibration measurement)…
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to add a term that represents average calibration error to the loss function as disclosed by Büttner to the method disclosed by Lafond to achieve accurate results on novel data (Büttner, ¶6, obtaining well-calibrated, trustworthy probabilities for both in-domain samples as well as out-of-domain samples).
Lafond in view of Büttner fails to teach generating a distribution vector for a subset of the outcomes vector that corresponds to the subset of the set of feature vectors.
However, in the same field of endeavor, Du teaches generating a distribution vector for a subset of the outcomes vector that corresponds to the subset of the set of feature vectors (¶27, For example, from event sequence 110, computing system 101 determines that three outcomes are possible, with respective probabilities 0.8, 0.1, and 0.1).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to generate a distribution vector for the outcomes vector as disclosed by Du in the method disclosed by Lafond in view of Büttner to allow a user to prioritize more likely outcomes (Du, ¶49, analysis application 102 displays the most probable future paths by preserving the most probable event at each prediction step).
Lafond in view of Büttner and Du fails to teach updating weights based on the Bregman divergence.
However, in the same field of endeavor, Bongaarts teaches updating weights based on the Bregman divergence (¶106, loss functions can include mean squared error, L2 loss, Kullback-Leibler divergence – loss functions are used in backpropagation to update neural network weights).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use a Bregman divergence to update neural network weights as disclosed by Bongaarts in the method disclosed by Lafond in view of Büttner and Du to detect when there is a poor fit between model output and training data (Bongaarts, ¶106, Error can be determined during training using a loss function in which higher values represent a poorer fit between CM outputs and training-data output values).
Regarding claim 17, it is an apparatus implementing a method similar to that of claim 1 and is rejected on the same grounds – see above.
Regarding claim 6, Lafond in view of Büttner and Du fails to teach: The method of claim 1, further comprising: using a Kullback-Leibler divergence as the Bregman divergence.
However, Bongaarts teaches: using a Kullback-Leibler divergence as the Bregman divergence (¶106, loss functions can include mean squared error, L2 loss, Kullback-Leibler divergence).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use a Kullback-Leibler divergence to update neural network weights as disclosed by Bongaarts in the method disclosed by Lafond in view of Büttner and Du to detect when there is a poor fit between model output and training data (Bongaarts, ¶106, Error can be determined during training using a loss function in which higher values represent a poorer fit between CM outputs and training-data output values).
Regarding claim 19, Lafond in view of Büttner and Du fails to teach: The apparatus of claim 17, wherein the at least one processor is further configured by the computer executable instructions to facilitate: using a Kullback-Leibler divergence as the Bregman divergence.
However, Bongaarts teaches: using a Kullback-Leibler divergence as the Bregman divergence (¶106, loss functions can include mean squared error, L2 loss, Kullback-Leibler divergence).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use a Kullback-Leibler divergence to update neural network weights as disclosed by Bongaarts in the system disclosed by Lafond in view of Büttner and Du to detect when there is a poor fit between model output and training data (Bongaarts, ¶106, Error can be determined during training using a loss function in which higher values represent a poorer fit between CM outputs and training-data output values).
Regarding claim 20, Lafond further teaches: The apparatus of claim 19, wherein the at least one processor is further configured by the computer executable instructions to facilitate: receiving a novel feature vector; and producing a novel outcomes vector by running the neural network on the novel feature vector (¶278, The online learning procedure 600 is executed asynchronously and continuously in real-time every time data to process becomes available, i.e. when an unlabelled feature vector 612 is obtained).
Claim(s) 4 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lafond in view of Büttner, Du and Bongaarts as applied to claim 1 above, and further in view of Malhotra et al. (US 20200012921 A1), herein Malhotra.
Regarding claim 4, Lafond in view of Büttner, Du and Bongaarts fails to teach: The method of claim 1, further comprising: the at least one hardware processor producing an equipment outcomes vector by deploying the trained neural network on an equipment feature vector; and performing preventative maintenance on equipment, based on review of the equipment outcomes vector.
However, in the same field of endeavor, Malhotra teaches: The method of claim 1, further comprising: the at least one hardware processor producing an equipment outcomes vector by deploying the trained neural network on an equipment feature vector; and performing preventative maintenance on equipment, based on review of the equipment outcomes vector (¶23, Estimating remaining useful life (RUL) for equipment using sensor data streams is useful to enable condition based maintenance, and avoid catastrophic shutdowns due to impending failures).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use an equipment outcomes vector to determine when maintenance should be done as disclosed by Malhotra in the method disclosed by Lafond in view of Büttner, Du and Bongaarts in order to avoid equipment failure (Malhotra, ¶23, avoid catastrophic shutdowns due to impending failures).
Claim(s) 5 is/are rejected under 35 U.S.C. 103 as being unpatentable Lafond in view of Büttner, Du and Bongaarts as applied to claim 1 above, and further in view of Kojima et al. (US 20220138375 A1), herein Kojima.
Regarding claim 5, Lafond in view of Büttner, Du and Bongaarts fails to explicitly teach: The method of claim 1, further comprising: using a squared loss as the Bregman divergence.
However, in the same field of endeavor, Kojima teaches: using a squared loss as the Bregman divergence (¶60, The L2 divergence is defined as a squared error of two probability density functions. Which divergence should be used depends on the problem).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use squared loss to measure divergence as disclosed by Kojima in the method disclosed by Lafond in view of Büttner, Du and Bongaarts to get fast and accurate results (Kojima, ¶86, to suppress a computation time and accurately estimate the parameters of the model representing a probability distribution of censored data).
Claim(s) 8 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lafond in view of Büttner, Du and Bongaarts as applied to claim 1 above, and further in view of Tsamoura et al. (US 20220391704 A1).
Regarding claim 8, Lafond in view of Büttner, Du and Bongaarts fails to teach: The method of claim 1, wherein selecting the subset of the set of feature vectors comprises drawing the subset from a collection of random subsets of D, wherein each of the random subsets is of a same size k.
However, in the same field of endeavor, Tsamoura teaches The method of claim 1, wherein selecting the subset of the set of feature vectors comprises drawing the subset from a collection of random subsets of D, wherein each of the random subsets is of a same size k (¶104, Next, random subsets of the list are selected of size p).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to add constant size random subset selection as disclosed by Tsamoura to the method disclosed by Lafond in view of Büttner, Du and Bongaarts to potentially improve model input quality (Tsamoura, ¶100, improve the chances of providing meaningful feedback to the machine learning model).
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 HARRISON CHAN YOUNG KIM whose telephone number is (571)272-0713. The examiner can normally be reached Monday - Thursday 10:00 am - 7:00 pm.
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, Cesar Paula can be reached at (571) 272-4128. 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.
/HARRISON C KIM/Examiner, Art Unit 2145
/CESAR B PAULA/Supervisory Patent Examiner, Art Unit 2145
Prosecution Insights