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 .
Response to Amendment
The amendment filed 02/12/2026 has been entered. Claims 1-8, 10-17 and 19 are presented for examination.
Claim Rejections - 35 USC § 112
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.
Claim 7 is 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.
The following term lacks antecedent basis: “the weighted”.
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.
Claim 19 is rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter. The claim(s) does/do not fall within at least one of the four categories of patent eligible subject matter.
Claim 19 recited a computer program product comprising instructions without reciting any associated hardware structure is a software per se. Accordingly, software per se is directed to a non-statutory subject matter. Therefore claim 19 is rejected under 35 USC 101 for being directed to a non-statutory subject matter.
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, 10-11, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Wulff et al. (“Multiple Imputation by Chained Equations in Praxis”, 2017, pages 41-45, 50-52) hereafter Wulff, and in view of Chang et al. ("Multiple imputation for analysis of incomplete data in distributed health data networks", Nature Communication, 2021, pages 1-11) hereafter Chang.
Wulff was cited in the previous office action.
Regarding claim 1, Wulff teaches a method of training a machine learning model comprising:
receiving a training dataset having a plurality of original items for training a first machine learning model (Figure 1, The reception of a “training dataset” is taught by the Incomplete data set being fed into the process in Figure 1.
PNG
media_image1.png
241
684
media_image1.png
Greyscale
Page 45, “The real life data used in the example stem from The European Social Survey (ESS).”, The “plurality of original items for training” is taught by the ESS. Page 46, “MICE needs an imputation model and an analysis model. While the imputation model is used for filling in the missing values, the purpose of the analysis model is to analyze the imputed data afterwards.”, The “first machine learning model” is taught by the analysis model);
each original item comprising a plurality of features, wherein at least one original item has at least one unknown feature (Table I, The “plurality of features” for each item in the training dataset is taught in Table I by “happy, flappl, agea, gndr,” etc. The existence of “at least one original item” wherein there is at least one unknown feature is taught in Table 1 by the “Number Missing” column.)
using the plurality of features of the at least one original item on at least one additional machine learning model to generate at least one score for the at least one unknown feature of the at least one original item, wherein the at least one additional machine learning model differs from the first machine learning model using the training dataset (page 46, “MICE needs an imputation model and an analysis model. While the imputation model is used for filling in the missing values, the purpose of the analysis model is to analyse the imputed data afterwards.”, The “at least on additional machine learning model” is taught by the imputation model. The “at least one score” is taught by the filled-in missing values during imputation.), and wherein the additional machine learning model is trained with data not accessed for training the first machine learning model (page 42, “MICE runs through an iterative process: In the first iteration, the imputation model for the variable with the least missing values is estimated using only complete data. Next, the variable with the second least missing values is imputed using the complete data and the imputed values from the last iteration.”, Because multiple iterations are used for imputation, the imputation model uses incomplete data whereas the analysis model uses the filled in data. Thus, MICE teaches training the first model and the additional model with different data.);
generating an estimated distribution using the first machine learning model and the additional machine learning model (page 50, “After the imputation, we have m complete datasets. We estimate the model (e.g. linear regression) on each of the m datasets and combine the estimates to one combined result.”, The “first machine learning model” is taught by the analysis model (e.g. linear regression). The “estimated distribution” that is generated is taught by the m complete datasets. The “additional machine learning model” is taught by the imputation model.);
receiving an estimate of a confidence interval of the at least one score, wherein the receiving of the estimate of the confidence interval is from the first machine learning model and the additional machine learning model (page 52, “Table IV presents the results from the stepwise model selection procedure after multiple imputation.”. Table IV - “Lower CI” and “Upper CI”, The CI is computed following the use of the imputation and analysis models);
generating a predicted feature by applying the estimate of the confidence interval on the at least one score (page 52, “This procedure performs stepwise variable selection on each imputed data set. If a given variable is selected its estimate is different from zero while it is zero if the variable is not selected. Thus, coefficients of variables selected in only a few imputed data sets are shrunk towards zero when combined.”, Application of the confidence intervals is taught in Wulff by the usage of the confidence interval bounds for determining which variable coefficients are to be included in the predictive model. That is, variable coefficients estimated to be 0 or with a confidence interval that includes 0 are excluded in the predictive regression model.); and
assigning the predicted feature to the at least one unknown feature (page 50, “After the imputation, we have m complete datasets. We estimate the model (e.g., linear regression) on each of the m datasets and combine the estimates to one combined result.”, The assignment of the predicted feature to the unknown feature is taught by the imputations.).
Wulff did not specially teach wherein the additional machine learning model is not the first machine learning model intended for training and is trained using data that is at least partially unavailable to the first machine learning model intended for training.
However, Chang teaches wherein the additional machine learning model is not the first machine learning model intended for training and is trained using data that is at least partially unavailable to the first machine learning model intended for training (page 1, “We develop communication-efficient distributed multiple imputation methods for incomplete data that are horizontally partitioned. Since subject-level data are not shared or transferred outside of each site in the proposed methods, they enhance protection of patient privacy and have the potential to strengthen public trust in analysis of sensitive health data”. Page 2 “To enhance protection of patient privacy, we investigate distributed MI methods for handling missing data that do not require sharing individual level data between sites. Under MAR, one straightforward privacy-preserving MI approach for horizontally partitioned incomplete data would be to conduct MI within each institution/site and then perform the distributed analysis”. Page 10 “Distributed MI for general missing data patterns. Since it is commonly encountered in practice to have multiple variables with missing values, it is of particular interest to develop privacy-preserving distributed MI methods for general missing data patterns. A very popular method for handling general missing data patterns is the MI by chained equation, known as MICE in short20. Without loss of generality, we assume that the first q (q < p) covariates, i.e., (X1, . . . , Xq), have missing values. The MICE algorithm starts with an initial imputation. For example, the missing values of Xj can be imputed by random samples from the observed data. Then, for each j= 1, …, q, the missing values of Xj are imputed by a MI method, assuming that the imputed values of all the other variables were actually observed. A sweep of imputations for all q missing variables form an iteration, and multiple iterations are fulfilled until the distribution of imputed values becomes stationary before the first imputed dataset is sampled. Multiple iterations are carried out between each imputed dataset to alleviate autocorrelations. A total of M imputed datasets is collected and used to fit the analysis model separately, and the results are combined by Rubin’s rule. Readers referred to ref. 20 for more details about MICE.).
It would have been obvious to one of an ordinary skill in the art before the time of the effective filling date of the claimed invention to have included the concept of the additional machine learning model is not the first machine learning model intended for training, and is trained using data that is at least partially unavailable to the first machine learning model intended for training as suggested in Chang into Wulff’s system because both of these system are addressing the need of handling missing data. By incorporating the teaching of Chang into Wulff would enhance protection of patient privacy and have the potential to strengthen public trust in analysis of sensitive health data (Chang, abstract).
Regarding claim 2, the combination of Wulff and Chang teaches the method of claim 1, further comprising:
generating an additional predicted feature by applying the estimate of the confidence interval on the at least one score (page 52, “This procedure performs stepwise variable selection on each imputed data set. If a given variable is selected its estimate is different from zero while it is zero if the variable is not selected. Thus, coefficients of variables selected in only a few imputed data sets are shrunk towards zero when combined.”, Application of the confidence intervals is taught in Wulff by the usage of the confidence interval bounds for determining which variable coefficients are to be included in the predictive model. That is, variable coefficients estimated to be 0 or with a confidence interval that includes 0 are excluded in the predictive regression model); and
adding a synthetic item comprising substantially the plurality of features and the additional predicted feature assigned to the at least one unknown feature to the training dataset (page 52, “Reporting the regression results based on imputed data is not different from regression results based on regular data. However, when using multiple imputation to account for missing data, the data analyst must supply additional information either in the main paper / report or in the supplement materials. Because many of the decisions we make when imputing data may affect the end-result, it is reasonable to state these decisions.”, Here it is noted in the source that the imputed results are used in a regression, i.e., imputed results were necessarily used in the training process for the regression thus the source teaches the claim element of “adding a synthetic item.... to the training dataset”. Per the specification, “synthetic items” are construed as being samples that are the combination of original features and generated missing features. (Specification 16:31, “Additionally, a plurality of synthetic items, shown in 541, each comprising substantially the plurality of features and a predicted feature from the plurality of predicted features, assigned to the at least one unknown feature, may be added to the training dataset for some, or each of the original items”. Per the specification, “comprising substantially” is construed as comprising of between .1% and 20% of the total features. (Specification 19:4, “As used herein the terms “about” or “substantially” refer to functionally equivalent, which ay range from ±20% to ±0.1% according to the domain, and may include other modifications in non-numerical context.”).
Regarding claims 10-11, these are system claims corresponding to method claims 1-2. Therefore, they are rejected for the same reason as method claims 1-2 above.
Regarding claim 19, this is a program product claim corresponding to method claim 1. Therefore, it is rejected for the same reason as method claim 1 above.
Claims 5 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over Wulff et al. (“Multiple Imputation by Chained Equations in Praxis”, 2017, pages 41-45, 50-52) hereafter Wulff, in view of Chang et al. ("Multiple imputation for analysis of incomplete data in distributed health data networks", Nature Communication, 2021, pages 1-11) hereafter Chang, as applied in claims 1 and 10 above, and further in view of Jiang (“Simple Weighted Average Ensemble Machine Learning”, October 2020).
Jiang was cited in the previous office action.
Regarding claim 5, the combination of Wulff and Change teaches the method of claim 1 further comprising using the plurality of features of the at least one original item on at least one further additional machine learning model to generate at least one additional score for the at least one unknown feature of the at least one original item; receiving an additional estimate of the confidence interval of the at least one additional score (Table IV, The table teaches a confidence interval that is computed for each score.); and
and adding a synthetic item comprising substantially the plurality of features and the additional predicted feature assigned to the at least one unknown feature to the training dataset (page 52, “Reporting the regression results based on imputed data is not different from regression results based on regular data. However, when using multiple imputations to account for missing data, the data analyst must supply additional information either in the main paper / report or in the supplement materials. Because many of the decisions we make when imputing data may affect the end-result, it is reasonable to state these decisions.”, Here it is noted in the source that the imputed results are used in a regression, i.e., imputed results were necessarily used in the training process for the regression thus the source teaches the claim element of “adding a synthetic item.... to the training dataset”).
The combination of Wulff and Chang did not specially teach generating an additional predicted feature by applying a weighted averaging using the additional estimate of the confidence interval on the at least one additional score.
However, Jiang teaches generating an additional predicted feature by applying a weighted averaging using the additional estimate of the confidence interval on the at least one additional score (third step heading – lines 16 to 21, Jiang teaches ensemble learning and the practice of using a weighted average to combine multiple estimates into a single prediction.)
##### WEIGHTED AVERAGE
weighted_average = (0.3*XGB_pred+
0.2*DecisionTree_pred+
0.2*RandomForest_pred+
0.15*MLPC_pred+
0.15*KNN_pred)
It would have been obvious to one of an ordinary skill in the art before the time of the effective filling date of the claimed invention to have included the concept of generating an additional predicted feature by applying a weighted averaging using the additional estimate of the confidence interval on the at least one additional score as suggested in Jiang into the combination of Wulff and Chang’s system because all of these systems are directed towards statistical computation. It would have been obvious for one of ordinary skill in statistics and statistical computing to implement Jiang’s weighted averaging of multiple machine learning algorithm predictions into Wulff and Chang’s disclosed system to impute missing features as described in Wulff and Chang’s system.
Regarding claim 14, this is a system claim corresponding to method claim 5. Therefore, it is rejected for the same reason as claim 5 above.
Claims 6 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Wulff et al. (“Multiple Imputation by Chained Equations in Praxis”, 2017, pages 41-45, 50-52) hereafter Wulff, in view of Chang et al. ("Multiple imputation for analysis of incomplete data in distributed health data networks", Nature Communication, 2021, pages 1-11) hereafter Chang, as applied in claims 1 and 10 above, and further in view of Liu (“Applying Domain Knowledge to Clinical Predictive Models”, 2016, pp. 3).
Liu was cited in the previous office action.
Regarding claim 6, the combination of Wulff and Chang teaches the method of claim 1. The combination of Wulff and Chang did not specifically teach adjusting the predicted feature by applying domain knowledge after receiving the estimate of the at least one score and the confidence interval from the at least one additional machine learning model for the predicted feature, using the plurality of features.
However, Liu teaches adjusting the predicted feature by applying domain knowledge after receiving the estimate of the at least one score and the confidence interval from the at least one additional machine learning model for the predicted feature, using the plurality of features (Abstract, “Clinical predictive models are useful in predicting a patient's risk of developing adverse outcomes and in guiding patient therapy. In this thesis, we explored two different ways to apply domain knowledge to improve clinical predictive models.”).
It would have been obvious to one of an ordinary skill in the art before the time of the effective filling date of the claimed invention to have included the concept of adjusting the predicted feature by applying domain knowledge after receiving the estimate of the at least one score and the confidence interval from the at least one additional machine learning model for the predicted feature, using the plurality of features as suggested in Liu into the combination of Wulff and Chang’s system because all of these systems are directed towards predictive modeling. It would have been obvious for one of ordinary skill in statistical computing and machine learning to implement Liu’s application of domain knowledge into Wulff and Chang’s disclosed system before the filing date of the claimed invention to impose domain-specific limits to the predicted feature imputations described in Wulff and Chang to increase the accuracy of the imputations.
Regarding claim 15, this is system claim corresponding to method claim 6. Therefore, it is rejected for the same reason as method claim 6 above.
Claims 7 is rejected under 35 U.S.C. 103 as being unpatentable over Wulff et al. (“Multiple Imputation by Chained Equations in Praxis”, 2017, pages 41-45, 50-52) hereafter Wulff, in view of Chang et al. ("Multiple imputation for analysis of incomplete data in distributed health data networks", Nature Communication, 2021, pages 1-11) hereafter Chang, as applied in claim 1 above, and further in view of Olson et al., (“Improved prediction of Canada lynx distribution through regional model”, January 2021, pp. 1671) hereafter Olson.
Olson was cited in the previous office action.
Regarding claim 7, the combination of Wulff and Chang teaches the method of claim 1. The combination of Wulff and Chang did not specifically teach the weighted averaging is based on the respective area under the receiver operating characteristic curve of the at least one additional machine learning model.
However, Olson teaches the weighted averaging is based on the respective area under the receiver operating characteristic curve of the at least one additional machine learning model (page 1671, “This resulted in 60 models per scenario, which were combined into a weighted average based on area under the curve (AUC) of the receiver operating characteristic (ROC), so that better performing models contributed more to the final ensemble, with the threshold for inclusion greater or equal to the median AUC calculated from all 60 models.”).
It would have been obvious to one of an ordinary skill in the art before the time of the effective filling date of the claimed invention to have included the concept of having the weighted averaging is based on the respective area under the receiver operating characteristic curve of the at least one additional machine learning model as suggested in Olson into the combination of Wulff and Chang’s system because all of these systems are directed towards predictive modeling. It would have been obvious for one of ordinary skill in statistical computing and machine learning to implement Olson’s weighted averaging by utilizing the ROC into Wulff and Chang’s disclosed system before the filing date of the claimed invention. It would have been obvious because one of ordinary skill would be motivated to utilize an ensemble of models to create a more accurate feature imputation in Wulff and Chang.
Claims 8 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Wulff et al. (“Multiple Imputation by Chained Equations in Praxis”, 2017, pages 41-45, 50-52) hereafter Wulff, in view of Chang et al. ("Multiple imputation for analysis of incomplete data in distributed health data networks", Nature Communication, 2021, pages 1-11) hereafter Chang, as applied in claims 1 and 10 above, and further in view of Carney et al. (“Confidence and Prediction Intervals for Neural Network Ensembles”, 1999, pp. 3) hereafter Carney.
Carney was cited in the previous office action.
Regarding claim 8, the combination of Wulff and Chang teaches the method of claim 1. The combination of Wulff and Chang did not specifically teach receiving an estimate of the confidence interval is selected from a group of methods consisting of: receiving the estimate of the confidence interval from the at least one additional machine learning model, an analysis of the training data and model performance, and modulating the plurality of features and applying the at least one additional machine learning model thereon.
However, Carney teaches receiving an estimate of the confidence interval is selected from a group of methods consisting of: receiving the estimate of the confidence interval from the at least one additional machine learning model, an analysis of the training data and model performance, and modulating the plurality of features and applying the at least one additional machine learning model thereon (page 3, “To form our confidence intervals, we need to estimate the variance of the distribution P(f(x) | φcom(x)). However, we have no direct access to this distribution, and we do not know what the true regression f(x) is. Using the outputs of the bootstrap re-sampled networks in the ensemble, we can however approximate it.” I.e., the source teaches using a machine learning model (neural networks) to estimate (approximate) the variance of a distribution which in turn provides an estimate confidence interval. Moreover, the source proceeds with discussion of modulating the inputs to the predictive model and reassessing the confidence interval of the distribution: “Using a large number of bootstrap networks for the ensemble (we use 200, but fewer would suffice) we divide the ensemble into M smaller ensembles (we use 8 groups of 25 networks each). This gives us a set of M φcom(x) values.” The M values are denoted in the source by Ϛ thereafter, see equation (7). Carney proceeds to describe bootstrapping the modulated values: “…we undergo one more iteration of bootstrapping. We form P (we use P = 1000) bootstrap re-sampled sets of Ϛ.” Finally, Carney proceeds to describe recomputing the variance using the modulated samples: “We calculate a variance measure for each of these sets and then calculate an average of these to provide a smoother, lower variance estimate of the variance of the distribution…” and calculating the confidence interval: “Now that we have a good variance estimate for the distribution… we can calculate a confidence interval in the usual fashion.”).
It would have been obvious to one of an ordinary skill in the art before the time of the effective filling date of the claimed invention to have included the concept of having receiving an estimate of the confidence interval is selected from a group of methods consisting of: receiving the estimate of the confidence interval from the at least one additional machine learning model, an analysis of the training data and model performance, and modulating the plurality of features and applying the at least one additional machine learning model thereon in Carney into the combination of Wulff and Chang’s system because all of these systems are directed towards predictive modeling. It would have been obvious for one of ordinary skill in statistical computing and machine learning to implement Carney’s computation of confidence intervals for neural network ensembles into the disclosed systems of Wulff and Chang before the filing date of the claimed invention. It would have been obvious because one of ordinary skill would be motivated to compute an accurate confidence interval for the machine learning models for use with the systems described in Wulff and Chang.
Regarding claim 17, this is system claim corresponding to method claim 8. Therefore, it is rejected for the same reason as method claim 8 above.
Claim 16 is rejected under 35 U.S.C. 103 as being unpatentable over Wulff et al. (“Multiple Imputation by Chained Equations in Praxis”, 2017, pages 41-45, 50-52) hereafter Wulff, in view of Chang et al. ("Multiple imputation for analysis of incomplete data in distributed health data networks", Nature Communication, 2021, pages 1-11) hereafter Chang, in view of Jiang (“Simple Weighted Average Ensemble Machine Learning”, October 2020), as applied in claim 14, and further in view of Olson et al., (“Improved prediction of Canada lynx distribution through regional model”, January 2021, pp. 1671) hereafter Olson.
Regarding claim 16, the combination of Wulff, Chang and Jiang teach the method of claim 14. The combination of Wulff, Chang and Jiang did not specifically teach wherein the weighted averaging is based on the respective area under a receiver operating characteristic curve of the at least one additional machine learning model.
However, Olson teaches the weighted averaging is based on the respective area under a receiver operating characteristic curve of the at least one additional machine learning model (page 1671, “This resulted in 60 models per scenario, which were combined into a weighted average based on area under the curve (AUC) of the receiver operating characteristic (ROC), so that better performing models contributed more to the final ensemble, with the threshold for inclusion greater or equal to the median AUC calculated from all 60 models.”).
It would have been obvious to one of an ordinary skill in the art before the time of the effective filling date of the claimed invention to have included the concept of having the weighted averaging is based on the respective area under a receiver operating characteristic curve of the at least one additional machine learning model as suggested in Olson into the combination of Wulff ,Chang and Jiang’s system because all of these systems are directed towards predictive modeling. It would have been obvious for one of ordinary skill in statistical computing and machine learning to implement Olson’s weighted averaging by utilizing the ROC into Wulff, Chang and Jiang’s disclosed systems before the filing date of the claimed invention to utilize an ensemble of models to create a more accurate feature imputation in Wulff, Chang and Jiang.
Allowable Subject Matter
Claims 3-4 and 12-13 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.
Response to Arguments
Applicants’ arguments, filed 02/26/2026, with respect to the rejection(s) of claim(s) 1-8, 10-17 and 19 under 35 USC 102 and 35 USC 103 are moot in view of the new ground of rejection (see rejection above).
Applicants’ arguments, filed 02/26/2026, with respect to the rejection(s) of claim(s) 1-8, 10-17 and 19 under 35 USC 101 as directed to a judicial exception without significantly more are persuasive. Therefore, the rejection is withdrawn.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Eilert et al (US 11,755,884 B2) disclose distributed machine learning with privacy protection.
Sun et al (US 2022/0327428 A1) disclose executing machine learning models.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JENNIFER N WELCH whose telephone number is (571)272-7212. The examiner can normally be reached M-T 5:30AM - 4:00PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, DAVID WILEY can be reached at (571) 272-4150. 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.
JENNIFER N. WELCH
Supervisory Patent Examiner
Art Unit 2143
/JENNIFER N WELCH/Supervisory Patent Examiner, Art Unit 2143
Prosecution Insights