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 .
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 stand rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception without significantly more.
Step 1 analysis:
In the instant case, the claims are directed to a method, an apparatus, and a tangible non-transitory computer-readable medium. Thus, each of the claims falls within one of the four statutory categories (i.e., process, machine, manufacture, or composition of matter).
Step 2A analysis:
Based on the claims being determined to be within of the four categories (Step 1), it must be determined if the claims are directed to a judicial exception (i.e., law of nature, natural phenomenon, and abstract idea), in this case the claims fall within the judicial exception of an abstract idea. Specifically the abstract ideas of Mathematical Concepts (including mathematical relationships, formulas, and/or calculations) and Mental processes (including observation, evaluation, judgment and/or opinion).
Independent Claim 1 (and Claims 13 and 20 as being analogous):
Step 2A: Prong 1 analysis:
“determining, by the device, a measured utility of the respective data on each of the plurality of trainer clients”- Examiner interprets “determining” to be equivalent in purpose to “calculating” a measured utility, as the measured utility is analogous to a usefulness value as described in the specification. Therefore, the limitation recites calculating a value in view of the broadest reasonable interpretation in light of the specification, which amounts to a mathematical concept. The device mentioned will be addressed at Prong 2 below;
“selecting, by the device, specific trainer clients from among the plurality of trainer clients that have a corresponding measured utility greater than a given utility threshold…”– this limitation recites performing simple arithmetic comparison by comparing values against a given threshold, which is interpreted, in view of the broadest reasonable interpretation in light of the specification, as performing an evaluation and judgment, being a mental process, based on mathematical relationships and values, which are directed to mathematical concepts. The device mentioned will be addressed at Prong 2 below.
Step 2A: Prong 2 analysis:
This judicial exception is not integrated into a practical application because it only recites these additional elements:
“training, by a device, a machine learning model using asynchronous federated learning with a plurality of trainer clients and respective data on the plurality of trainer clients” – this limitation recites using a device used for training a machine learning model on data from a given dataset, as well as for determining and selecting purposes as explained above, therefore, this device is considered a generic computer component merely used as a tool to implement the abstract ideas and the training of the model. Furthermore, the training of the model is recited at a high level of generality, therefore, it amounts to applying the judicial exception to the field of use of machine learning for training a model to perform a task (see MPEP 2106.05(h));
“…to use for training the machine learning model using asynchronous federated learning” – this limitation recites using the trainer clients selected for asynchronous federated learning model training, which under broadest reasonable interpretation in light of the specification, amounts to applying the judicial exception to the field of use of machine learning for training a federated learning model to perform a task (see MPEP 2106.05(h)).
Accordingly, these additional elements do not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. The claims are directed to an abstract idea.
Step 2B analysis:
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements recited at Claims 1, 13 and 20 above amounts to a generic computer component merely used as a tool to implement the abstract ideas and the training of the model, and applying the judicial exception to a field of use.
Viewed as a whole, these additional claim element(s) do not provide meaningful limitation(s) to transform the abstract idea into a patent eligible application of the abstract idea such that the claim(s) amounts to significantly more than the abstract idea itself. Therefore, the claim(s) are rejected under 35 U.S.C. 101 as being directed to non-statutory subject matter.
Claim 2: this limitation clarifies how the selection process uses number comparisons to determine exclusions. This is directed to the abstract idea of mathematical concept i.e., arithmetic comparison. In addition, it recites using the abstract idea for training asynchronous federated learning, which under broadest reasonable interpretation in light of the specification, amounts to applying the judicial exception to the field of use of training a federated learning model to perform a task (see MPEP 2106.05(h)).
Claim 3: this limitation clarifies how the selection process is achieved based on datasets being replaced, which amounts to the mental process as described above.
Claim 4: this limitation clarifies the selection process using number comparisons to determine priority. This is directed to the abstract idea of mathematical concept i.e., arithmetic comparison to determine the highest value in a ranking.
Claim 5: this limitation clarifies how the measured utility is calculated based on a loss metric value. This is directed to the abstract idea of mathematical concept i.e., calculating a value based on another value.
Claim 6: this limitation clarifies how the measured utility is calculated based on a loss metric value. This is directed to the abstract idea of mathematical concept i.e., calculating a value based on another value.
Claim 7: this limitation clarifies that the utility threshold given is a fixed value. This is directed to the abstract idea of mathematical concept i.e., a number.
Claim 8: this limitation clarifies that the utility threshold is a varying value. This is directed to the abstract idea of mathematical concept i.e., calculating a value from a function or formula.
Claim 9: this limitation adds calculating a statistical value of usefulness for the overall group of trainer clients. This is directed to the abstract idea of mathematical concept i.e., calculating a value from another value. Further, it clarifies how the dynamic utility threshold is calculated by the statistical value above. This is directed to the abstract idea of mathematical concept i.e., calculating a value from another value. Further, it recites using the calculation step for training a model, which amounts to applying the judicial exception to the field of use of training a machine learning model to perform a task (see MPEP 2106.05(h)).
Claim 10: this limitation clarifies how the statistical value is calculated. This is directed to the abstract idea of mathematical concept i.e., calculating a value from a function or formula.
Claim 11: this limitation clarifies how the process of calculating the utility threshold is changed. This is directed to the abstract idea of a mental process i.e., a person switching their approach to a problem when results are poor.
Claim 12: this limitation clarifies that the utility threshold given is a fixed value, and changes due to results from the training. This is directed to the abstract idea of a mathematical process i.e., calculating a value based on data.
Claim 13: this claim is rejected under the same rationale as Claim 1, mutatis mutandis.
Claim 14: this claim is rejected under the same rationale as Claim 2, mutatis mutandis.
Claim 15: this claim is rejected under the same rationale as Claim 3, mutatis mutandis.
Claim 16: this claim is rejected under the same rationale as Claim 4, mutatis mutandis.
Claim 17: this claim is rejected under the same rationale as Claim 5, mutatis mutandis.
Claim 18: this claim is rejected under the same rationale as Claim 8, mutatis mutandis.
Claim 19: this claim is rejected under the same rationale as Claim 9, mutatis mutandis.
Claim 20: this claim is rejected under the same rationale as Claim 1, mutatis mutandis.
Claim Rejections - 35 USC § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 1-8, 13-18, and 20 are rejected under 35 U.S.C. 102 (a)(1) and 102 (a)(2) as being anticipated by Anand et al (US Pub. No. US 20230394320 A1- hereinafter Anand).
Referring to Claim 1, Anand teaches a method, comprising:
training, by a device, a machine learning model using asynchronous federated learning with a plurality of trainer clients and respective data on the plurality of trainer clients (see Anand at [Abstract]: “A federated model is trained on respective local training datasets of respective multiple edge devices”. Further, see [0073] “The federated learning can also be asynchronous, e.g., split learning may be used”. Examiner interprets the edge devices as the claimed ‘trainer clients’ and their respective training datasets to the claimed ‘respective data’, since they are used to train an asynchronous federated learning model);
determining, by the device, a measured utility of the respective data on each of the plurality of trainer clients (see Anand at [0091]: “Filtering 260 may be performed in order to select a subset of training inputs for which training 270 is most effective, e.g., for which a greatest improvement in performance of the current federated model 270 is expected.” Further, see [0093]: “As another example, filtering 260 may be based on a confidence score of the current federated model 290 for the training input 211-212.” Examiner interprets the confidence score to be equivalent to the claimed “measured utility”. The confidence score is used in the same intention as the current application as based on this score, they determine the training to be effective, which is equivalent to ‘utility’); and
selecting, by the device, specific trainer clients from among the plurality of trainer clients that have a corresponding measured utility greater than a given utility threshold to use for training the machine learning model using asynchronous federated learning (see Anand at [0093]: “For example, if the confidence score does not exceed a threshold, the training input 211-212 may be included in the subset 241-242 of filtered training inputs.” Further, see [0096]: “A training input may be included if its loss contribution exceeds a given threshold, or belongs to the top-K losses for a given K, for example. It is noted that using a loss function typically implies that model inputs for which the determined model output does not match the training model output are included, and may also imply that training model inputs for which the confidence score does not exceed a threshold are included.” Examiner interprets the confidence score is directly associated with the loss function value, as is the claimed “measured utility.” Low confidence is equivalent to high loss, as is high utility measured. Therefore, the selected datasets, or “trainer clients”, are prioritized as useful to the federated learning model).
Referring to Claim 2, Anand teaches the method of claim 1, wherein selecting the specific trainer clients specifically blocks certain trainer clients from among the plurality of trainer clients with corresponding measured utilities less than the given utility threshold from being used for training the machine learning model using asynchronous federated learning (see Anand at [0008]: “Training inputs for which the current federated model provides the correct output with high confidence may be left out of the subset, however.” Examiner interprets that each training dataset is not included, or “blocked”, when its confidence score is high. When a confidence score is too high, it has too low of a loss value, and thus is not useful, similar to how low measured utility has low loss value and is also not considered useful).
Referring to Claim 3, Anand teaches the method of claim 1, wherein selecting the specific trainer clients is in response to previous trainer clients finishing training to replace the previous trainer clients to meet a concurrency of the asynchronous federated learning (see Anand at [0020]: “In an embodiment, a training input may be included in the subset of items to be trained on, that was not included in this subset in a previous iteration. Thus, training inputs need not be discarded from the set of inputs to use forever. For example, in one iteration the current model may work well for a certain training input, whereas in a later iteration, e.g., due to model updates from another edge device, the model may work less well, based on which the training input may be selected again. In particular, determining the subset of filtered training inputs may be performed in such a way that it does not depend on what subset of training inputs was selected in previous iterations, e.g., the whole local dataset may be filtered.” Examiner interprets each iteration as a separate selection process called “filtering.” When filtering, a new dataset is included to replace older, less useful, datasets);
Referring to Claim 4, Anand teaches the method of claim 1, wherein selecting comprises:
prioritizing, from within the specific trainer clients, a particular trainer client with a highest corresponding measured utility as compared to other trainer clients of the specific trainer clients (see Anand at [0097]: “Adaptive filtering can be performed, e.g. … by selecting the top-N contributors to the loss function for the current federated model, or a combination of these strategies”. Examiner interprets selecting the top N contributors as prioritizing the ones having a ranking list from 1-5, 1 being the highest corresponding measured utility).
Referring to Claim 5, Anand teaches the method of claim 1, further comprising:
measuring the measured utility of the respective data on each of the plurality of trainer clients by acquiring a loss metric from a training process by each of the plurality of trainer clients (see Anand at [0096]: “A training input may be included if its loss contribution exceeds a given threshold, or belongs to the top-K losses for a given K, for example. It is noted that using a loss function typically implies that model inputs for which the determined model output does not match the training model output are included, and may also imply that training model inputs for which the confidence score does not exceed a threshold are included.” Examiner interprets the calculation of confidence score is based on the loss function value, as is the claimed “measured utility.” Low confidence is equivalent to high loss, as is high utility measured. Therefore, the calculation of each, confidence score and measured utility, is based on a loss metric of each of its respective datasets).
Referring to Claim 6, Anand teaches the method of claim 1, wherein the measured utility of the respective data on each of the plurality of trainer clients comprises a training-based loss metric (see Anand at [0096]: “A training input may be included if its loss contribution exceeds a given threshold, or belongs to the top-K losses for a given K, for example. It is noted that using a loss function typically implies that model inputs for which the determined model output does not match the training model output are included, and may also imply that training model inputs for which the confidence score does not exceed a threshold are included.” Examiner interprets the calculation of confidence score is based on the loss function value, as is the claimed “measured utility.” Low confidence is equivalent to high loss, as is high utility measured. Therefore, the calculation of each, confidence score and measured utility, is based on a training loss metric of each of its respective datasets.).
Referring to Claim 7, Anand teaches the method of claim 1, wherein the given utility threshold is static (see Anand at [0093]: “For confidence scores on a scale from 0, no confidence; to 1, full confidence, the threshold can be at most or at least 0.3 or at most or at least 0.7, for example.” Examiner interprets the confidence, or “utility,” threshold to be fixed numbers, as it exemplifies 0.3 or 0.7, which are thus, static).
Referring to Claim 8, Anand teaches the method of claim 1, further comprising: adjusting the given utility threshold dynamically (see Anand at [0095]: “Instead of using a fixed threshold for all matching model inputs, it is also possible to use a threshold that depends on a degree of matching, e.g., a difference between the training model output and the determined model output, for example.” Examiner interprets the threshold to be based on a “degree of matching,” which varies from dataset to dataset. Thus, the threshold is dynamic).
Referring to Claim 13, this claim is rejected under the same rationale as Claim 1, mutatis mutandis.
Referring to Claim 14, this claim is rejected under the same rationale as Claim 2, mutatis mutandis.
Referring to Claim 15, this claim is rejected under the same rationale as Claim 3, mutatis mutandis.
Referring to Claim 16, this claim is rejected under the same rationale as Claim 4, mutatis mutandis.
Referring to Claim 17, this claim is rejected under the same rationale as Claim 5, mutatis mutandis.
Referring to Claim 18, this claim is rejected under the same rationale as Claim 8, mutatis mutandis.
Referring to Claim 20, this claim is rejected under the same rationale as Claim 1, mutatis mutandis.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 9-10, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Anand in view of Lumezanu et al (US Pub. No. 2019/0098050- hereinafter Lumezanu).
Referring to Claim 9, Anand teaches the method of claim 8, further comprising:
calculating a statistical utility value of the plurality of trainer clients based on measured utilities of the respective data on current trainer clients currently selected for training the machine learning model (see Anand at [0089]: “As a result of model application 250, apart from model outputs 221-222, also confidence scores (also known as classification scores; not shown in the figure) for the respective model outputs may be obtained, e.g., a confidence score may indicate a probability that the model output is correct, e.g., a probability that a classification or other discrete model output is correct, or a confidence interval for a continuous model output, etc.” Examiner interprets a confidence score that indicates a probability that the model output is correct as a statistical calculation of confidence in a given dataset. In this case, the confidence interval is based on a series of outputs by each dataset, calculating an estimated range of confidence for the series of outputs, which is analogous to calculating a statistical utility value).
However, Anand fails to teach:
adjusting the given utility threshold dynamically based on the statistical utility value.
Lumezanu teaches, in an analogous system, adjusting the given utility threshold dynamically based on the statistical utility value (see Lumezanu at [0034-0035]: “Block 408 therefore groups the training data by target and computes statistics (including, e.g., mean, median, 75.sup.th percentile, etc.) for each group. Testing is then performed using a dynamic similarity threshold for each target and the threshold is updated after each packet. The dynamic similarity threshold may be determined as, for example, the mean, median, or 75.sup.th percentile for training errors at the target. When testing, when a packet is found to be legitimate (e.g., part of a valid connection), then the mean, median, or 75.sup.th percentile value can be recalculated across all legitimate packets seen so far, including those used in training and those seen during testing”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Anand with the above teachings of Lumezanu by calculating a statistical utility value of the plurality of trainer clients based on measured utilities of the respective data, as taught by Anand, and dynamically adjusting a utility threshold, as taught by Lumezanu. The modification would have been obvious because one of ordinary skill in the art would be motivated to test the training data for errors using a dynamic threshold based on the computed statistics of the training data (as suggested by Lumezanu at 0034-0035).
Referring to Claim 10, the combination of Anand and Lumezanu teaches the method of claim 9, further comprising:
computing the statistical utility value based on one or more of: a mean, a median, a quartile, or a percentile (see Lumezanu at [0034-0035]: “Block 408 therefore groups the training data by target and computes statistics (including, e.g., mean, median, 75.sup.th percentile, etc.) for each group. Testing is then performed using a dynamic similarity threshold for each target and the threshold is updated after each packet. The dynamic similarity threshold may be determined as, for example, the mean, median, or 75.sup.th percentile for training errors at the target. When testing, when a packet is found to be legitimate (e.g., part of a valid connection), then the mean, median, or 75.sup.th percentile value can be recalculated across all legitimate packets seen so far, including those used in training and those seen during testing”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Anand with the above teachings of Lumezanu by calculating a statistical utility value of the plurality of trainer clients based on measured utilities of the respective data, as taught by Anand, and dynamically adjusting a utility threshold, as taught by Lumezanu. The modification would have been obvious because one of ordinary skill in the art would be motivated to dynamically adjust the training data based on statistics to determine legitimate or utile training data during the testing phase (as suggested by Lumezanu at 0034-0035).
Referring to Claim 19, this claim is rejected under the same rationale as Claim 9, mutatis mutandis.
Claims 11-12 are rejected under 35 U.S.C. 103 as being unpatentable over Anand in view of Guo et al (CN 115618963 A- hereinafter Guo).
Referring to Claim 11, Anand teaches the method of claim 8, further comprising:
changing methodologies for adjusting the given utility threshold during continued training of the machine learning model using asynchronous federated learning (see Anand at [0073]: “The federated learning can also be asynchronous, e.g., split learning may be used. In that sense, the training being performed iteratively merely means that the aggregation device performs repeated updates to the current federated model and that an edge device repeatedly receives a current model and determines a model update for it”. Further, at [0095]: “Instead of using a fixed threshold for all matching model inputs, it is also possible to use a threshold that depends on a degree of matching, e.g., a difference between the training model output and the determined model output, for example”. Therefore, Examiner interprets the iterative updates using asynchronous federated learning to be equivalent to adjusting the utility threshold. However, Anand fails to explicitly teach the change of methodologies using asynchronous federated learning).
Guo, in an analogous system, teaches changing methodologies using asynchronous federated learning (see Guo at Abstract: “an asynchronous federated learning asynchronous training method based on optimization direction guidance”, “improving the effectiveness of the single wheel aggregate the guide of the client model with higher sample diversity, using the model increment asynchronous updating mechanism based on the training state and the training decision based on the model difference; improving the model updating real time and training fairness. The invention optimizes and improves the training efficiency of wireless federated learning from two directions of data heterogeneous and resource heterogeneous, and through model increment asynchronous updating mechanism and client training decision”. Therefore, Examiner interprets Guo’s use of asynchronous federated learning to optimize training in real time by taking decisions to improve the efficiency to be analogous to change methodologies for the purpose of optimizing the training of the model).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Anand with the above teachings of Guo by having a dynamic threshold for training purposes of a model, as taught by Anand, and determining certain decisions to change methodologies for optimizing the training of the model, as taught by Guo. The modification would have been obvious because one of ordinary skill in the art would be motivated to improving the model updating in real time and training fairness, thereby optimizing and improving training efficiency (as suggested by Guo at Abstract).
Referring to Claim 12, Anand teaches the method of claim 1, further comprising:
starting with a static utility threshold as the given utility threshold (see Anand at [0093]: “For confidence scores on a scale from 0, no confidence; to 1, full confidence, the threshold can be at most or at least 0.3 or at most or at least 0.7, for example.” Examiner interprets the confidence, or “utility,” threshold to be fixed numbers, as it exemplifies 0.3 or 0.7, which are thus, static); and
adjusting the given utility threshold dynamically based on continued training of the machine learning model using asynchronous federated learning (see Anand at [0095]: “Instead of using a fixed threshold for all matching model inputs, it is also possible to use a threshold that depends on a degree of matching, e.g., a difference between the training model output and the determined model output, for example.” Examiner interprets the threshold to be based on a “degree of matching,” which varies from dataset to dataset. Thus, the threshold is dynamic).
However, Anand fails to explicitly teach the decision of starting with a static threshold, and then changing to a dynamic threshold using asynchronous federated learning.
Guo, in an analogous system, teaches the decision of starting with a static threshold, and then changing to a dynamic threshold using asynchronous federated learning (see Guo at Abstract: “an asynchronous federated learning asynchronous training method based on optimization direction guidance”, “improving the effectiveness of the single wheel aggregate the guide of the client model with higher sample diversity, using the model increment asynchronous updating mechanism based on the training state and the training decision based on the model difference; improving the model updating real time and training fairness. The invention optimizes and improves the training efficiency of wireless federated learning from two directions of data heterogeneous and resource heterogeneous, and through model increment asynchronous updating mechanism and client training decision”. Therefore, Examiner interprets Guo’s use of asynchronous federated learning to optimize training in real time by taking decisions to improve the efficiency to be analogous to change methodologies such as starting with a fixed threshold and then taking the decision in real time to adjust it for purposes of optimizing the training of the model).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Anand with the above teachings of Guo by having a dynamic threshold for training purposes of a model, as taught by Anand, and determining certain decisions to change methodologies such as starting with a fixed threshold and then adjusting it for optimizing the training of the model, as taught by Guo. The modification would have been obvious because one of ordinary skill in the art would be motivated to improving the model updating in real time and training fairness, thereby optimizing and improving training efficiency (as suggested by Guo at Abstract).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to LUIS A SITIRICHE whose telephone number is (571)270-1316. The examiner can normally be reached M-F 9am-6pm.
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 Yi can be reached at (571) 270-7519. 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.
/LUIS A SITIRICHE/Primary Examiner, Art Unit 2126
Prosecution Insights