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
This Office action has been issued in response to amendment filed on 07/06/2026, Claims (8-17, 19-20 and 22), (23-24, 26-28), 32, 35 and 39 are pending. Applicants' arguments have been carefully and respectfully considered and addressed. Accordingly, this action has been made FINAL necessitated by amendment.
Claims (8-17, 19-20 and 22), (23-24, 26-28), 32, 35 and 39 are presented for examination.
Claims 1-7 are cancelled.
Response to Arguments
Regarding Applicant arguments, the arguments were fully considered and are moot in view of the new ground rejection wherein Nadiger in view Glodek and further in view of Sutherland for teaching the amended claims. Wherein the reference Sutherland is part of the cited references. Wherein Sutherland teaches dataset (pages. 6-7, Introduction section, page. 14, section 2.2, page. 15, page. 17, section 2.4, page. 30, page. 31, page. 44, section 6.1 wherein Sutherland describes applying machine learning on collected objects or datasets to define vectors model based on the concentration of several chemicals and applying a classifier to a region and that the classifier marks as relevant. Sutherland applies multiple density estimator for mixture distributions and observes samples dataset from those distributions. Wherein the samples dataset represents local datasets as they are collected objects. Section 6.1 of page. 44 describes how Sutherland uses samples from an image-level distribution of local features, and classified images based on those sets of features).
Allowable Subject Matter
Claims (8-17, 19-20, 22, 32) and 35 are directed to allowable subject matter if the cited rejection and objections are addressed.
PNG
media_image1.png
326
720
media_image1.png
Greyscale
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 of this title, 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 23-24, 26, 28, 32, 39 are rejected under Nadiger in view Glodek and further in view of Sutherland.
Regarding claim 23, Nadiger teaches A method for using federated learning to develop a machine learning model, the method, performed by a distributed node associated with a local dataset, comprising: receiving a seed version of a machine-learning model, wherein the seed version of the machine-learning model has been developed using a machine-learning algorithm (Nadiger, pg. 126, "Referring to figure 5, the server sends a base global model to all the clients to start the NPC learning. [ ... ] The base global model was built offline using the method described in [6].") developing a node version of the machine-learning model, based on the seed version of the machine-learning model and the associated local dataset (Nadiger, Figure 3, "Local ML training", "Local Training Data")
Nadiger does not teach generating a distribution density estimate, the distribution density estimate being an estimate of a distribution density of the…dataset;
However in analogous art of federated learning method, Glodek teaches generating a distribution density estimate, the distribution density estimate being an estimate of a distribution density of the…dataset; communicating the generated distribution density estimate to a management function’ communicating a representation of the node version of the machine-learning model to the management function (page. 2, sections 2 and 2.1 wherein Glodek describes a probability density function that provides a density estimate of data distribution and incorporates a density function that is defined over continuous variables).
It would have been obvious to a person in the ordinary skill in the art before the effective filing date of the claimed invention to combine Nadiger with Glodek by incorporating the method of generating a distribution density estimate, the distribution density estimate being an estimate of a distribution density of the local dataset; communicating the generated distribution density estimate to a management function’ communicating a representation of the node version of the machine-learning model to the management function of Glodek into the method of using federated learning to develop a machine learning model, the method, performed by a distributed node associated with a local dataset, comprising: receiving a seed version of a machine-learning model, wherein the seed version of the machine-learning model has been developed using a machine-learning algorithm of Nadiger for the purpose of increase the robustness and accuracy of density estimation. (Glodek: Abstract).
Nadiger does not teach generating a distribution density estimate, the distribution density estimate being an estimate of a distribution density of the local dataset;
However in analogous art of federated learning method, Sutherland teaches generating a distribution density estimate, the distribution density estimate being an estimate of a distribution density of the local dataset (pages. 6-7, Introduction section, page. 14, section 2.2, page. 15, page. 17, section 2.4, page. 30, page. 31, page. 44, section 6.1 wherein Sutherland describes applying machine learning on collected objects or datasets to define vectors model based on the concentration of several chemicals and applying a classifier to a region and that the classifier marks as relevant. Sutherland applies multiple density estimator for mixture distributions and observes samples dataset from those distributions. Wherein the samples dataset represents local datasets as they are collected objects. Section 6.1 of page. 44 describes how Sutherland uses samples from an image-level distribution of local features, and classified images based on those sets of features).
It would have been obvious to a person in the ordinary skill in the art before the effective filing date of the claimed invention to combine Nadiger with Sutherland by incorporating the method of generating a distribution density estimate, the distribution density estimate being an estimate of a distribution density of the local dataset of Sutherland into the method of using federated learning to develop a machine learning model, the method, performed by a distributed node associated with a local dataset, comprising: receiving a seed version of a machine-learning model, wherein the seed version of the machine-learning model has been developed using a machine-learning algorithm of Nadiger Glodek for the purpose of increasing the robustness and accuracy of density estimation. (Sutherland: Abstract).
PNG
media_image2.png
542
776
media_image2.png
Greyscale
Regarding claim 24, Nadiger as modified by Glodek and Sutherland teach wherein the representation of distribution of data within the local data set dataset comprises any one of a Gaussian mixture model (GMM), a Euclidean distance, a L-2 distance, a maximum mean discrepancy (MMD), or a Jsensen-Renyi divergence (Abstract, section 2.1 on page. 2, page. 4, paragraph 1, wherein Glodek incorporates GMM and compare the accuracies of the three densities, the integral of the absolute difference to the original density are taken over the interval), the representation of distribution of data within the local data set dataset further comprises a quantity of labels per predetermined category in the local dataset (Sutherland: Page. 9, “One common method for representing distributions is the use of histograms; many distances are then simple to compute.” Examiner notes that a histogram is a count of labels per category”).
Regarding claim 26, Nadiger as modified by Glodek and Sutherland teach receiving form the management function at least one hyper parameter that is designed for a learning group to which the distributed node is assigned; and using the hyper parameter to develop a node version of the machine-learning model (page. 126, Global model regularization factor") [that is designed for a learning group to which the distributed node is assigned], "Input: [ ... ] Previous Global Model Factor" )."
Regarding claim 28, Nadiger as modified by Glodek and Sutherland teach wherein the step of communicating the representation of the node version of the machine-learning model to the management function comprises communicating the representation of the node version of the machine-learning model to a group management function of a learning group to which the distributed node is assigned, and the method further comprises receiving, from the management function, an instruction of how to communicate a representation of the node version of the machine-learning model to the management function (Nadiger, Figure 3, Client Model arrows pointed from the client nodes to the Server node for aggregation. The server, as shown in Figure 1, where each group of clients is managed),", (Nadiger, pg. 125, "The client-server interaction happens after certain 'K' number of episodes of the game on the client. In this work, K is set to be 25 episodes. Once the client sends its NPC model weights, it then waits till a new global model is available. Once the new global model is available, it updates the NPC weights as per the model replacement policy. Then, it resumes play with human-proxy player for next K episodes. The global model is generated on the server through a federation policy.")."
Regarding claim 32, Nadiger as modified by Glodek teaches A non-transitory computer readable storage medium storing a computer program comprising instructions which, when executed on at least one processor, cause the at least one processor to perform the method of claim 8 (Nadiger, Fig. 1, "Client Node", "Server Node")."
Regarding claim 39, the claim is similar in scope to claim 8 therefore the claims are rejected under similar rationale.
Claims 27 are rejected under Nadiger in view Glodek and further in view of Ghosh.
Regarding claim 27, Nadiger and Glodek do not teach wherein the distributed node is assigned to a learning group on the basis of a similarity of its representations of distribution of data to representations of distribution of data in local datasets associated with other distributed nodes.
However, in analogous art of federated learning method, Ghosh teaches wherein the distributed node is assigned to a learning group on the basis of a similarity of its representations of distribution of data to representations of distribution of data in local datasets associated with other distributed nodes ((Ghosh, pg. 4, "The second step of the modular algorithms deals with clustering the compute nodes based on their local ERMs.")."
It would have been obvious to a person in the ordinary skill in the art before the effective filing date of the claimed invention to combine Ghosh with Nadiger and Glodek by incorporating the method of wherein the distributed node is assigned to a learning group on the basis of a similarity of its representations of distribution of data to representations of distribution of data in local datasets associated with other distributed nodes of Ghosh into the method of using federated learning to develop a machine learning model, the method, performed by a distributed node associated with a local dataset, comprising: receiving a seed version of a machine-learning model, wherein the seed version of the machine-learning model has been developed using a machine-learning algorithm of Nadiger and Glodek for the purpose of enabling multiple clients to collaborate train a shared model under the coordination of an edge server (Ghosh: Abstract).
Conclusion
THIS ACTION IS MADE FINAL. 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 extension fee 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 HASSAN MRABI whose telephone number is (571)272-8875. The examiner can normally be reached on Monday-Friday, 7:30am-5pm. Alt, Friday, EST.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Viker Lamardo can be reached on 571-270-5871. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/HASSAN MRABI/Examiner, Art Unit 2144
Prosecution Insights