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 .
DETAILED OFFICE ACTION
Status of Claims
Claims 1-20 are rejected and pending in this Office Action.
Double Patenting
1. The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP §§ 706.02(l)(1) - 706.02(l)(3) for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/process/file/efs/guidance/eTD-info-I.jsp.
2. Claims 1,2,3,4,5,6,7,8,9,10,11,12&14,13,15,16,17,18,19 and 20 are rejected on the ground of non-statutory double patenting as being unpatentable over claims 1,2,1,1,3,4,5,5-6,7,8,9,10,13,11,14,15,16,17 and 18 respectively of U.S. Patent No. 12,248,412. Although the claims at issue are not identical, they are not patentably distinct from each other because of the following similarity of the Claim limitations:
Present Application
U.S. Patent No. 12,248,412
As per claim 1, A method, comprising: hosting, by a first computing device, a master version of an artificial neural network (ANN);
hosting, by the first computing device, a master version of a feature dictionary;
receiving, by the first computing device, encoded features from a second computing device; decoding, by the first computing device, the received encoded features according to the master version of the feature dictionary; and
training, by the first computing device, the master version of the ANN based on the decoded features using machine learning.
As per claim 1, A method, comprising: hosting, by a first computing device, a master version of an artificial neural network (ANN); hosting, by the first computing device, a master version of a feature dictionary; receiving, by the first computing device, encoded features from a second computing device; decoding, by the first computing device, the received encoded features according to the master version of the feature dictionary; training, by the first computing device, the master version of the ANN based on the decoded features using machine learning; …
As per claim 2, The method of claim 1,
comprising transmitting, by the first computing device, the trained master version of the ANN to the second computing device.
As per claim 2,The method of claim 1, comprising transmitting, by the first computing device, the trained master version of the ANN to the second computing device.
As per claim 3, The method of claim 1,
comprising receiving, by the first computing device, the local version of the feature dictionary from the second computing device.
As per claim 1,… receiving, by the first computing device, the local version of the feature dictionary from the second computing device; …
As per claim 4,The method of claim 3, comprising changing, by the first computing device, the master version of the feature dictionary based on the received local version of the feature dictionary.
As per claim 1, …and changing, by the first computing device, the master version of the feature dictionary based on the received local version of the feature dictionary.
As per claim 5, The method of claim 4,
wherein the decoding comprises decoding the encoded features according to the changed master version of the feature dictionary.
As per claim 3, The method of claim 1,
wherein the decoding comprises decoding the encoded features according to the changed master version of the feature dictionary.
As per claim 6, The method of claim 1,
comprising transmitting, by the first computing device, the changed master version of the feature dictionary to the second computing device.
As per claim 4,The method of claim 1, comprising transmitting, by the first computing device, the changed master version of the feature dictionary to the second computing device.
As per claim 7, A method, comprising: hosting, by a second computing device, a local version of an artificial neural network (ANN); hosting, by the second computing device, a local version of a feature dictionary; extracting, by the second computing device, features from user data hosted by the second computing device;
encoding, by the second computing device, the extracted features according to the local version of the feature dictionary;
transmitting, by the second computing device, the encoded features to a first computing device that hosts a master version of the ANN so that the encoded features are decoded by a master version of the feature dictionary hosted by the first computing device and then used as input to train the master version of the ANN using machine learning.
As per claim 5, A method, comprising: hosting, by a second computing device, a local version of an artificial neural network (ANN); hosting, by the second computing device, a local version of a feature dictionary; extracting, by the second computing device, features from user data hosted by the second computing device; encoding, by the second computing device, the extracted features according to the local version of the feature dictionary; transmitting, by the second computing device, the encoded features to a first computing device that hosts a master version of the ANN; …
As per claim 8, The method of claim 7,
comprising: determining, by the second computing device, whether the extracted features are included in the local version of the feature dictionary;
in response to the local version of the feature dictionary including the extracted features, encoding the extracted features according to the local version of the feature dictionary; and in response to the local version of the feature dictionary not including the extracted features, changing the local version of the feature dictionary based on the extracted features and then encoding the extracted features according to the changed local version of the feature dictionary.
As per claim 5, … determining, by the second computing device, whether the extracted features are included in the local version of the feature dictionary; and in response to the local version of the feature dictionary including the extracted features, encoding the extracted features according to the local version of the feature dictionary.
As per claim 6,The method of claim 5, comprising: in response to the local version of the feature dictionary not including the extracted features, changing the local version of the feature dictionary based on the extracted features and then encoding the extracted features according to the changed local version of the feature dictionary.
As per claim 9, The method of claim 8,
comprising: transmitting the changed local version of the feature dictionary to the first computing device so that the master version of the feature dictionary is changed according to the local version of the feature dictionary.
As per claim 7, The method of claim 6, comprising: transmitting the changed local version of the feature dictionary to the first computing device so that the master version of the feature dictionary is changed according to the local version of the feature dictionary.
As per claim 10, The method of claim 9,
comprising: receiving a changed master version of the feature dictionary that was changed according to the local version of the feature dictionary; and changing the local version of the feature dictionary based on the changed master version of the feature dictionary.
As per claim 8, The method of claim 7, comprising: receiving a changed master version of the feature dictionary that was changed according to the local version of the feature dictionary; and changing the local version of the feature dictionary based on the changed master version of the feature dictionary.
As per claim 11, The method of claim 7,
comprising: receiving the trained master version of the ANN; and changing the local version of the ANN based on the trained master version of the ANN.
As per claim 9, The method of claim 5, comprising: receiving the trained master version of the ANN; and changing the local version of the ANN based on the trained master version of the ANN.
As per claim 12, A system, comprising: memory configured to: store a master version of an artificial neural network (ANN); and store a master version of a feature dictionary; a processor configured to: decode encoded features, received from a second computing device, according to the master version of the feature dictionary; and train the master version of the ANN based on the decoded features using machine learning.
As per claim 14, The system of claim 12,
wherein the transceiver of the first computing device is configured to receive the local version of the feature dictionary from the second computing device, and wherein the processor of the first computing device is configured to: change the master version of the feature dictionary based on the received local version of the feature dictionary; and decode the received encoded features according to the changed master version of the feature dictionary.
As per claim 10, A system, comprising: a second computing device; and a first computing device communicatively coupled to the second computing device, comprising: memory configured to: store a master version of an artificial neural network (ANN); and store a master version of a feature dictionary; transceiver configured to receive encoded features from the second computing device; and a processor configured to: decode the received encoded features according to the master version of the feature dictionary; and train the master version of the ANN based on the decoded features using machine learning; wherein the transceiver of the first computing device is configured to receive the local version of the feature dictionary from the second computing device, and wherein the processor of the first computing device is configured to: change the master version of the feature dictionary based on the received local version of the feature dictionary; and decode the received encoded features according to the changed master version of the feature dictionary.
As per claim 13, The system of claim 12,
wherein the transceiver of the first computing device is configured to transmit the trained master version of the ANN to the second computing device.
As per claim 13, The system of claim 12, wherein the transceiver of the first computing device is configured to transmit the changed master version of the feature dictionary to the second computing device.
As per claim 15, The system of claim 14,
wherein the transceiver of the first computing device is configured to transmit the changed master version of the feature dictionary to the second computing device.
As per claim 11, The system of claim 10, wherein the transceiver of the first computing device is configured to transmit the trained master version of the ANN to the second computing device.
As per claim 16, The system of claim 12,
wherein the second computing device comprises: memory configured to: store user data; store a local version of the ANN; and store the local version of the feature dictionary; a processor configured to: extract features from the stored user data; and encode the extracted features according to the local version of the feature dictionary; and a transceiver configured to transmit the encoded features to the first computing device so that the encoded features are decoded by the master version of the feature dictionary and then used as input to train the master version of the ANN using the machine learning.
As per claim 14, The system of claim 10, wherein the second computing device comprises: memory configured to: store user data; store a local version of the ANN; and store the local version of the feature dictionary; a processor configured to: extract features from the stored user data; and encode the extracted features according to the local version of the feature dictionary; and a transceiver configured to transmit the encoded features to the first computing device so that the encoded features are decoded by the master version of the feature dictionary and then used as input to train the master version of the ANN using the machine learning.
As per claim 17, The system of claim 16,
wherein the processor of the second computing device is further configured to: determine whether the extracted features are included in the local version of the feature dictionary; in response to the local version of the feature dictionary including the extracted features, encode the extracted features according to the local version of the feature dictionary; and in response to the local version of the feature dictionary not including the extracted features, change the local version of the feature dictionary based on the extracted features and then encode the extracted features according to the changed local version of the feature dictionary.
As per claim 15, The system of claim 14, wherein the processor of the second computing device is further configured to: determine whether the extracted features are included in the local version of the feature dictionary; in response to the local version of the feature dictionary including the extracted features, encode the extracted features according to the local version of the feature dictionary; and in response to the local version of the feature dictionary not including the extracted features, change the local version of the feature dictionary based on the extracted features and then encode the extracted features according to the changed local version of the feature dictionary.
As per claim 18, The system of claim 17,
wherein the transceiver of the second computing device is configured to transmit the changed local version of the feature dictionary to the first computing device so that the master version of the feature dictionary is changed according to the local version of the feature dictionary.
As per claim 16, The system of claim 15, wherein the transceiver of the second computing device is configured to transmit the changed local version of the feature dictionary to the first computing device so that the master version of the feature dictionary is changed according to the local version of the feature dictionary.
As per claim 19, The system of claim 18,
wherein the transceiver of the second computing device is configured to receive the changed master version of the feature dictionary, and wherein the processor of the second computing device is configured to change the local version of the feature dictionary based on the changed master version of the feature dictionary.
As per claim 17, The system of claim 16, wherein the transceiver of the second computing device is configured to receive the changed master version of the feature dictionary, and wherein the processor of the second computing device is configured to change the local version of the feature dictionary based on the changed master version of the feature dictionary.
As per claim 20, The system of claim 16,
wherein the transceiver of the second computing device is configured to receive the trained master version of the ANN, and wherein the processor of the second computing device is configured to change the local version of the ANN based on the trained master version of the ANN.
As per claim 18, The system of claim 14, wherein the transceiver of the second computing device is configured to receive the trained master version of the ANN, and wherein the processor of the second computing device is configured to change the local version of the ANN based on the trained master version of the ANN.
A. ) The features of claim 7 of the current application that are not present in claim 5 of the Patent No. 12,248,412 are :
“so that the encoded features are decoded by a master version of the feature dictionary hosted by the first computing device and then used as input to train the master version of the ANN using machine learning.”
However, within analogous art, QIANG YANG et al. (NPL Doc: “Federated Machine Learning: Concept and Applications,” January 2019, ACM Transactions on Intelligent Systems and Technology, Vol. 10, No. 2, Article 12,Pages 12:1-12:13) teaches “so that the encoded features are decoded by a master version of the feature dictionary hosted by the first computing device and then used as input to train the master version of the ANN using machine learning.”( Page 12:9- Fig. 3 AND 2.4 Architecture for a Federated Learning System- Participants update their respective model with the decrypted gradients.” AND Page 12:9 – “…• Step 3: A and B compute encrypted gradients and add an additional mask, respectively. B also computes encrypted loss. A and B send encrypted values to C. • Step 4: C decrypts and send the decrypted gradients and loss back to A and B. A and B unmask the gradients and update the model parameters accordingly….” AND the training of the neural network/machine learning of decrypted data taught within Page 12:9- Fig. 4 and Page 12:9- “…Part 2. Encrypted model training. After determining the common entities, we can use these common entities’ data to train the machine-learning model. The training process can be divided into the following four steps (as shown in Figure 4).• Step 1: Collaborator C creates encryption pairs and sends a public key to A and B. • Step 2: A and B encrypt and exchange the intermediate results for gradient and loss calculations. • Step 3: A and B compute encrypted gradients and add an additional mask, respectively. B also computes encrypted loss. A and B send encrypted values to C. • Step 4: C decrypts and send the decrypted gradients and loss back to A and B. A and B unmask the gradients and update the model parameters accordingly. Here, we illustrate the training process using linear regression and homomorphic encryption as an example. To train a linear regression model with gradient descent methods, we need secure computations of its loss and gradients. Assuming learning rate η, regularization parameter λ, dataset…” ) .
The combination of the known elements is achieved by the System and method for Federated Machine Learning: Concept and Applications mentioned by QIANG YANG et al. within the Feature Dictionary For Bandwidth Enhancement taught in claim 5 of Patent No. 12,248,412.
Therefore, the results would have been predictable to one of ordinary skill in the art.
Based on the above findings, it would have been obvious to one of ordinary skill before the
effective filing date of the invention to add the elements taught in QIANG YANG et al. to the system taught in claim 5 of Patent No. 12,248,412 as no more “than the predictable use of prior-art elements according to their established functions.”
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
2. Claims 13,14,15,18,19 and 20 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
3. Claims 13,14,15,18,19 and 20 recite the limitation "the transceiver". There is insufficient antecedent basis for this limitation in the claim.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries set forth in Graham v. John Deere Co., 383 U.S. 1, 148 USPQ 459 (1966), that are applied for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b) (2) (C) for any potential 35 U.S.C. 102(a) (2) prior art against the later invention.
4. Claims 1,2,3,4,5,6,7,11,12,13,14 and 15 are rejected under 35 U.S.C 103 as being patentable over McMahan et al. (USPUB 20190227980) in view of QIANG YANG et al. (NPL Doc: “Federated Machine Learning: Concept and Applications,” January 2019, ACM Transactions on Intelligent Systems and Technology, Vol. 10, No. 2, Article 12,Pages 12:1-12:13).
As per claim 1, McMahan et al. teaches A method, comprising: hosting, by a first computing device, a master version of an artificial neural network (ANN) ( FIG. 1 and FIG. 2 teaches the first computing device ( FIG. 2-SERVER – 210 ) and the hosting of the master version of the neural network ( FIG. 1- SERVER 104 and MODEL-106) further taught within Paragraphs [0127-0131]- “…The one or more server computing device(s) 104 can be configured to access machine-learned model 106, and to provide model 106 to a plurality of client computing devices 102. Model 106 can be, for instance, a linear regression model, logistic regression model, a support vector machine model, a neural network (e.g. convolutional neural network, recurrent neural network, etc.), or other suitable model. I…” AND Paragraph [0133]-“ FIG. 2 depicts an example computing system 200 that can be used to implement the methods and systems of the present disclosure. The system 200 can be implemented using a client-server architecture that includes a server 210 (e.g., one or more server computing devices)…”) ; hosting, by the first computing device, a master version of a feature dictionary ( FIG. 2 teaches GLOBAL PARAMETERS/master version of the feature dictionary /centralized repository of documents stored within the SERVER ( 210) and further taught within Paragraphs [0137-0138]- “…the instructions 216 can be executed by the one or more processors 212 to implement a global updater 220. The global updater 220 can be configured to receive one or more local updates and to determine an updated machine-learned model based at least in part on the local updates….”) ; receiving, by the first computing device, encoded features from a second computing device ( FIG. 1 – FIG. 2 – CLIENT DEVICES / second computing device and encoding further taught within Paragraphs [0030-0031]- “…ach client computing device can then provide the local update to the one or more server computing devices. For example, in some implementations, the local update can be clipped before being provided to the one or more server computing devices, as described herein. In various implementations, the local update can be provided as one or more vectors, matrices, parameters, or other formats, and may be encoded before being provided to the one or more server computing devices….”) ;
McMahan et al. does not explicitly teach decoding, by the first computing device, the received encoded features according to the master version of the feature dictionary; and training, by the first computing device, the master version of the ANN based on the decoded features using machine learning.
However, within analogous art, QIANG YANG et al. teaches decoding, by the first computing device, the received encoded features according to the master version of the feature dictionary ( Page 12:9- Fig. 3 AND 2.4 Architecture for a Federated Learning System- Participants update their respective model with the decrypted gradients.” AND Page 12:9 – “…• Step 3: A and B compute encrypted gradients and add an additional mask, respectively. B also computes encrypted loss. A and B send encrypted values to C. • Step 4: C decrypts and send the decrypted gradients and loss back to A and B. A and B unmask the gradients and update the model parameters accordingly….”) ; and training, by the first computing device, the master version of the ANN based on the decoded features using machine learning ( the training of the neural network/machine learning of decrypted data taught within Page 12:9- Fig. 4 and Page 12:9- “…Part 2. Encrypted model training. After determining the common entities, we can use these common entities’ data to train the machine-learning model. The training process can be divided into the following four steps (as shown in Figure 4).• Step 1: Collaborator C creates encryption pairs and sends a public key to A and B. • Step 2: A and B encrypt and exchange the intermediate results for gradient and loss calculations. • Step 3: A and B compute encrypted gradients and add an additional mask, respectively. B also computes encrypted loss. A and B send encrypted values to C. • Step 4: C decrypts and send the decrypted gradients and loss back to A and B. A and B unmask the gradients and update the model parameters accordingly. Here, we illustrate the training process using linear regression and homomorphic encryption as an example. To train a linear regression model with gradient descent methods, we need secure computations of its loss and gradients. Assuming learning rate η, regularization parameter λ, dataset…” ).
One of ordinary skill in the art would have been motivated to combine the teaching of QIANG YANG et al. within the modified teaching of the Training User-Level Differentially Private Machine-Learned Models mentioned by McMahan et al. because the Federated Machine Learning: Concept and Applications mentioned by QIANG YANG et al. provides a system and method for implementing federated learning within neural network / machine learning infrastructure.
Therefore, it would have been obvious for one in the ordinary skills in the art before the effective filing date of the claimed invention to implement the Federated Machine Learning: Concept and Applications mentioned by QIANG YANG et al. within the modified teaching of the Training User-Level Differentially Private Machine-Learned Models mentioned by McMahan et al. for implementation of a system and method for federated learning within neural network / machine learning infrastructure.
As per claim 2, Combination of McMahan et al. and QIANG YANG et al. teach claim 1,
McMahan et al. teaches comprising transmitting, by the first computing device, the trained master version of the ANN to the second computing device ( Paragraph [0018]- “Federated learning offers several distinct advantages compared to performing learning at a centralized server. For example, information of the model update may be less sensitive than the data itself. Thus, user data that is privacy sensitive can remain at the user's computing device and need not be uploaded to the server. Rather, only the less sensitive model update can be transmitted. As another advantage, federated learning leverages the computational power of a large number of computing devices (e.g., user mobile devices). In addition, in some implementations, the federated learning framework can be modified to allow for the creation and use of personalized models for each user at their respective device” AND FIG. 3- 304 and Paragraph [0149]- “ At (304), the method (300) can include providing, by the one or more server computing devices, the machine-learned model to the selected client computing devices, and at (306), the method (300) can include receiving, by the selected client computing devices, the machine-learned model. In some implementations, the machine-learned model can be, for example, a global set of parameters.”) .
As per claim 3, Combination of McMahan et al. and QIANG YANG et al. teach claim 1,
McMahan et al. teaches comprising receiving, by the first computing device, the local version of the feature dictionary from the second computing device ( FIG. 3- 310- PROVIDE LOCAL UPDATE and 312-316 AND Paragraphs [0151-0153]- “At (310), the method (300) can include providing, by each selected client computing device, the local update to the one or more server computing devices, and at (312), the method (300) can include receiving, by the one or more server computing devices, the local updates. In some implementations, the local updates can be, for example, expressed in one or more matrices, vectors, or other suitable format….”) .
As per claim 4, Combination of McMahan et al. and QIANG YANG et al. teach claim 3,
McMahan et al. teaches comprising changing, by the first computing device, the master version of the feature dictionary based on the received local version of the feature dictionary ( FIG. 3- 316- DETERMINE UPDATED MACHINE-LEARNED MODEL and Paragraph [0156]- “At (316), the method (300) can include determining an updated machine-learned model based at least in part on the bounded-sensitivity data-weighted average of the local updates. For example, in some implementations, the bounded-sensitivity data-weighted average of the local updates can be added to the machine-learned model to determine the updated machine-learned model. …” ) .
As per claim 5, Combination of McMahan et al. and QIANG YANG et al. teach claim 4,
Within analogous art, QIANG YANG et al. teaches wherein the decoding comprises decoding the encoded features according to the changed master version of the feature dictionary ( Decryption of the encrypted data taught within Page 12:14- “…we can exploit a federated-learning framework. In particular, we can use the encryption mechanism of federated learning and encrypt the user list at each party and then take the intersection of the encrypted list in the federation. The decryption of the final result gives the list of multiparty borrowers without exposing the other “good” users to the other party. As we will see below, this operation corresponds to the vertical federated learning framework…”) .
As per claim 6, Combination of McMahan et al. and QIANG YANG et al. teach claim 1,
McMahan et al. teaches comprising transmitting, by the first computing device, the changed master version of the feature dictionary to the second computing device ( FIG. 3- 318- PROVIDE UPDATED MACHINE LEARNED MODEL and Paragraphs [0156-0157]- “ … the method (300) can include providing, by the one or more server computing devices, the updated machine-learned model to one or more client computing devices, and at (320), the method (300) can include receiving, by the one or more client computing devices, the updated machine-learned model…”) .
As per claim 7, McMahan et al. teaches A method, comprising: hosting, by a second computing device, a local version of an artificial neural network (ANN) ( FIG. 1 and FIG. 2 teaches the first computing device ( FIG. 2-SERVER – 210 ) and the hosting of the master version of the neural network ( FIG. 1- SERVER 104 and MODEL-106) further taught within Paragraphs [0127-0131]- “…The one or more server computing device(s) 104 can be configured to access machine-learned model 106, and to provide model 106 to a plurality of client computing devices 102. Model 106 can be, for instance, a linear regression model, logistic regression model, a support vector machine model, a neural network (e.g. convolutional neural network, recurrent neural network, etc.), or other suitable model. I…” AND Paragraph [0133]-“ FIG. 2 depicts an example computing system 200 that can be used to implement the methods and systems of the present disclosure. The system 200 can be implemented using a client-server architecture that includes a server 210 (e.g., one or more server computing devices)…”) ; hosting, by the second computing device, a local version of a feature dictionary ( FIG. 2 teaches GLOBAL PARAMETERS/master version of the feature dictionary /centralized repository of documents stored within the SERVER ( 210) and further taught within Paragraphs [0137-0138]- “…the instructions 216 can be executed by the one or more processors 212 to implement a global updater 220. The global updater 220 can be configured to receive one or more local updates and to determine an updated machine-learned model based at least in part on the local updates….”) ; extracting, by the second computing device, features from user data hosted by the second computing device ( FIG. 3 & 5 and Paragraphs [0127-0128]- “ Client computing devices 102 can each be configured to determine one or more local updates associated with model 106 based at least in part on training data 108. For instance, training data 108 can be data that is respectively stored locally on the client devices 102. The training data 108 can include audio files, image files, video files, a typing history, location history, and/or various other suitable data. In some implementations, the training data 108 can be any data derived through a user interaction with a client computing device 102….”) ; encoding, by the second computing device, the extracted features according to the local version of the feature dictionary( FIG. 1 – FIG. 2 – CLIENT DEVICES / second computing device and encoding further taught within Paragraphs [0030-0031]- “…ach client computing device can then provide the local update to the one or more server computing devices. For example, in some implementations, the local update can be clipped before being provided to the one or more server computing devices, as described herein. In various implementations, the local update can be provided as one or more vectors, matrices, parameters, or other formats, and may be encoded before being provided to the one or more server computing devices….”) ;
McMahan et al. does not explicitly teach transmitting, by the second computing device, the encoded features to a first computing device that hosts a master version of the ANN so that the encoded features are decoded by a master version of the feature dictionary hosted by the first computing device and then used as input to train the master version of the ANN using machine learning.
However, within analogous art, QIANG YANG et al. teaches transmitting, by the second computing device, the encoded features to a first computing device that hosts a master version of the ANN so that the encoded features are decoded by a master version of the feature dictionary hosted by the first computing device and then used as input to train the master version of the ANN using machine learning( Page 12:9- Fig. 3 AND 2.4 Architecture for a Federated Learning System- Participants update their respective model with the decrypted gradients.” AND Page 12:9 – “…• Step 3: A and B compute encrypted gradients and add an additional mask, respectively. B also computes encrypted loss. A and B send encrypted values to C. • Step 4: C decrypts and send the decrypted gradients and loss back to A and B. A and B unmask the gradients and update the model parameters accordingly….” AND the training of the neural network/machine learning of decrypted data taught within Page 12:9- Fig. 4 and Page 12:9- “…Part 2. Encrypted model training. After determining the common entities, we can use these common entities’ data to train the machine-learning model. The training process can be divided into the following four steps (as shown in Figure 4).• Step 1: Collaborator C creates encryption pairs and sends a public key to A and B. • Step 2: A and B encrypt and exchange the intermediate results for gradient and loss calculations. • Step 3: A and B compute encrypted gradients and add an additional mask, respectively. B also computes encrypted loss. A and B send encrypted values to C. • Step 4: C decrypts and send the decrypted gradients and loss back to A and B. A and B unmask the gradients and update the model parameters accordingly. Here, we illustrate the training process using linear regression and homomorphic encryption as an example. To train a linear regression model with gradient descent methods, we need secure computations of its loss and gradients. Assuming learning rate η, regularization parameter λ, dataset…” ) .
One of ordinary skill in the art would have been motivated to combine the teaching of QIANG YANG et al. within the modified teaching of the Training User-Level Differentially Private Machine-Learned Models mentioned by McMahan et al. because the Federated Machine Learning: Concept and Applications mentioned by QIANG YANG et al. provides a system and method for implementing federated learning within neural network / machine learning infrastructure.
Therefore, it would have been obvious for one in the ordinary skills in the art before the effective filing date of the claimed invention to implement the Federated Machine Learning: Concept and Applications mentioned by QIANG YANG et al. within the modified teaching of the Training User-Level Differentially Private Machine-Learned Models mentioned by McMahan et al. for implementation of a system and method for federated learning within neural network / machine learning infrastructure.
As per claim 11, Combination of McMahan et al. and QIANG YANG et al. teach claim 7,
Combination of McMahan et al. and QIANG YANG et al. teaches comprising: receiving the trained master version of the ANN; and changing the local version of the ANN based on the trained master version of the ANN.
Within analogous art, Di Pietro et al. teaches comprising: receiving the trained master version of the ANN ( FIG. 6C showing the computer Device ( Router – FAR-1) training machine learning further taught within Paragraphs [0041-0042] and the feature being distributed to the other device within Nodes (11) shown FIG. 8E and further taught within Paragraphs [0071-0072]) ; and changing the local version of the ANN based on the trained master version of the ANN ( the LM ( learning machine) that processes the ANN is mentioned to route the feature to the nodes taught within FIG. 9 and Paragraph [0117]) .
As per claim 12, McMahan et al. teaches A system, comprising: memory configured to: store a master version of an artificial neural network (ANN) ( FIG. 2- SERVER – MEMORY and Paragraphs [0134-0136]- “…The system 200 includes a server 210, such as a web server. The server 210 can be implemented using any suitable computing device(s). The server 210 can have one or more processors 212 and one or more memory devices 214. The server 210 can be implemented using one server device or a plurality of server devices….”) ; and store a master version of a feature dictionary ( FIG. 2- 218- GLOBAL PARAMETERS DATA) ;
McMahan et al. does not explicitly teach a processor configured to: decode encoded features, received from a second computing device, according to the master version of the feature dictionary; and train the master version of the ANN based on the decoded features using machine learning.
However, within analogous art, QIANG YANG et al. teaches a processor configured to: decode encoded features, received from a second computing device, according to the master version of the feature dictionary ( Page 12:9- Fig. 3 AND 2.4 Architecture for a Federated Learning System- Participants update their respective model with the decrypted gradients.” AND Page 12:9 – “…• Step 3: A and B compute encrypted gradients and add an additional mask, respectively. B also computes encrypted loss. A and B send encrypted values to C. • Step 4: C decrypts and send the decrypted gradients and loss back to A and B. A and B unmask the gradients and update the model parameters accordingly….”) ; and train the master version of the ANN based on the decoded features using machine learning ( the training of the neural network/machine learning of decrypted data taught within Page 12:9- Fig. 4 and Page 12:9- “…Part 2. Encrypted model training. After determining the common entities, we can use these common entities’ data to train the machine-learning model. The training process can be divided into the following four steps (as shown in Figure 4).• Step 1: Collaborator C creates encryption pairs and sends a public key to A and B. • Step 2: A and B encrypt and exchange the intermediate results for gradient and loss calculations. • Step 3: A and B compute encrypted gradients and add an additional mask, respectively. B also computes encrypted loss. A and B send encrypted values to C. • Step 4: C decrypts and send the decrypted gradients and loss back to A and B. A and B unmask the gradients and update the model parameters accordingly. Here, we illustrate the training process using linear regression and homomorphic encryption as an example. To train a linear regression model with gradient descent methods, we need secure computations of its loss and gradients. Assuming learning rate η, regularization parameter λ, dataset…” ) .
One of ordinary skill in the art would have been motivated to combine the teaching of QIANG YANG et al. within the modified teaching of the Training User-Level Differentially Private Machine-Learned Models mentioned by McMahan et al. because the Federated Machine Learning: Concept and Applications mentioned by QIANG YANG et al. provides a system and method for implementing federated learning within neural network / machine learning infrastructure.
Therefore, it would have been obvious for one in the ordinary skills in the art before the effective filing date of the claimed invention to implement the Federated Machine Learning: Concept and Applications mentioned by QIANG YANG et al. within the modified teaching of the Training User-Level Differentially Private Machine-Learned Models mentioned by McMahan et al. for implementation of a system and method for federated learning within neural network / machine learning infrastructure.
As per claim 13, Combination of McMahan et al. and QIANG YANG et al. teach claim 12,
McMahan et al. teaches wherein the transceiver of the first computing device is configured to transmit the trained master version of the ANN to the second computing device ( Paragraph [0135]- “…server 210 can also include a network interface used to communicate with one or more client devices 230 over the network 242. The network interface can include any suitable components for interfacing with one more networks, including for example, transmitters, receivers, ports, controllers, antennas, or other suitable components….”) .
As per claim 14, Combination of McMahan et al. and QIANG YANG et al. teach claim 12,
McMahan et al. teaches wherein the transceiver of the first computing device is configured to receive the local version of the feature dictionary from the second computing device ( FIG. 3 and 4 and Paragraph [0031]- “…The one or more server computing devices can then receive the local updates from the selected client computing devices, and determine a differentially private aggregate of the local updates. For example, in some implementations, a bounded-sensitivity data-weighted average of the local updates can be determined…” AND Paragraph [0132]- “… the one or more server computing devices 104 can receive each local update from client device 102, and can determine differentially private aggregate of the local updates. For example, in some implementations, a bounded-sensitivity data-weighted average of the local updates can be determined. Further, the one or more server computing devices 104 can determine an updated machine-learned model based at least in part on the differentially private aggregate of the local updates….”) , and wherein the processor of the first computing device is configured to: change the master version of the feature dictionary based on the received local version of the feature dictionary ( Update of Machine learned model data received from Client computing device within the Server computing device taught within FIG. 3 AND Paragraphs [0151-0152]- “…include determining a differentially private aggregate of the local updates. For example, in some implementations, the differentially private aggregate can be a bounded-sensitivity data-weighted average of the local updates can be determined. …”) ; and decode the received encoded features according to the changed master version of the feature dictionary( Paragraphs [0030-0031]- “…the local update can be clipped before being provided to the one or more server computing devices, as described herein. In various implementations, the local update can be provided as one or more vectors, matrices, parameters, or other formats, and may be encoded before being provided to the one or more server computing devices… server computing devices can then receive the local updates from the selected client computing devices, and determine a differentially private aggregate of the local updates. For example, in some implementations, a bounded-sensitivity data-weighted average of the local updates can be determined…”) .
As per claim 15, Combination of McMahan et al. and QIANG YANG et al. teach claim 12,
McMahan et al. teaches wherein the transceiver of the first computing device is configured to transmit the changed master version of the feature dictionary to the second computing device
( FIG. 3- 318- PROVIDE UPDATED MACHINE LEARNED MODEL and Paragraphs [0156-0157]- “ … the method (300) can include providing, by the one or more server computing devices, the updated machine-learned model to one or more client computing devices, and at (320), the method (300) can include receiving, by the one or more client computing devices, the updated machine-learned model…”) .
4. Claim 11 is rejected under 35 U.S.C 103 as being patentable over McMahan et al. (USPUB 20190227980) in view of QIANG YANG et al. (NPL Doc: “Federated Machine Learning: Concept and Applications,” January 2019, ACM Transactions on Intelligent Systems and Technology, Vol. 10, No. 2, Article 12,Pages 12:1-12:13) in further view of Di Pietro et al. (USPUB 20150195146).
As per claim 11, Combination of McMahan et al. and QIANG YANG et al. teach claim 7,
Combination of McMahan et al. and QIANG YANG et al. teaches comprising: receiving the trained master version of the ANN; and changing the local version of the ANN based on the trained master version of the ANN.
Within analogous art, Di Pietro et al. teaches comprising: receiving the trained master version of the ANN ( FIG. 6C showing the computer Device ( Router – FAR-1) training machine learning further taught within Paragraphs [0041-0042] and the feature being distributed to the other device within Nodes (11) shown FIG. 8E and further taught within Paragraphs [0071-0072]) ; and changing the local version of the ANN based on the trained master version of the ANN ( the LM ( learning machine) that processes the ANN is mentioned to route the feature to the nodes taught within FIG. 9 and Paragraph [0117]) .
One of ordinary skill in the art would have been motivated to combine the teaching of Di Pietro et al. within the combined modified teaching of the Federated Machine Learning: Concept and Applications mentioned by QIANG YANG et al. and the Training User-Level Differentially Private Machine-Learned Models mentioned by McMahan et al. because the Feature aggregation in a computer network mentioned by Di Pietro et al.
provides a system and method for implementing aggregated data processing of learning machines within computer networks.
Therefore, it would have been obvious for one in the ordinary skills in the art before the effective filing date of the claimed invention to implement the Feature aggregation in a computer network mentioned by Di Pietro et al. within the combined modified teaching of the Federated Machine Learning: Concept and Applications mentioned by QIANG YANG et al. and the Training User-Level Differentially Private Machine-Learned Models mentioned by McMahan et al. for implementation of a system and method for aggregated data processing of learning machines within computer networks.
It is noted that any citations to specific, pages, columns, lines, or figures in the prior art references and any interpretation of the reference should not be considered to be limiting in any way. A reference is relevant for all it contains and may be relied upon for all that it would have reasonably suggested to one having ordinary skill in the art. See MPEP 2123.
Examiner’s Notes
5. The Examiner acknowledges the following prior arts below as pertinent to the current applications claim limitations and inventive concept, although the following prior arts shown below were not relied upon to address the limitations within the claim , they are analogous art mentioning the inventive concept key points on (Federated Distribution , neural networking/ machine learning, central/master data/features , servers – clients communication devices etc.).
1) Micah J. Sheller et al.,"Multi-institutional Deep Learning Modeling Without Sharing Patient Data: A Feasibility Study on Brain Tumor Segmentation," 26th January 2019, Lecture Notes in Computer Science ((LNIP,volume 11383), Pages 92-99
2) Felix Sattler et al.,"Robust and Communication-Efficient Federated Learning From Non-i.i.d. Data," 1st November 2019, IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, VOL. 31, NO. 9, SEPTEMBER 2020,Pages 3400-3410.
3) Adrian Nilsson et al.,"A Performance Evaluation of Federated Learning Algorithms,"DIDL ’18, December 10–11, 2018, Rennes, France,Pages 1-6.
4) Davy Preuveneers et al.,"Chained Anomaly Detection Models for Federated Learning: An Intrusion Detection Case Study," 18th December 2018, Appl. Sci. 2018, 8, 2663; doi:10.3390,Pages 1-13.
5) H. Brendan McMahan et al.,"Communication-Efficient Learning of Deep Networks from Decentralized Data," February 2016, Proceedings of the 20th International Conference on Artificial Intelligence and Statistics (AISTATS) 2017, Fort Lauderdale, Florida, USA. JMLR: W&CP volume 54,Pages 1-8.
6) Seyed Ali Osia et al.,"Deep Private-Feature Extraction,28th February 2018, arXiv:1802.03151v2,Pages 1-8.
7) Micah J. Sheller et al. ,"Multi-institutional Deep Learning Modeling Without Sharing Patient Data: A Feasibility Study on Brain TumorSegmentation," 26th January 2019, Lecture Notes in Computer Science ((LNIP,volume 11383)) Pages 92-100.
8) Heydari (USPUB 20220011834)
9) METTLER MAY (CN 201680064109)
10) Malik et al. (USPUB 20210117780)
11) Mohassel et al. (USPUB 20200242466)
12) Cella et al. (USPUB 20190339688 )
13) Cella et al. (USPUB 20190324439 )
14) Arditi (USPUB 20190147331 )
15) Vasseur et al. (USPUB 20150193694)
16) Cruz Mota et al. (USPUB 20150193695 )
17) Vasseur et al. (USPUB 20150193697 )
18) Drew et al. (USPUB 20140344195 )
Conclusion
6. The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Refer to PTO-892, Notice of Reference Cited for a listing of analogous art.
7. Any inquiry concerning this communication or earlier communications from the examiner should be directed to OMAR S ISMAIL whose telephone number is (571)272-9799 and Fax # is (571)273-9799. The examiner can normally be reached on M-F 9:00am-6:00pm.
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, David C. Payne can be reached on (571) 272-3024. The fax phone number for the organization where this application or proceeding is assigned is (571)273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free)? If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/OMAR S ISMAIL/
Primary Examiner, Art Unit 2635
