Notice of Pre-AIA or AIA Status
This Non-Final communication is in response to Application No. 18/417,005 filed on 01/19/2024. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Election/Restrictions
The Requirement for Restriction/Election issued on 10/29/2025 is withdrawn. This Non-Final office action addresses pending claims 1-20.
Priority
Acknowledgment is made of applicant’s claim for foreign priority under 35 U.S.C. 119 (a)-(d). The certified copy has been filed in parent Application No. TW112105224, filed on 2/14/2023.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-2, 5-16, 18-20 are rejected under 35 U.S.C 103 as being unpatentable over Kourtellis (NPL ‘FLaaS: Federated Learning as a Service’ (2020)) in view of Abreha (NPL ‘Federated Learning in Edge Computing: A Systematic Survey’ (2022).
Regarding claim 1, Kourtellis teaches:
A federated learning booking platform with a clustered architecture for training based on booking information input by a first user end, comprising: (Page 4 left column “Front-End: Main interface for customers (e.g., app or service developers), to bootstrap, configure, and terminate the service. It runs a GUI, processes front-end API calls from customers (or the GUI), and calls functions on the Controller to execute customer requests.”)
a second user end having a dataset; and (page 4 bottom of right column “Jointly-trained FL modeling between group of apps for existing ML problem: In the following scenario, we assume a group of two or more apps, i 2 A, installed on device k 2 K, are interested in collaborating and building a common FL model with FLaaS. This model will be shared among all apps but will be built jointly on each application’s local data.”)
Kourtellis does not teach:
a training end comprising a main server and a sub-server operated under an assignment from the main server, wherein the main server comprises a service server and the sub-server comprises a backup server, the service server is configured to receive the booking information and communicate with the backup server and the first user end based on the booking information, the backup server is configured to store the dataset of the second user end.
However, Abreha does:
a training end comprising a main server and a sub-server operated under an assignment from the main server, wherein the main server comprises a service server and the sub-server comprises a backup server, the service server is configured to receive the booking information and communicate with the backup server and the first user end based on the booking information, the backup server is configured to store the dataset of the second user end. (Fig 2b shows the Cloud Center/Scheduler (main server) which has edge node (sub-server). The Cloud Center/Scheduler oversees the training and scheduling of tasks and it can be seen that the edge nodes receive data from a second user.)
Kourtellis and Abreha are considered analogous art to the claimed invention because they are in the same field of endeavor being federated learning. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the federated learning service of Kourtellis with the edge computing of Abreha. One would want to do this as Kourtellis already has the capabilities of edge computing (Fig. 1) but does not go into the details.
Regarding claim 2, Kourtellis in view of Abreha teaches claim 1 as outlined above. Kourtellis further teaches:
the training end further comprises a booking interface configured to provide the first user end to transmit the booking information, enabling the service server to receive and communicate with the backup server and the first user end based on the booking information. (Page 4 left column “Front-End: Main interface for customers (e.g., app or service developers), to bootstrap, configure, and terminate the service. It runs a GUI, processes front-end API calls from customers (or the GUI), and calls functions on the Controller to execute customer requests.” Where Abreha teaches the backup server as mentioned above.)
Regarding claim 5, Kourtellis in view of Abreha teaches claim 2 as outlined above. Kourtellis further teaches:
the booking interface is a booking web page provided by the service server or a web server. (Page 4 left column “Front-End: Main interface for customers (e.g., app or service developers), to bootstrap, configure, and terminate the service. It runs a GUI, processes front-end API calls from customers (or the GUI), and calls functions on the Controller to execute customer requests.”)
Regarding claim 6, Kourtellis teaches:
A federated learning booking system with a clustered architecture, comprising: (Abstract)
a first user end configured to input booking information; (Page 4 left column “Front-End: Main interface for customers (e.g., app or service developers), to bootstrap, configure, and terminate the service. It runs a GUI, processes front-end API calls from customers (or the GUI), and calls functions on the Controller to execute customer requests.”)
a second user end having a dataset; and (page 4 bottom of right column “Jointly-trained FL modeling between group of apps for existing ML problem: In the following scenario, we assume a group of two or more apps, i 2 A, installed on device k 2 K, are interested in collaborating and building a common FL model with FLaaS. This model will be shared among all apps but will be built jointly on each application’s local data.”)
Kourtellis does not teach:
a training end comprising a main server and a sub-server operated under an assignment from the main server, wherein the main server comprises a service server and the sub-server comprises a backup server, the service server is configured to receive the booking information and communicate with the backup server and the first user end based on the booking information, the backup server is configured to store the dataset of the second user end.
However, Abreha does:
a training end comprising a main server and a sub-server operated under an assignment from the main server, wherein the main server comprises a service server and the sub-server comprises a backup server, the service server is configured to receive the booking information and communicate with the backup server and the first user end based on the booking information, the backup server is configured to store the dataset of the second user end. (Fig 2b shows the Cloud Center/Scheduler (main server) which has edge node (sub-server). The Cloud Center/Scheduler oversees the training and scheduling of tasks and it can be seen that the edge nodes receive data from a second user.)
Kourtellis and Abreha are considered analogous art to the claimed invention because they are in the same field of endeavor being federated learning. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the federated learning service of Kourtellis with the edge computing of Abreha. One would want to do this as Kourtellis already has the capabilities of edge computing (Fig. 1) but does not go into the details.
Regarding claim 7, Kourtellis in view of Abreha teaches claim 6 as outlined above. Kourtellis further teaches:
the service server transmits corresponding control information and a first initial training model to the backup host and the first user end based on the booking information. (Page 4 left column “Controller: Takes as input commands received from Front-End, executes the required steps to configure the service, e.g., initialize the model, set appropriate permissions, etc. Once the service starts, the Controller is in charge of monitoring service health, budget, and terminating execution of ML modeling when requested.” And where Abreha teaches the backup host as mentioned above.)
Regarding claim 8, Kourtellis in view of Abreha teaches claim 7 as outlined above. Abreha further teaches:
the backup host trains by using the first initial training model and the dataset based on the control information to generate and return a first training model to the service server, the first user end trains by using the first initial training model based on the control information to generate and return a second training model to the service server, the service server receives and performs computations by using the first training model and the second training model to generate and return a third training model to the backup host and the first user end. (Page 9 “Local model training: The participants receive the global model w^t_G, where t denotes the current iteration index, and each participant updates the local model parameters w^t_i based on their local data and device. The objective of client i is therefore to obtain an optimal parameter w^t_i at the t time iteration at the minimum value of the loss function … Finally, each local model’s updated parameters are sent again back to the FL parameter server; Global model aggregation: The centralized server receives the local parameters from each participant and aggregates the local models from the participants, then sends the updated global model parameters w^t+1_G back to all the participating clients to minimize the global loss function”)
Regarding claim 9, Kourtellis in view of Abreha teaches claim 8 as outlined above. Abreha further teaches:
when the first user end re-enters the booking information into the booking interface, the service server performs a test by using the first training model, the second training model and the third training model to generate a test result, used as a second initial training model, the service server transmits the control information and the second initial training model to the backup host and the first user end based on the booking information, the backup host trains by using the second initial training model and the dataset based on the control information to generate and return a fourth training model to the service server, the first user end trains by using the second initial training model based on the control information to generate and return a fifth training model to the service server, the service server performs computation by using the fourth training model and the fifth training model to generate and return a sixth training model to the backup host and the first user end. (Page 9 “Furthermore, Steps 2 and 3 are repeated in the iteration process until the global loss function achieves the optimal accuracy.” These steps are referring to the steps outlined in claim 8)
Regarding claim 10, Kourtellis in view of Abreha teaches claim 6 as outlined above. Abreha further teaches:
a number of the second user end is greater than or equal to a number of the first user end. (We can see this in Fig. 2b.)
Regarding claim 11, Kourtellis in view of Abreha teaches claim 6 as outlined above. Abreha further teaches:
the dataset of the second user end is stored on the training end, or the second user end uses the backup host of the training end for storing the dataset. (Figure 2b shows the second users data (data producer) going to the sub-server (edge node) and the main server (cloud center)).
Regarding claim 12, Kourtellis in view of Abreha teaches claim 11 as outlined above. Abreha further teaches:
when the second user end stores the dataset on the backup host at fixed or non-fixed time, the second user end transmits path data of the dataset and backup instructions to the service server. (Page 27 5th paragraph “To determine the primary and dual variables, the traditional distributed approach sends the local stochastic gradients to a central node, which then aggregates the gradients from each client”)
Regarding claim 13, Kourtellis in view of Abreha teaches claim 12 as outlined above. Abreha further teaches:
the dataset obtained by the backup host is either a copy of the dataset or a data shortcut of the dataset. (Page 2 3rd paragraph “Reliability: Clients send their datasets via different communication network connections to the remotely located cloud server in conventional centralized model-training architectures. Therefore, the wireless communications and core network connections between clients and servers affect DL model training and inferences to a large extent. Hence, the connection has to be reasonably reliable even when there is an interruption in the network. Nevertheless, a centralized architecture faces system performance degradation and possible failure because of the unreliable wireless connection between the client and server, which can significantly affect the model;”)
Regarding claim 14, Kourtellis in view of Abreha teaches claim 6 as outlined above. Abreha further teaches:
the first user end stores desired training data on the training end, or the first user end uses the backup host of the training end for storing data. (Page 28 last paragraph: “Traditional distributed learning systems provide access to the entire training dataset to a central server”)
Regarding claim 15, Kourtellis in view of Abreha teaches claim 7 as outlined above. Kourtellis further teaches:
the control information is configured to execute an application, allowing the backup host and the first user end to connect to the service server via a secured mechanism. (Start of section 3 “FLaaS aims at providing to single applications an easy way to use FL, without the costly process of developing and tuning the algorithms, as well as to enable multiple applications to collaboratively build models with minimal efforts”)
Regarding claim 16, Kourtellis in view of Abreha teaches claim 15 as outlined above. Kourtellis further teaches:
the service server possesses a host account of the backup host and use the host account as an authentication mechanism. (Page 4 section 4.2 left column “PERMISSION APIs enable customers to specify if and which other customers (e.g., apps) can access said data, or how other customers can access the model for inference, or to build models to solve new ML problems.”)
Regarding claim 18, Kourtellis in view of Abreha teaches claim 9 as outlined above. Abreha further teaches:
the first user end and the second user end obtain final training models from folders of hosts, servers, electronic devices used for training, or cloud storage. (We can see on figure 2b that the central server is in the cloud and thus the users receive the models from the cloud.)
Regarding claim 19, Kourtellis in view of Abreha teaches claim 6 as outlined above. Abreha further teaches:
the second user end either store data on the backup host or use the backup host as a primary data storage space. (Page 28 last paragraph “Traditional distributed learning systems provide access to the entire training dataset to a central server. To obtain an efficient model, the server splits the dataset into subsets and distributes them to participating devices based on the distributions”)
Regarding claim 20, Kourtellis teaches:
A federated learning booking method with a clustered architecture, comprising: (Abstract)
communicating with a backup host of a training end and a first user end by a service server based on booking information when the service server of the training end receives the booking information from the first user end, wherein the backup host is assigned to operate by the service server, and the backup host obtains a dataset from a second user end; (Page 4 left column “Front-End: Main interface for customers (e.g., app or service developers), to bootstrap, configure, and terminate the service. It runs a GUI, processes front-end API calls from customers (or the GUI), and calls functions on the Controller to execute customer requests.” And page 4 bottom of right column “Jointly-trained FL modeling between group of apps for existing ML problem: In the following scenario, we assume a group of two or more apps, i 2 A, installed on device k 2 K, are interested in collaborating and building a common FL model with FLaaS. This model will be shared among all apps but will be built jointly on each application’s local data.”)
transmitting control information and a first initial training model to the backup host and the first user end from the service server; (Page 4 left column “Controller: Takes as input commands received from Front-End, executes the required steps to configure the service, e.g., initialize the model, set appropriate permissions, etc. Once the service starts, the Controller is in charge of monitoring service health, budget, and terminating execution of ML modeling when requested.” )
Kourtellis does not teach:
The backup host
conducting training by the backup host using the first initial training model and the dataset, generating and returning a first training model to the service server;
conducting training by the first user end using the initial training model based on the control information, generating and returning a second training model to the service server; and
generating a third training model by operating the first training model and the second training model in the service server, returning the third training model to the backup host and the first user end, wherein the second user end obtains the third training model from the backup host;
wherein the service server performs a test by using the first training model, the second training model and the third training model to generate a test result, used as a second initial training model, when the first user end re-enters the booking information into the booking interface, the service server transmits the control information and the second initial training model to the backup host and the first user end.
However, Abreha does:
The backup host (Fig 2b shows the Cloud Center/Scheduler (main server) which has edge node (sub-server). The Cloud Center/Scheduler oversees the training and scheduling of tasks and it can be seen that the edge nodes receive data from a second user.)
conducting training by the backup host using the first initial training model and the dataset, generating and returning a first training model to the service server; (Page 9 “Local model training: The participants receive the global model w^t_G, where t denotes the current iteration index, and each participant updates the local model parameters w^t_i based on their local data and device. The objective of client i is therefore to obtain an optimal parameter w^t_i at the t time iteration at the minimum value of the loss function”)
conducting training by the first user end using the initial training model based on the control information, generating and returning a second training model to the service server; and (Page 9 “Local model training: The participants receive the global model w^t_G, where t denotes the current iteration index, and each participant updates the local model parameters w^t_i based on their local data and device. The objective of client i is therefore to obtain an optimal parameter w^t_i at the t time iteration at the minimum value of the loss function”)
generating a third training model by operating the first training model and the second training model in the service server, returning the third training model to the backup host and the first user end, wherein the second user end obtains the third training model from the backup host; ( page 9 “Finally, each local model’s updated parameters are sent again back to the FL parameter server; Global model aggregation: The centralized server receives the local parameters from each participant and aggregates the local models from the participants, then sends the updated global model parameters w^t+1_G back to all the participating clients to minimize the global loss function”)
wherein the service server performs a test by using the first training model, the second training model and the third training model to generate a test result, used as a second initial training model, when the first user end re-enters the booking information into the booking interface, the service server transmits the control information and the second initial training model to the backup host and the first user end. (Page 9 “Furthermore, Steps 2 and 3 are repeated in the iteration process until the global loss function achieves the optimal accuracy.” These steps are referring to the steps outlined in the previous limitations)
Kourtellis and Abreha are considered analogous art to the claimed invention because they are in the same field of endeavor being federated learning. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the federated learning service of Kourtellis with the edge computing of Abreha. One would want to do this as Kourtellis already has the capabilities of edge computing (Fig. 1) but does not go into the details.
Claims 3 and 4 are rejected under 35 U.S.C 103 as being unpatentable over Kourtellis in view of Abreha and Georgiadis (US 2023/0058972 A1).
Regarding claim 3, Kourtellis in view of Abreha teaches claim 1 as outlined above. Neither Kourtellis or Abreha teaches:
the service server encrypts and communicates with the first user end by a hypertext transfer protocol secure (HTTPS).
However Georgiadis does:
the service server encrypts and communicates with the first user end by a hypertext transfer protocol secure (HTTPS). ([0053] “In step 220, each client network may communicate its weights to the federated learning computer program. For example, the client networks may communicate encrypted tokens and the weights using TCP/IP, HTTP, a virtual private network, SSH, or other suitable means for communication between the client network and the federated learning computer program.”)
Kourtellis, Abreha and Georgiadis are considered analogous art to the claimed invention because they are in the same field of endeavor being federated learning. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to substitute the communication method of Kourtellis and Abreha with the communication method of Georgiadis. One would want to do this to have more secure communication.
Regarding claim 4, Kourtellis in view of Abreha teaches claim 2 as outlined above. Neither Kourtellis or Abreha teaches:
the booking interface encrypts and communicates with the service server by a secure shell protocol (SSH).
However, Georgiadis does:
the booking interface encrypts and communicates with the service server by a secure shell protocol (SSH). ([0053] “In step 220, each client network may communicate its weights to the federated learning computer program. For example, the client networks may communicate encrypted tokens and the weights using TCP/IP, HTTP, a virtual private network, SSH, or other suitable means for communication between the client network and the federated learning computer program.”)
Kourtellis, Abreha and Georgiadis are considered analogous art to the claimed invention because they are in the same field of endeavor being federated learning. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to substitute the communication method of Kourtellis and Abreha with the communication method of Georgiadis. One would want to do this to have more secure communication.
Claim 17 are rejected under 35 U.S.C 103 as being unpatentable over Kourtellis in view of Abreha and Dai (US 2024/0054350 A1).
Regarding claim 17, Kourtellis in view of Abreha teaches claim 8 as outlined above. Neither Kourtellis or Abreha teaches:
notifications are sent to the first user end and second user end by emails when the training is completed.
However, Dai does:
notifications are sent to the first user end and second user end by emails when the training is completed. ([0055] “Other applications 316 may also include communication applications, such as email, texting, voice, social networking, and IM applications that allow a user to send and receive emails, calls, texts, and other notifications through network 360.”)
Kourtellis, Abreha and Dai are considered analogous art to the claimed invention because they are in the same field of endeavor being federated learning. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the federated learning system of Kourtellis and Abreha with the notification method of Dai. One would want to do this to be able to notify users when training is done.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DANIEL P GRUSZKA whose telephone number is (571)272-5259. The examiner can normally be reached M-F 9:00 AM - 6:00 PM ET.
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, Li Zhen can be reached at (571) 272-3768. 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.
/DANIEL GRUSZKA/Examiner, Art Unit 2121
/Li B. Zhen/Supervisory Patent Examiner, Art Unit 2121
Prosecution Insights