DETAILED ACTION
1. 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
2. Receipt of Applicant’s Preliminary Amendment filed on 08/27/2024 is acknowledged. The preliminary amendment includes the amending of the specification, the amending of claims 1-9, and the addition of claim 10.
Information Disclosure Statement
3. The information disclosure statement (IDS) submitted on 08/27/2024 has been received, entered into the record, and considered. The submission is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Rejections - 35 USC § 101
4. 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.
5. Claims (4-6 & 10) and (8) are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Under the 2019 PEG, when considering subject matter eligibility under 35 U.S.C. § 101, it must be determined whether the claim is directed to one of the four statutory categories of invention, i.e., process, machine, manufacture, or composition of matter (step 1). If the claim does fall within one of the statutory categories, it must then be determined whether the claim is directed to a judicial exception (i.e., law of nature, natural phenomenon, and abstract idea) (step 2A prong 1), and if so, it must additionally be determined whether the claim is integrated into a practical application (step 2A prong 2). If an abstract idea is present in the claim without integration into a practical application, any element or combination of elements in the claim must be sufficient to ensure that the claim amounts to significantly more than the abstract idea itself (step 2B).
In the instant case, claims (4-6 & 10) and (8) are directed to a device and method respectively. Thus, each of the claims falls within one of the four statutory categories. However, the claims also fall within the judicial exception of an abstract idea.
Under Step 2A Prong 1, the test is to identify whether the claims recite a judicial exception. The examiner notes that the claimed invention recites an abstract idea in that the instant application recites a mental processes, specifically providing information.
The examiner further notes that claims (4-6 & 10) and (8) recite a device and method for providing information which is similar to themes defined above of method of mental processes such as performing the providing information, and is similar to the abstract idea identified in the 2019 PEG in grouping “c” in that the claims recite certain methods of mental processes such as performing the providing of information. The limitations, substantially comprising the body of the claim, recite a process of providing information. The examiner notes that the claimed invention provides information. Because the limitations above closely follow the steps in providing information, and the steps of the claims involve mental processes, the claim recites an abstract idea consistent with the “mental processes” grouping set forth in the 2019 PEG.
Claim 4:
A secure federated learning device comprising processing circuitry configured to: processing circuitry configured to: obtain confidential information of information that specifies a plurality of worker models from a plurality of model learning devices;
obtain confidential information of information that specifies an aggregate model that is an aggregation of the plurality of worker models without obtaining the plurality of worker models through secure computation using the confidential information of the information that specifies the plurality of worker models; and
provide the information that specifies the aggregate model or the confidential information of the information that specifies the aggregate model to the plurality of model learning devices.
These limitations, as drafted, is an apparatus that, under its broadest reasonable interpretation, covers the performance of mental processes specifically providing information. Providing information has long before the modern computer was invented, and continues to be predominantly a product of human endeavor. The instant application recites providing information. Moreover, the executing a secure computation to obtain confidential information of information that specifies an aggregate model can be performed by a human via their mind and/or pen & paper. Because the limitations above closely follow the steps of providing information, and the steps involved human judgments, observations and evaluations that can be practically or reasonably performed in the human mind and/or pen & paper, the claim recites an abstract idea consistent with the “mental process” grouping set forth in the 2019 PEG.
The mere nominal recitation of generic computing components such as a “secure federated learning device”, “processing circuitry”, and “model learning devices” do not take the claim out of certain methods of mental processes grouping. Therefore, the limitation recites an abstract idea.
If the claims recite the judicial exception of an abstract idea, it must then be determined under Step 2A Prong 2 whether the judicial exception is integrated into a practical application. The Examiner notes that considerations under Step 2A Prong 2 comprise most the consideration previously evaluated in the context of Step 2B. The Examiner submits that the considerations discussed previously determined that the claim does not recite “significantly more” at Step 2B would be evaluated the same under Step 2A Prong 1 and result in the determination that the claim does not integrate the abstract idea into a practical application.
The instant application fails to integrate the judicial exception into a practical application because the instant application merely recites words “apply it” (or an equivalent) with the judicial exception or merely includes instructions to implement an abstract idea. The instant application is directed to an apparatus instructing the reader to implement the identified apparatus of mental processes of providing information. The elements of the claim do not themselves amount to an improvement to the computer, to a technology or another technical field. Moreover, the obtaining of confidential information that specifies a plurality of worker models is a data gathering operation that is an insignificant data gathering operation that does not integrate the abstract idea into a practical application. Furthermore, the receiving of a label description is a data gathering operation that is an insignificant data gathering operation that does not integrate the abstract idea into a practical application. Additionally, the providing of information to model learning devices is a data transmission operation that is an insignificant data transmission operation that does not integrate the abstract idea into a practical application.
Here, the claim elements entirely comprise the abstract idea, leaving little if any aspects of the claim for further consideration under Step 2A Prong 2. In short, the claims have failed to integrate a practical application (see at least 84 Fed. Reg. (4) at 55). Under the 2019 PEG, this supports the conclusion that the claim is directed to an abstract idea, and the analysis proceeds to Step 2B.
While many considerations in Step 2A need not be reevaluated in Step 2B because the outcome will be the same. Here, on the basis of the additional elements other than the abstract idea, considered individually and in combination as discussed above, the Examiner respectfully submits that the claim 4 does not contain any additional elements that individually or as an ordered combination amount to an inventive concept and the claims are ineligible.
With respect to the dependent claims do not recite anything that is found to render the abstract idea as being transformed into a patent eligible invention. The dependent claims are merely reciting further embellishments of the abstract idea and do not claim anything that amounts to significantly more than the abstract idea itself.
With respect to the dependent claims, they have been considered and are not found to be reciting anything that amounts to being significantly more than the abstract idea. Claims 5-6 and 10 are directed to further embellishments of the central theme of the abstract idea in that the claims are directed to further embellishments of the providing information of the steps of claim 4 and do not amount to significantly more.
Specifically, claim 5 recites the determination of whether or not confidential information has been obtained in order to perform a subsequent computation which can be performed by a human via their mind and/or pen & paper and does not amount to significantly more.
Moreover, claim 6 recites the obtaining of synchronization information which is a data gathering operation that is an insignificant data gathering operation that does not integrate the abstract idea into a practical application. Moreover, the use of synchronization information to determine whether or not confidential information has been obtained in order to perform a subsequent computation which can be performed by a human via their mind and/or pen & paper and does not amount to significantly more.
Furthermore, claim 10 recites a non-transitory computer-readable recording medium storing a program for causing a computer to function as a secure federated learning device. The mere nominal recitation of generic computing components such as a “non-transitory computer-readable recording medium”, “computer” and “secure federated learning device” do not take the claim out of certain methods of mental processes grouping. Therefore, the limitation recites an abstract idea.
Claim 8:
A secure federated learning method using a secure federated learning device, the method comprising: obtaining confidential information of information that specifies a plurality of worker models from a plurality of model learning devices;
obtaining confidential information of information that specifies an aggregate model that is an aggregation of the plurality of worker models without obtaining the plurality of worker models through secure computation using the confidential information of the information that specifies the plurality of worker models; and
providing the information that specifies the aggregate model or the confidential information of the information that specifies the aggregate model to the plurality of model learning devices.
These limitations, as drafted, is an apparatus that, under its broadest reasonable interpretation, covers the performance of mental processes specifically providing information. Providing information has long before the modern computer was invented, and continues to be predominantly a product of human endeavor. The instant application recites providing information. Moreover, the executing a secure computation to obtain confidential information of information that specifies an aggregate model can be performed by a human via their mind and/or pen & paper. Because the limitations above closely follow the steps of providing information, and the steps involved human judgments, observations and evaluations that can be practically or reasonably performed in the human mind and/or pen & paper, the claim recites an abstract idea consistent with the “mental process” grouping set forth in the 2019 PEG.
The mere nominal recitation of generic computing components such as a “secure federated learning device” and “model learning devices” do not take the claim out of certain methods of mental processes grouping. Therefore, the limitation recites an abstract idea.
If the claims recite the judicial exception of an abstract idea, it must then be determined under Step 2A Prong 2 whether the judicial exception is integrated into a practical application. The Examiner notes that considerations under Step 2A Prong 2 comprise most the consideration previously evaluated in the context of Step 2B. The Examiner submits that the considerations discussed previously determined that the claim does not recite “significantly more” at Step 2B would be evaluated the same under Step 2A Prong 1 and result in the determination that the claim does not integrate the abstract idea into a practical application.
The instant application fails to integrate the judicial exception into a practical application because the instant application merely recites words “apply it” (or an equivalent) with the judicial exception or merely includes instructions to implement an abstract idea. The instant application is directed to an apparatus instructing the reader to implement the identified apparatus of mental processes of providing information. The elements of the claim do not themselves amount to an improvement to the computer, to a technology or another technical field. Moreover, the obtaining of confidential information that specifies a plurality of worker models is a data gathering operation that is an insignificant data gathering operation that does not integrate the abstract idea into a practical application. Furthermore, the receiving of a label description is a data gathering operation that is an insignificant data gathering operation that does not integrate the abstract idea into a practical application. Additionally, the providing of information to model learning devices is a data transmission operation that is an insignificant data transmission operation that does not integrate the abstract idea into a practical application.
Here, the claim elements entirely comprise the abstract idea, leaving little if any aspects of the claim for further consideration under Step 2A Prong 2. In short, the claims have failed to integrate a practical application (see at least 84 Fed. Reg. (4) at 55). Under the 2019 PEG, this supports the conclusion that the claim is directed to an abstract idea, and the analysis proceeds to Step 2B.
While many considerations in Step 2A need not be reevaluated in Step 2B because the outcome will be the same. Here, on the basis of the additional elements other than the abstract idea, considered individually and in combination as discussed above, the Examiner respectfully submits that the claim 4 does not contain any additional elements that individually or as an ordered combination amount to an inventive concept and the claims are ineligible.
Claim Rejections - 35 USC § 102
6. 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.
7. 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.
8. Claims 1, 4, and 7-10 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Stripelis et al. (Article entitled “Secure Neuroimaging Analysis using Federated Learning with Homomorphic Encryption”, dated 2021).
9. Regarding claim 1, Stripelis teaches a model learning device comprising:
A) a storage configured to store local learning data (Page 5, Section 5, Figure 1); and
B) processing circuitry configured to: obtain information that specifies an aggregate model or confidential information of the information that specifies the aggregate model from a secure federated learning device (Page 5, Section 5);
C) update the aggregate model through machine learning using the local learning data to obtain a worker model (Page 5, Section 5);
D) obtain confidential information of information that specifies the worker model (Page 5, Section 5); and
E) provide the confidential information of the information that specifies the worker model to the secure federated learning device (Page 5, Section 5).
The examiner notes that Stripelis teaches “a storage configured to store local learning data” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model” (Page 5, Section 5). The examiner further notes that Figure 1 depicts learners (i.e. clients) that each store training data. The examiner further notes that Stripelis teaches “processing circuitry configured to: obtain information that specifies an aggregate model or confidential information of the information that specifies the aggregate model from a secure federated learning device” as “the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key” (Page 5, Section 5). The examiner further notes that the transmission of the new community model to all learners (i.e. clients) teaches the specification of the aggregate model. The examiner further notes that Stripelis teaches “update the aggregate model through machine learning using the local learning data to obtain a worker model” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model…the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the decryption (via a private key) at each learner of the sent community model (i.e. aggregate model) to each results in that sent decrypted community model (i.e. aggregate model) being updated (resulting in a “worker model” being obtained) at each learner in an iterative manner. The examiner further notes that Stripelis teaches “obtain confidential information of information that specifies the worker model” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the encryption of the locally trained model (i.e. worker model) teaches the claimed obtaining. The examiner further notes that Stripelis teaches “provide the confidential information of the information that specifies the worker model to the secure federated learning device” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the transmission of the encryption of the locally trained model (i.e. worker model) to the controller teaches the claimed providing.
Regarding claim 4, Stripelis teaches a secure federated learning device comprising:
A) processing circuitry configured to: obtain confidential information of information that specifies a plurality of worker models from a plurality of model learning devices (Page 5, Section 5);
B) obtain confidential information of information that specifies an aggregate model that is an aggregation of the plurality of worker models without obtaining the plurality of worker models through secure computation using the confidential information of the information that specifies the plurality of worker models (Page 5, Section 5); and
C) provide the information that specifies the aggregate model or the confidential information of the information that specifies the aggregate model to the plurality of model learning devices (Page 5, Section 5).
The examiner notes that Stripelis teaches “processing circuitry configured to: obtain confidential information of information that specifies a plurality of worker models from a plurality of model learning devices” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the transmission of an encrypted local model from each learner to a controller teaches the claimed obtaining. The examiner further notes that Stripelis teaches “obtain confidential information of information that specifies an aggregate model that is an aggregation of the plurality of worker models without obtaining the plurality of worker models through secure computation using the confidential information of the information that specifies the plurality of worker models” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the controller computing a new encrypted model (i.e. aggregate model) without ever decrypting the local models (i.e. worker models) teaches the claimed obtaining. The examiner further notes that Stripelis teaches “provide the information that specifies the aggregate model or the confidential information of the information that specifies the aggregate model to the plurality of model learning devices” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the transmission of the computed new community model (i.e. aggregate model) from the controller to each learner teaches the claimed providing.
Regarding claim 7, Stripelis teaches a model learning method comprising:
A) obtaining information that specifies an aggregate model or confidential information of the information that specifies the aggregate model from a secure federated learning device (Page 5, Section 5);
B) updating the aggregate model through machine learning using local learning data stored in a storage to obtain a worker model (Page 5, Section 5);
C) obtain confidential information of information that specifies the worker model (Page 5, Section 5); and
D) providing the confidential information of the information that specifies the worker model to the secure federated learning device (Page 5, Section 5).
The examiner notes that Stripelis teaches “obtaining information that specifies an aggregate model or confidential information of the information that specifies the aggregate model from a secure federated learning device” as “the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key” (Page 5, Section 5). The examiner further notes that the transmission of the new community model to all learners (i.e. clients) teaches the specification of the aggregate model. The examiner further notes that Stripelis teaches “updating the aggregate model through machine learning using local learning data stored in a storage to obtain a worker model” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model…the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the decryption (via a private key) at each learner of the sent community model to each results in that sent decrypted community model being updated (resulting in a “worker model” for each learner) at each learner in an iterative manner. The examiner further notes that Stripelis teaches “obtaining confidential information of information that specifies the worker model” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the encryption of the locally trained model (i.e. worker model) teaches the claimed obtaining. The examiner further notes that Stripelis teaches “providing the confidential information of the information that specifies the worker model to the secure federated learning device” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the transmission of the encryption of the locally trained model (i.e. worker model) to the controller teaches the claimed providing.
Regarding claim 8, Stripelis teaches a secure federated learning method comprising:
A) obtaining confidential information of information that specifies a plurality of worker models from a plurality of model learning devices (Page 5, Section 5);
B) obtaining confidential information of information that specifies an aggregate model that is an aggregation of the plurality of worker models without obtaining the plurality of worker models through secure computation using the confidential information of the information that specifies the plurality of worker models (Page 5, Section 5); and
C) providing the information that specifies the aggregate model or the confidential information of the information that specifies the aggregate model to the plurality of model learning devices (Page 5, Section 5).
The examiner notes that Stripelis teaches “obtaining confidential information of information that specifies a plurality of worker models from a plurality of model learning devices” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the transmission of an encrypted local model from each learner to a controller teaches the claimed obtaining. The examiner further notes that Stripelis teaches “obtaining confidential information of information that specifies an aggregate model that is an aggregation of the plurality of worker models without obtaining the plurality of worker models through secure computation using the confidential information of the information that specifies the plurality of worker models” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the controller computing a new encrypted model (i.e. aggregate model) without every decrypting the local models (i.e. worker models) teaches the claimed obtaining. The examiner further notes that Stripelis teaches “providing the information that specifies the aggregate model or the confidential information of the information that specifies the aggregate model to the plurality of model learning devices” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the transmission of the computed new community model (i.e. aggregate model) from the controller to each learner teaches the claimed providing.
Regarding claim 9, Stripelis further teaches a model learning device comprising:
A) A non-transitory computer-readable recording medium storing a program for causing a computer to function as the model learning device according to claim 1 (Page 5, Section 5, Figure 1).
The examiner further notes that Stripelis teaches “A non-transitory computer-readable recording medium storing a program for causing a computer to function as the model learning device according to claim 1” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the iterative federative learning depicted in Figure 1 includes a non-transitory computer-readable recording medium.
Regarding claim 10, Stripelis further teaches a secure federated learning device comprising:
A) A non-transitory computer-readable recording medium storing a program for causing a computer to function as the secure federated learning device according to claim 4 (Page 5, Section 5, Figure 1).
The examiner further notes that Stripelis teaches “A non-transitory computer-readable recording medium storing a program for causing a computer to function as the secure federated learning device according to claim 4” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the iterative federative learning depicted in Figure 1 includes a non-transitory computer-readable recording medium.
Claim Rejections - 35 USC § 103
10. 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.
11. 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.
12. 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.
13. Claim 2 is rejected under 35 U.S.C. 103 as being unpatentable over Stripelis et al. (Article entitled “Secure Neuroimaging Analysis using Federated Learning with Homomorphic Encryption”, dated 2021) as applied to claims 1, 4, and 7-10 above, in view of Zhang et al. (U.S. PGPUB 2024/0054354), and further in view of Park et al. (U.S. PGPUB 2022/0383197).
14. Regarding claim 2, Stripelis does not explicitly teach a model learning device comprising:
A) wherein the processing circuitry is further: configured to determine whether or not it is necessary to update the aggregate model to newly obtain the worker model;
B) wherein when it is determined that it is not necessary to update the aggregate model to newly obtain the worker model, without updating the aggregate model to newly obtain the worker model;
C) when it is determined that it is necessary to update the aggregate model to newly obtain the worker model, the processing circuitry updates the aggregate model through machine learning using the local learning data to obtain the worker model.
Zhang, however, teaches “wherein the processing circuitry is further: configured to determine whether or not it is necessary to update the aggregate model to newly obtain the worker model” as “The client may request, by using the request message, to participate in federated learning from the first server, to request to use local data to perform model training, and influence a global model by using the local data, to obtain a model that carries a personalized feature of the client” (Paragraph 148), “For example, when the data stored on the client exceeds a preset data amount, the request message may be sent to the federated learning system, and the ELB in the system routes the request message to the adapted first server, to implement federated learning through collaboration between the first server and the client” (Paragraph 149), and “For another example, the client may collect personalized training data that is different from training data of another device. After collecting the training data, the client may actively send the request message to the federated learning system, to request to use the locally stored data to perform federated learning, and to obtain a model adapted to the training data. The client synchronizes the model to the system, to complete federated learning, so that a global model in the cluster can adaptively learn various types of data collected by the clients, and output precision of the global model is improved” (Paragraph 150), “wherein when it is determined that it is not necessary to update the aggregate model to newly obtain the worker model, without updating the aggregate model to newly obtain the worker model” as “The client may request, by using the request message, to participate in federated learning from the first server, to request to use local data to perform model training, and influence a global model by using the local data, to obtain a model that carries a personalized feature of the client” (Paragraph 148), “For example, when the data stored on the client exceeds a preset data amount, the request message may be sent to the federated learning system, and the ELB in the system routes the request message to the adapted first server, to implement federated learning through collaboration between the first server and the client” (Paragraph 149), and “For another example, the client may collect personalized training data that is different from training data of another device. After collecting the training data, the client may actively send the request message to the federated learning system, to request to use the locally stored data to perform federated learning, and to obtain a model adapted to the training data. The client synchronizes the model to the system, to complete federated learning, so that a global model in the cluster can adaptively learn various types of data collected by the clients, and output precision of the global model is improved” (Paragraph 150) and “when it is determined that it is necessary to update the aggregate model to newly obtain the worker model, the processing circuitry updates the aggregate model through machine learning using the local learning data to obtain the worker model” as “The client may request, by using the request message, to participate in federated learning from the first server, to request to use local data to perform model training, and influence a global model by using the local data, to obtain a model that carries a personalized feature of the client” (Paragraph 148), “For example, when the data stored on the client exceeds a preset data amount, the request message may be sent to the federated learning system, and the ELB in the system routes the request message to the adapted first server, to implement federated learning through collaboration between the first server and the client” (Paragraph 149), and “For another example, the client may collect personalized training data that is different from training data of another device. After collecting the training data, the client may actively send the request message to the federated learning system, to request to use the locally stored data to perform federated learning, and to obtain a model adapted to the training data. The client synchronizes the model to the system, to complete federated learning, so that a global model in the cluster can adaptively learn various types of data collected by the clients, and output precision of the global model is improved” (Paragraph 150), “Operation 402: The first server separately sends a training configuration parameter and the global model to the at least one client for the request message” (Paragraph 151), “After receiving the training configuration parameter and the global model that are delivered by the first server, the client (or referred to as a first client) may train the global model by using a training set, to obtain a trained global model. The training set may include a plurality of training samples. The training sample may be data collected by the client, or may be data received by the client” (Paragraph 155), and “Operation 404: The client feeds back model update parameters to the first server” (Paragraph 157)
The examiner further notes that Zhang teaches the concept of a client in a federated system determining whether or not to produce a “worker model” via a received global model. The combination would result in the learners (i.e. clients) of Stripelis to determine whether or not produce its worker models.
It would have been obvious to one of ordinary skill in the art before the effective filing date of instant invention to combine the teachings of the cited references because teaching Zhang’s would have allowed Stripelis’s to provide a method for improving efficiency in federated learning, as noted by Zhang (Paragraph 4).
Stripelis and Zhang do not explicitly teach:
B) the processing circuitry acquires information that specifies a new aggregate model or confidential information of the information that specifies the new aggregate model from the secure federated learning device after a waiting time has elapsed.
Park, however, teaches “the processing circuitry acquires information that specifies a new aggregate model or confidential information of the information that specifies the new aggregate model from the secure federated learning device after a waiting time has elapsed” as “This application claims benefit of and priority to U.S. Provisional Patent Application Ser. No. 63/195,517, entitled “Federated Learning Using Secure Centers of Client Device Embeddings,” filed Jun. 1, 2021, and assigned to the assignee hereof, the contents of which are hereby incorporated by reference in its entirety” (Paragraph 1) and “the set of client devices to be used in training the global machine learning model may be selected based on one or more additional criteria. For example, client devices with higher usage and data acquisition may be selected over client devices with lower usage and data acquisition, as these devices may provide additional data that can be used to improve the quality of the global machine learning model. In some aspects, client devices may be selected based on an amount of time elapsed since the client devices last participated in training or updating the global machine learning model…a request to update the global machine learning model is transmitted to each respective client device in the selected set of client devices. The request generally includes information defining the global machine learning model. This information may include a plurality of model parameters, a plurality of secure centers, and information defining a radius of a secure hypersphere associated with each of the plurality of secure centers” (Paragraphs 69-70 of 63/195517).
The examiner further notes that although Zhang teaches different thresholds to determine if it is necessary for a client to produce a “worker model”, there is no explicit teaching of a time thresholds. Nevertheless, Park teaches the concept of using an elapsed waiting time period for transmitting a global model to clients for subsequent federated learning. The combination would result in expanding Stripelis to have a waiting period before using learners in its federated learning.
It would have been obvious to one of ordinary skill in the art before the effective filing date of instant invention to combine the teachings of the cited references because teaching Park’s would have allowed Stripelis’s and Zhang’s to provide a method for using additional data to train a global model, as noted by Park (Paragraph 69 of 63/195517).
15. Claim 3 is rejected under 35 U.S.C. 103 as being unpatentable over Stripelis et al. (Article entitled “Secure Neuroimaging Analysis using Federated Learning with Homomorphic Encryption”, dated 2021) as applied to claims 1, 4, and 7-10 above, in view of Fukumoto et al. (JP 2018032344A) (Machine Translation Provided).
16. Regarding claim 3, Stripelis does not explicitly teach a model learning device comprising:
A) wherein the processing circuitry further provides the secure federated learning device with plain text synchronization information indicating that the model learning device has provided the secure federated learning device with the confidential information of the information that specifies the worker model.
Fukomoto, however, teaches “wherein the processing circuitry further provides the secure federated learning device with plain text synchronization information indicating that the model learning device has provided the secure federated learning device with the confidential information of the information that specifies the worker model” as “First, the vertex management unit 33 (distributed processing management unit) of the processing server 30 (worker) monitors the distributed processing unit 20 (vertex) belonging to itself, thereby calculating / transmitting a certain distributed processing unit 20 (vertex). It is detected that fn is completed (step S10). Then, the vertex management unit 33 transmits a calculation / transmission processing completion report with the identification number of the distributed processing unit 20 (vertex) and the step number (n) of the super step at that time to the management server 10. (Step S11)” (Page 7).
The examiner further notes that Fukomoto teaches the concept of a worker sending a report (i.e. the undefined claimed plain text synchronization information in the broadest reasonable interpretation) to a management server. The combination would result in the workers of Stripelis to send reports to the controller that indicates the workers have provided the controller with the confidential information of the information that specifies the worker model.
It would have been obvious to one of ordinary skill in the art before the effective filing date of instant invention to combine the teachings of the cited references because teaching Fukomoto’s would have allowed Stripelis’s to provide a method for improving efficiency and speed in distributed processing, as noted by Fukomoto (Page 7).
17. Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over Stripelis et al. (Article entitled “Secure Neuroimaging Analysis using Federated Learning with Homomorphic Encryption”, dated 2021) as applied to claims 1, 4, and 7-10 above, in view of Choudhary et al. (U.S. PGPUB 2019/0385043).
18. Regarding claim 5, Stripelis further teaches a secure federated learning device comprising:
B) the processing circuitry obtains the confidential information of the information that specifies the aggregate model that is an aggregation of the worker models through secure computation using the confidential information of the information that specifies the worker model (Page 5, Section 5).
The examiner notes that Stripelis teaches “the processing circuitry obtains the confidential information of the information that specifies the aggregate model that is an aggregation of the worker models through secure computation using the confidential information of the information that specifies the worker model” as “In a centralized federated learning environment, each learner trains on its local dataset for an assigned number of local epochs, and upon completion of its local training task, it sends the local model to the federation controller to compute the new community model. In an encrypted centralized federation environment, the procedure is similar with the addition of three pivotal key steps: encryption, encrypted-aggregation, decryption. During the encryption step, every learner encrypts its locally trained model with an HE scheme using the public key, and sends the encrypted model (ciphertext) to the controller. For each learner, its encrypted model is treated as a vector of ciphertext objects, each object corresponding to a model array. With this approach, the encrypted data is represented as a (concatenated) collection of flattened data-vectors, each of them representing the local data for a particular learner. The controller receives all the encrypted local models, and then performs the encrypted weighted-aggregation to compute the new encrypted community model without ever decrypting any of the individual models. Subsequently, the controller sends the new community model to all the learners, and the learners decrypt it using the private key. Once the decryption is complete, the learners train the (new) decrypted model on their local data set and the entire procedure repeats” (Page 5, Section 5). The examiner further notes that the controller computing a new encrypted model (i.e. aggregate model) without ever decrypting the local models (i.e. worker models) teaches the claimed obtaining.
Stripelis does not explicitly teach:
A) wherein the processing circuitry is further configured to determine whether or not the processing circuitry has obtained the confidential information of the information that specifies the worker model from a predetermined model learning device;
B) wherein, when it is determined that the confidential information of the information that specifies the worker model has been obtained from the predetermined model learning device.
Chaudhary, however, teaches “wherein the processing circuitry is further configured to determine whether or not the processing circuitry has obtained the confidential information of the information that specifies the worker model from a predetermined model learning device” as “As indicated by the first timeline 402a and the second timeline 402b, the asynchronous training system 106 delays completion of a training iteration until receiving modified parameter indicators from certain client devices. In particular, upon detecting that one or more client devices have not sent a set of modified parameter indicators to the server(s) 102 in a threshold number of training iterations, the asynchronous training system 106 waits (e.g., waits a threshold time) for the one or more client devices to respond” (Paragraph 89) and “As indicated by the first timeline 402a for the initial training iteration, the asynchronous training system 106 receives modified-parameter-indicator sets 404a, 404b, and 404c from client devices 406a, 406b, and 406c by a first time 410” (Paragraph 90) and “wherein, when it is determined that the confidential information of the information that specifies the worker model has been obtained from the predetermined model learning device” as “As indicated by the first timeline 402a and the second timeline 402b, the asynchronous training system 106 delays completion of a training iteration until receiving modified parameter indicators from certain client devices. In particular, upon detecting that one or more client devices have not sent a set of modified parameter indicators to the server(s) 102 in a threshold number of training iterations, the asynchronous training system 106 waits (e.g., waits a threshold time) for the one or more client devices to respond” (Paragraph 89) and “As indicated by the first timeline 402a for the initial training iteration, the asynchronous training system 106 receives modified-parameter-indicator sets 404a, 404b, and 404c from client devices 406a, 406b, and 406c by a first time 410” (Paragraph 90).
The examiner further notes that Chaudhary teaches the concept of a server in a federated system determining whether or not it has received parameters from clients. The combination would result in the controller (i.e. server) of Stripelis to determine whether or not it has received the worker models from its learners (i.e. clients).
It would have been obvious to one of ordinary skill in the art before the effective filing date of instant invention to combine the teachings of the cited references because teaching Chaudhary’s would have allowed Stripelis’s to provide a method for ensuring slower clients are not excluded in federated learning, as noted by Chaudhary (Paragraph 89).
19. Claim 6 is rejected under 35 U.S.C. 103 as being unpatentable over Stripelis et al. (Article entitled “Secure Neuroimaging Analysis using Federated Learning with Homomorphic Encryption”, dated 2021) as applied to claims 1, 4, and 7-10 above, in view of Choudhary et al. (U.S. PGPUB 2019/0385043) as applied to claim 5 above, and further in view of Fukumoto et al. (JP 2018032344A) (Machine Translation Provided).
20. Regarding claim 6, Stripelis and Choudhary do not explicitly teach a model learning device comprising:
A) wherein the processing circuitry acquires plain text synchronization information indicating that the model learning device has provided the secure federated learning device with the confidential information of the information that specifies the worker model; and
B) uses the synchronization information to determine whether or not the confidential information of the information that specifies the worker model has been obtained from the predetermined model learning device.
Fukomoto, however, teaches “wherein the processing circuitry acquires plain text synchronization information indicating that the model learning device has provided the secure federated learning device with the confidential information of the information that specifies the worker model” as “First, the vertex management unit 33 (distributed processing management unit) of the processing server 30 (worker) monitors the distributed processing unit 20 (vertex) belonging to itself, thereby calculating / transmitting a certain distributed processing unit 20 (vertex). It is detected that fn is completed (step S10). Then, the vertex management unit 33 transmits a calculation / transmission processing completion report with the identification number of the distributed processing unit 20 (vertex) and the step number (n) of the super step at that time to the management server 10. (Step S11)” (Page 7) and “uses the synchronization information to determine whether or not the confidential information of the information that specifies the worker model has been obtained from the predetermined model learning device” as “First, the vertex management unit 33 (distributed processing management unit) of the processing server 30 (worker) monitors the distributed processing unit 20 (vertex) belonging to itself, thereby calculating / transmitting a certain distributed processing unit 20 (vertex). It is detected that fn is completed (step S10). Then, the vertex management unit 33 transmits a calculation / transmission processing completion report with the identification number of the distributed processing unit 20 (vertex) and the step number (n) of the super step at that time to the management server 10. (Step S11). Next, the management server 10 (master) that has received the calculation / transmission processing completion report, the adjacent synchronization processing unit 11 performs the adjacent synchronization processing on the distributed processing unit 20 (vertex) indicated by the received calculation / transmission processing completion report. It is determined whether or not the condition is satisfied (step S12). Specifically, the adjacent synchronization processing unit 11 determines whether or not “the calculation / transmission processing f n of all vertices that are in contact with the own vertex and the input edge is completed” is satisfied” (Page 7).
The examiner further notes that Fukomoto teaches the concept of a worker sending a report (i.e. the undefined claimed plain text synchronization information in the broadest reasonable interpretation) to a management server. The combination would result in the workers of Stripelis to send reports to the controller that indicates the workers have provided the controller with the confidential information of the information that specifies the worker model. Moreover, the claimed use of the synchronization information to determine whether or not conditional information has been obtained is interpreted as an intended use type limitation, and as a result, the report of Fukomoto can be “used” to determine if the confidential information of Stripelis has been obtained.
It would have been obvious to one of ordinary skill in the art before the effective filing date of instant invention to combine the teachings of the cited references because teaching Fukomoto’s would have allowed Stripelis’s and Chaudhary’s to provide a method for improving efficiency and speed in distributed processing, as noted by Fukomoto (Page 7).
Conclusion
21. The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
U.S. PGPUB 2020/0285980 issued to Sharad et al. on 10 September 2020. The subject matter disclosed therein is pertinent to that of claims 1-10 (e.g., methods to perform federated training).
U.S. PGPUB 2021/0287080 issued to Moloney et al. on 16 September 2021. The subject matter disclosed therein is pertinent to that of claims 1-10 (e.g., methods to perform federated training).
U.S. PGPUB 2022/0108177 issued to Samek et al. on 07 April 2022. The subject matter disclosed therein is pertinent to that of claims 1-10 (e.g., methods to perform federated training).
Contact Information
22. Any inquiry concerning this communication or earlier communications from the examiner should be directed to Mahesh Dwivedi whose telephone number is (571) 272-2731. The examiner can normally be reached on Monday to Friday 8:20 am – 4:40 pm.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Charles Rones can be reached (571) 272-4085. The fax 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 20. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free).
Mahesh Dwivedi
Primary Examiner
Art Unit 2168
September 11, 2026
/MAHESH H DWIVEDI/Primary Examiner, Art Unit 2168
Prosecution Insights