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 .
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed 04/27/2026 in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 03/27/2026 has been entered.
The following action is in response to the amendment and remarks 03/27/2026.
By the amendment, claims 1, 8 and 15 have been amended. Claims 5 and 12 have been canceled.
Claims 1-4, 6-11, 13-17 and 21-23 are pending and have been considered below.
Response to Arguments
Claims 5 and 12 have been canceled rendering the corresponding rejections moot.
The 35 USC 101 rejection of claims 1-4, 6-11, 13-17 and 21-23 has been withdrawn in light of the amendment and applicant’s corresponding remarks (Remarks 03/27/2026 page 6).
The 35 USC 103 rejection of claims 1-4, 6-11, 13-17 and 21-23 has been updated and maintained below in light of the amendment and Applicant’s corresponding arguments (Remarks 03/27/2026 pages 6-7).
Applicant argues, regarding the 35 USC 103 rejection over Mathur in view of Yu, that Yu fails to teach or disclose using a latent dimension vector for calculating a frequency distribution associated with the local dataset as relied on the rejection. The Examiner respectfully disagrees.
The Examiner notes that the plurality of frequency distributions calculated from data is taught by Mathur (¶91, 99-101) and not by Yu. One cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. Yu is relied on only to teach that data used for calculation of the frequency distributions may be latent dimension vectors. Yu teaches that a plurality of statistical characteristics of a record dataset may be calculated as vector data comprising vector values (¶28-29: “These various statistical characteristics of the dataset may be provided as a vector data structure comprising vector values corresponding to the statistical characteristics and provides a description of the dataset.”). Yu discloses that the data is provided as a vector data structure comprising multiple data statistics (¶27). The claim does not preclude this interpretation as the claim does not provide further information regarding differences between a latent dimension vector and the interpretation presented by Yu. Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims (Specification ¶28, 30-38). The argument is not persuasive.
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 1-3, 6-10, 15-17 and 23 are rejected under 35 U.S.C. 103 as being unpatentable over MATHUR, US 2023/0169402 A1, effective filing 06/02/2020 (previously presented), in view of YU, US 2022/0269977 A1, published 08/25/2022 (previously presented).
Regarding claim 1, MATHUR discloses a method of determining similar nodes in a federated learning environment (¶3, ¶5), executable by a processor, comprising:
receiving, by a first node in the federated learning environment, data for calculating a frequency distribution for a local dataset associated with the first node (¶93: respective data based on a local dataset of a processing node, ¶99-101: for calculating frequency distribution);
calculating, by the first node, a frequency distribution associated with the local dataset (¶91: statistical distribution of local dataset of respective processing node, ¶99);
transmitting, to an aggregator, the calculated frequency distribution of the local dataset (¶99: performed at collaborative server);
receiving, from the aggregator, a plurality of frequency distributions respectively corresponding to other nodes in the federated learning environment (¶100); and
identifying, by the first node, a similarly between the first node and subset of the other nodes based on the received plurality of frequency distributions (¶100-101), and
responsive to the first node dropping off the federated learning environment (¶102: disregard node), replacing the first node with an active node selected from the subset of other nodes (¶102, ¶109: add new node, ¶110: iterative).
MATHUR fails to explicitly disclose wherein the data for calculating a frequency distribution for a local dataset associated with the first node is a latent dimension used in calculating the frequency distribution.
YU discloses methods for performing federated machine learning updates (¶6). In particular, YU discloses generating vectorized data structures of local datasets (¶57) having one or more latent dimensions (¶27-29) used in determining frequency distribution (¶29). Therefore it would have been obvious to one having ordinary skill in the art and the teachings of MATHUR and YU before them before the effective filing of the claimed invention to combine the use of latent dimension vectors for determining frequency distributions of a local dataset, as suggested by YU, with the data for calculating the frequency distribution of the local dataset of MATHUR. One would have been motivated to make this combination to ensure robustness when aggregating an environment with plural different data distributions, as suggested by YU (¶26-27).
Regarding claim 2, MATHUR and YU disclose the method of claim 1, and MATHUR further discloses wherein identifying the similarity between the first node and the subset of the other nodes is further based on a similarity score associated with the node being above a threshold value (¶101-103).
Regarding claim 3, MATHUR and YU disclose the method of claim 2, and MATHUR further discloses wherein the similarity score corresponds to a distance between the first node and the subset of other nodes (¶103).
Regarding claim 6, MATHUR and YU disclose the method of claim 1, and YU further discloses wherein using the latent dimension vector includes:
converting each record within the local dataset to a vector having one or more latent dimensions (¶28-29, ¶57).
Regarding claim 7, MATHUR and YU disclose the method of claim 6, and YU further discloses wherein the one or more latent dimensions correspond to one or more from among roundness, sharpness, and thickness associated with the entries in the dataset (¶28-29).
Regarding claims 8-10 and 13-14, claims 8-10 and 13-14 recite limitations similar to claims 1-3 and 6-7, respectively, and are similarly rejected.
Regarding claims 15-17, claims 15-17 recite limitations similar to claims 1-3, respectively, and are similarly rejected.
Regarding claim 23, MATHUR and YU disclose the method of claim 1, and MATHUR further discloses:
sending, by the aggregator, a set of frequency distributions to the other nodes to calculate respective similarity scores in batches (¶86: share data and updates among members of the set, Fig. 1).
Claims 4 and 11 are rejected under 35 U.S.C. 103 as being unpatentable over MATHUR in view of YU and in further view of KOEHLER, US 2022/0147869 A1, published 05/12/2022 (previously presented).
Regarding claim 4, MATHUR and YU disclose the method of claim 2, and MATHUR further discloses wherein the similarity score is calculated using one or more statistical distance measurements (¶15-16: Wasserstein Distance or Maximum Mean Discrepancy).
MATHUR and YU fail to disclose wherein the statistical distance measurement is selected from a group consisting of: Kullback-Leibler divergence, Jensen-Shannon distance, and Hellinger distance.
KOEHLER discloses methods for assessing uncertainty when training models (¶9-13). In particular, KOEHLER teaches using statistical measures to determine deviations/similarity between datasets such as through any of Kullback-Leibler divergence, Jensen-Shannon distance, Hellinger distance, and Wasserstein metrics (¶38). Therefore it would have been obvious to one having ordinary skill in the art and the teachings of MATHUR, YU and KOEHLER before them before the effective filing of the claimed invention to simply substitute, using known methods, the statistical scalar measure methods of Kullback-Leibler divergence, Jensen-Shannon distance, Hellinger distance, and Wasserstein metrics interchangeably, as suggested by KOEHLER, to achieve predictable results for calculating the similarity score of MATHUR and YU. One would have been motivated to make this substitution as doing so would simply substitute one know element for another to obtain predictable results (KSR, MPEP 2143.I.B.).
Regarding claim 11, claim 11 recites limitations similar to claim 4 and is similarly rejected.
Claims 21 and 22 are rejected under 35 U.S.C. 103 as being unpatentable over MATHUR in view of YU and in further view of TOPOREK, US 2023/0316141 A1 effective filing of 03/31/2022, (previously presented).
Regarding claim 21, MATHUR and YU disclose the method of claim 1, and MATHUR further discloses:
collaboratively training machine learning model, such as a neural network, by nodes in the federated learning environment including the first node and the subset of the other nodes, the nodes using local datasets (¶82, ¶93-96, ¶100-102).
Neither, MATHUR nor YU disclose wherein the machine learning model is an autoencoder.
TOPOREK discloses methods for training machine learning models using local nodes and data distributions (¶3-4). In particular, TOPOREK discloses that the trained machine learning model can be selected from a group including neural networks (NN), long short-term memory architecture (LTSM), generative adversarial networks (GAN), and/or variational encoder-decoder networks (VAR) (¶51-52). Therefore it would have been obvious to one having ordinary skill in the art and the teachings of MATHUR, YU and TOPOREK before them before the effective filing of the claimed invention to simply substitute, using known methods, the variational autoencoder, neural network, generative adversarial network or long short-term memory network interchangeably, as suggested by TOPOREK, with the neural network machine learning model of MATHUR and YU. One would have been motivated to make this substitution as doing so would simply substitute one know element for another to obtain predictable results (KSR, MPEP 2143.I.B.), and further suggested by TOPOREK (¶51).
Regarding claim 22, MATHUR, YU and TOPOREK disclose the method of claim 21, and MATHUR further discloses wherein the latent dimension vector is received by the first node from the collaboratively trained autoencoder (¶99-100).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Zoldi et al., US 2022/0188644 A1, latent-space misalignment measure of responsible AI for machine learning models
Wang, Wei, et al. "LDGAN: Latent determined ensemble helps removing IID data assumption and cross-node sampling in distributed GANs." 2022 26th International Conference on Pattern Recognition (ICPR). IEEE, 2022.
Gewers, Felipe L., et al. "Principal component analysis: A natural approach to data exploration." ACM Computing Surveys (CSUR) 54.4 (2021): 1-34.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ANDREW L TANK whose telephone number is (571)270-1692. The examiner can normally be reached Monday-Thursday 9a-6p.
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, Matthew Ell can be reached at 571-270-3264. 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.
/ANDREW L TANK/Primary Examiner, Art Unit 2141
Prosecution Insights