Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Detail Action
This office action is response to the application 19/047,870 filed on 02/07/2025. Claims 1-20 are pending in this communication.
Priority
This application claims priority from KOREA, REPUBLIC OF 10-2024-0020869 02/14/2024. Priority date has been accepted.
Claim Rejections - 35 USC § 103
The following is a quotation of AIA 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, 6, 11, 15 & 17 are rejected under AIA 35 U.S.C. 103 as being unpatentable over CUI; Tao et al. (US 2025/0267187 A1) in view of RODRIGUEZ MULET; Albert et al. (US 2024/0289635 A1)
Regarding Claim 1, CUI discloses a personalized federated learning method performed by a processor of a user terminal operating in conjunction with a server, the method comprising:
generating input data based on user data received through an interface of the user terminal; inputting the input data into a first learning model provided in the user terminal and training the first learning model using the corresponding output {[0029], “The FL (Federated Learning) server 110 performs device selection among the terminal devices A to E according to the received reports, to determine terminal devices participating in model training in the current iteration … a capability of the terminal device to upload the local training model”. Examiner’s note: cited ‘local training model’ is functioning as claimed ‘first learning model’};
…
receiving global parameters derived based on the local parameters from the server; and inputting the input data into the first learning model, to which the global parameters are applied, and a second learning model associated with the first learning model, and training the second learning model using the corresponding output {[0030], “in FIG. 1, the FL server 110 selects, in an N-th iteration, the terminal devices A, C and D to perform federated learning. Next, the FL server 110 distributes the global-model and a training parameter configuration to the selected terminal devices A, C and D. Then, the terminal devices A, C and D update the local-models to be consistent with the global-model, and train the local-models based on the local data”. Examiner’s note: ‘N-th iteration of federated learning’ is functioning as claimed ‘second learning model’}.
CUI, however, does not explicitly disclose
transmitting local parameters for weights of a neural network included in the first learning model to the server;
In an analogous reference RODRIGUEZ discloses
transmitting local parameters for weights of a neural network included in the first learning model to the server {[0021], “The local communication unit 104 transmits local model parameters and the preliminary local training information regarding the updated local model to the server 11. The local model parameters are parameters (weighting factor, bias, and the like) of the neural network for sharing parameters with the global model. Furthermore, the local communication unit 104 receives information regarding the global model from the server 11. The information regarding the global model is, for example, parameters of the global model”};
Before the effective filing date of the claimed invention, it would have been obvious to one with ordinary skill in the art to modify CUI’s technique of ‘personalized federated learning method trains a local model on terminal-specific data’ to ‘transmit parameters to a server to receive global updates, and then refines a second, associated local model using those global parameters’, as taught by RODRIGUEZ, in order to prepare training local models with global parameters. The motivation is - this architecture optimizes AI by providing data privacy, personalized accuracy, and reduced bandwidth. It allows devices to learn from a global network while keeping sensitive raw data local, ensuring that the AI adapts to specific user data without compromising security or needing a constant, high-speed cloud connection.
All references are inventions in analogous area but each invention teaches specific claimed limitation specifically and other references mutually cure each other’s deficiencies. When all claimed techniques are combined, they teach claimed invention. The Examiner notes that this motivation applies to all dependent and/or otherwise subsequently addressed claims unless addressed separately.
Regarding Claim 6, CUI as modified by RODRIGUEZ discloses all the features of claim 1. The combination further discloses
wherein the global parameter is calculated by the server based on a plurality of local parameters for each of the first learning models provided in different user terminals {CUI: [0030], “the FL server 110 distributes the global-model and a training parameter configuration to the selected terminal devices A, C and D. Then, the terminal devices A, C and D update the local-models to be consistent with the global-model, and train the local-models based on the local data”}.
Regarding Claim 11, CUI as modified by RODRIGUEZ discloses all the features of claim 1. The combination further discloses
wherein the training the first learning model and the training the second learning model are sequentially and repeatedly performed {CUI: [0030], “in FIG. 1, the FL server 110 selects, in an N-th iteration, the terminal devices A, C and D to perform federated learning” … [0044], “The FL participant entities receiving the common data, by taking the common data as the inputs of the local-models, sequentially input the common data into the local-models trained by local data, to obtain a series of output results, and return these output results to the first-level FL server entity”}.
Regarding claim 15, claim 15 is claim to an apparatus using the method of claim 1. Therefore, claim 15 is rejected for the reasons set forth for claim 1.
Regarding claim 17, claim 17 is claim to an apparatus using the method of claim 1. Therefore, claim 17 is rejected for the reasons set forth for claim 1.
Claim 2 is rejected under AIA 35 U.S.C. 103 as being unpatentable over CUI; Tao et al. (US 2025/0267187 A1) in view of RODRIGUEZ MULET; Albert et al. (US 2024/0289635 A1) and further in view of JIANG; Changjun et al. (US 2019/0073457 A1).
Regarding Claim 2, CUI as modified by RODRIGUEZ discloses all the features of claim 1. The combination, however, does not explicitly disclose
wherein the user data includes an image captured by a camera provided in the user terminal or user's touch pattern information input on a touch display provided in the user terminal.
In an analogous reference JIANG discloses
wherein the user data includes an image captured by a camera provided in the user terminal or user's touch pattern information input on a touch display provided in the user terminal {Fig. 1 & [0037], “analysis system of touch screen user keypress behavior pattern comprises a user data acquisition module (not shown), a data preprocessing module, a model training module and a user identity authentication module”}.
Before the effective filing date of the claimed invention, it would have been obvious to one with ordinary skill in the art to modify CUI’s technique as modified by RODRIGUEZ of personalized federated learning method trains a local model on terminal-specific data to transmit parameters to a server to receive global updates, and then refines a second, associated local model using those global parameters’ for ‘capturing user’s touch screen touching pattern and input in model training system’ by JIANG, in order to build local training model. The motivation is - by capturing biometric touch patterns locally, this method provides continuous, invisible authentication that adapts to your specific physical habits without ever sending your raw "digital fingerprint" to the cloud. This ensures high-precision security and a smoother user experience.
Claims 3 & 4 are rejected under AIA 35 U.S.C. 103 as being unpatentable over CUI; Tao et al. (US 2025/0267187 A1) in view of RODRIGUEZ MULET; Albert et al. (US 2024/0289635 A1) and further in view of JIANG; Changjun et al. (US 2019/0073457 A1) and YU; HaiYan et al. (US 2025/0391539 A1).
Regarding Claim 3, CUI as modified by RODRIGUEZ and further modified by JIANG discloses all the features of claim 2. The combination, however, does not disclose
when the user data is an image captured by the camera provided in the user terminal, dividing the image into a plurality of patches; and converting the plurality of divided patches into linear data through embedding based on positional information of the image.
In an analogous reference YU discloses
when the user data is an image captured by the camera provided in the user terminal, dividing the image into a plurality of patches; and converting the plurality of divided patches into linear data through embedding based on positional information of the image {[0074], “an image recognition system process … putting the obtained data into the learning model to further optimize the algorithm … [0078], “images of different specifications [Top View or Side View] … are used as input (Image Acquisition), and each image contains the calibration object and positioning for estimating the image scale factor; through the object detection network of the deep learning network … and segment the target (Image Segmentation)”}.
Before the effective filing date of the claimed invention, it would have been obvious to one with ordinary skill in the art to modify CUI’s technique as modified by RODRIGUEZ of personalized federated learning method trains a local model on terminal-specific data to transmit parameters to a server to receive global updates, and then refines a second, associated local model using those global parameters’ for ‘segmenting an image data and position to a learning model’, as disclosed by YU. The motivation is - this technique enables private spatial awareness by segmenting images and identifying object coordinates directly on the device. By combining global recognition logic with a refined local model, the system learns a user’s specific physical environment.
All references are inventions in analogous area but each invention teaches specific claimed limitation specifically and other references mutually cure each other’s deficiencies. When all claimed techniques are combined, they teach claimed invention. The Examiner notes that this motivation applies to all dependent and/or otherwise subsequently addressed claims unless addressed separately.
Regarding Claim 4, CUI as modified by RODRIGUEZ and further modified by JIANG discloses all the features of claim 2. The combination, however, does not disclose
when the user data is touch pattern information input on the touch display provided in the user terminal, deriving positional information and time information corresponding to a plurality of touch inputs included in the touch pattern information; and mapping the positional information and the time information for specific touch inputs and generating sequential data arranged in the order in which the touch inputs are applied {[0002], “FFQ can be used in large samples and can reflect the dose-dependent relationship between food types, intake and disease over a long period of time” … [0074], “an image recognition system process … putting the obtained data into the learning model to further optimize the algorithm … [0078], “images of different specifications [Top View or Side View] … are used as input (Image Acquisition), and each image contains the calibration object and positioning for estimating the image scale factor; through the object detection network of the deep learning network … and segment the target (Image Segmentation)”}.
Claim 5 is rejected under AIA 35 U.S.C. 103 as being unpatentable over CUI; Tao et al. (US 2025/0267187 A1) in view of RODRIGUEZ MULET; Albert et al. (US 2024/0289635 A1) and further in view of SHAO; Yunfeng et al. (US 2024/0394556 A1).
Regarding Claim 5, CUI as modified by RODRIGUEZ discloses all the features of claim 1. The combination, however, does not explicitly disclose
inputting the input data and receiving first output data as an output; deriving a first loss value of the first output data; and updating the neural network of the first learning model such that the first loss value is minimized.
In an analogous reference SHAO discloses
inputting the input data and receiving first output data as an output; deriving a first loss value of the first output data; and updating the neural network of the first learning model such that the first loss value is minimized {[0152], “training of the deep neural network is a process of reducing the loss as much as possible ... After a plurality of rounds of iterative training is performed on the selection model or after the selection model is converged, it may be considered that the training target is achieved” … [0253], “The client device calculates a task loss value based on the output results of the models and a real label”}.
Before the effective filing date of the claimed invention, it would have been obvious to one with ordinary skill in the art to modify CUI’s technique as modified by RODRIGUEZ of personalized federated learning method trains a local model on terminal-specific data to transmit parameters to a server to receive global updates, and then refines a second, associated local model using those global parameters’ for ‘calculating and reducing loss value of output data of the training models, as disclosed by SHAO. The motivation is - reducing loss value minimizes error, transforming raw data into high-precision intelligence. By refining internal weights, the learning model optimizes predictive accuracy and computational speed, ensuring it masters underlying patterns to solve complex, real-world problems effectively.
Allowable subject matter
Independent claims 12, 13 & 18 -19 are allowed. Dependent claims 14 & 20 are also allowed in virtue of independent claims. Therefore, claims 12-14 and 18-20 are allowed.
Claims 7 & 8 will be allowable if written in independent form with base method claim 1:
Reasons of allowance: what is missing from the prior arts is: inputting the input data and receiving the second output data generated as the output of the second learning model; deriving a second loss value of the second output data; and updating the first to third personalized parameters such that the second loss value is minimized.
Claim 16 will be allowable if written in independent form with base apparatus claim 15: Reasons of allowance: what is missing from the prior arts is: receiving a request for a specific service from a user through the interface; receiving a user authentication result from the server, determined based on the similarity between the output data and pre-registered user data stored in the server; and providing the service requested by the user on the screen of the user terminal if the user authentication result is determined to be successful.
Therefore, claims 7-10 & 16 are objected.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to QUAZI FAROOQUI whose telephone number is (571) 270-1034 or Quazi.farooqui@USPTO.GOV. The examiner can normally be reached on Monday-Friday 9:00 am to 5:30 pm, EST. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Bill Korzuch can be reached on (571) 272-7589 or William.Korzuch@USPTO.GOV. The fax phone number for Examiner Farooqui assigned is 571-270-2034.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-flee). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/QUAZI FAROOQUI/
Primary Examiner, Art Unit 2491
Prosecution Insights