DETAILED ACTION
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 .
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception without significantly more.
Regarding Claim 1:
Step 1 – Is the claim to a process, machine, manufacture, or composition of matter?
Yes
Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites the abstract ideas of:
neural network training information to be aggregated based, at least in part, on a relative contribution of the neural network training data to one or more performance metrics of the neural network. This limitation is directed to the abstract idea of a mathematical concepts, as aggregating the information based on the contribution is analogous to a mathematical calculation (see MPEP 2106.04(a)(2) I. C.)
Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application? – No, there are no additional elements that integrate the judicial exception into a practical application.
A processor comprising: one or more circuits. This limitation recites generic computer components such as processor and circuits, which invokes a computer merely as a tool for performing an existing process [see MPEP 2106.05(f)(2)] and therefore fails to integrate the exception into a practical application.
Step 2B – Does the claim recite any additional elements that amount to significantly more than the judicial exception? – No, there are no additional elements that amount to significantly more than the judicial exception.
A processor comprising: one or more circuits. This limitation invokes a computer merely as a tool for performing an existing process [see MPEP 2106.05(f)(2)] and therefore fails to amount to significantly more than the judicial exception.
Step 2A Prong Two and Step 2B:
Thus, the judicial exception is not integrated into a practical application (see MPEP 2106.04(d) I.), failing step 2A prong 2. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception under step 2B. The claim is ineligible.
Regarding Claim 2:
Step 1 – Is the claim to a process, machine, manufacture, or composition of matter?
Yes
Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites the abstract ideas of:
The processor of claim 1, wherein compute an importance value based at least in part, on the contribution of the neural network training data and one or more performance metrics of the neural network; This limitation is directed to the abstract idea of a mathematical concepts, as computing a value with the predetermined degree is analogous to a mathematical calculation (see MPEP 2106.04(a)(2) I. C.)
perform weighted averaging of neural network training information using the importance value. This limitation is directed to the abstract idea of a mathematical concepts, as performing weighted averaging with the predetermined degree is analogous to a mathematical calculation (see MPEP 2106.04(a)(2) I. C.)
Regarding Claim 3:
Step 1 – Is the claim to a process, machine, manufacture, or composition of matter?
Yes
Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Claim 3 does not recite abstract ideas other than the ones recited at claim 1.
Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application? – No, there are no additional elements that integrate the judicial exception into a practical application.
The processor of claim 1, wherein at least one portion of the neural network training data is to be used to generate at least one portion of the neural network training information. This limitation recites description of the network training data is used on generating the network training information. Therefore, this limitation amounts to merely indicating a field of use or technological environment [see MPEP 2106.05(h)] and fails to integrate the judicial exception into a practical application.
Step 2B – Does the claim recite any additional elements that amount to significantly more than the judicial exception? – No, there are no additional elements that amount to significantly more than the judicial exception.
The processor of claim 1, wherein at least one portion of the neural network training data is to be used to generate at least one portion of the neural network training information. This limitation recites description of the network training data is used on generating the network training information. Therefore, this limitation amounts to merely indicating a field of use or technological environment [see MPEP 2106.05(h)] and therefore fails to amount to significantly more than the judicial exception.
Step 2A Prong Two and Step 2B:
Thus, the judicial exception is not integrated into a practical application (see MPEP 2106.04(d) I.), failing step 2A prong 2. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception under step 2B. The claim is ineligible.
Regarding Claim 4:
Step 1 – Is the claim to a process, machine, manufacture, or composition of matter?
Yes
Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
The processor of claim 1, wherein the contribution of the neural network training data is estimated based, at least in part, on first gradients that are generated using one portion of the training information and second gradients that are generated using other portions of the training information. This limitation is directed to the abstract idea of a mental process as estimating the contribution is analogous to evaluation and judgment, which can be performed by the human mind, or by a human using pen and paper (see MPEP 2106.04(a)(2) III. C.). Also, this limitation is directed to the abstract idea of a mathematical concepts, as generating gradients is analogous to a mathematical calculation (see MPEP 2106.04(a)(2) I. C.)
Regarding Claim 5:
Step 1 – Is the claim to a process, machine, manufacture, or composition of matter?
Yes
Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Claim 5 does not recite abstract ideas other than the ones recited at claim 1.
Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application? – No, there are no additional elements that integrate the judicial exception into a practical application.
The processor of claim 1, wherein the training information is generated by a plurality of computing devices. This limitation invokes a computer merely as a tool for performing an existing process [see MPEP 2106.05(f)(2)] and therefore fails to amount to significantly more than the judicial exception.
Step 2B – Does the claim recite any additional elements that amount to significantly more than the judicial exception? – No, there are no additional elements that amount to significantly more than the judicial exception.
The processor of claim 1, wherein the training information is generated by a plurality of computing devices. This limitation invokes a computer merely as a tool to train a model for performing an existing process [see MPEP 2106.05(f)(2)] and therefore fails to amount to significantly more than the judicial exception.
Step 2A Prong Two and Step 2B:
Thus, the judicial exception is not integrated into a practical application (see MPEP 2106.04(d) I.), failing step 2A prong 2. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception under step 2B. The claim is ineligible.
Regarding Claim 6:
Step 1 – Is the claim to a process, machine, manufacture, or composition of matter?
Yes
Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
The processor of claim 1, wherein the neural network training information comprises weights of a plurality of neural networks. This limitation is directed to the abstract idea of a mathematical concept, as training information comprises weights, and the weights can be considered as a value under BRI. Thus, the limitation is analogous to mathematical relationships (see MPEP 2106.04(a)(2) I. C.).
Regarding Claim 7:
Step 1 – Is the claim to a process, machine, manufacture, or composition of matter?
Yes
Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Claim 7 does not recite abstract ideas other than the ones recited at claim 1.
Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application? – No, there are no additional elements that integrate the judicial exception into a practical application.
The processor of claim 1, wherein the one or more circuits are further to send, to a plurality of computing devices, aggregated neural network training information. This limitation recites an insignificant extra solution activity, as outputting information, under BRI, is mere data outputting per MPEP 2106.05(g)(3).
Step 2B – Does the claim recite any additional elements that amount to significantly more than the judicial exception? – No, there are no additional elements that amount to significantly more than the judicial exception.
The processor of claim 1, wherein the one or more circuits are further to send, to a plurality of computing devices, aggregated neural network training information. This limitation is directed to outputting/transmitting data, which the courts have recognized as well-understood, routine, conventional activity when they are claimed at a high level of generality or as insignificant extra-solution activity per MPEP 2106.05(d) II.i and therefore fails to amount to significantly more than the judicial exception.
Step 2A Prong Two and Step 2B:
Thus, the judicial exception is not integrated into a practical application (see MPEP 2106.04(d) I.), failing step 2A prong 2. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception under step 2B. The claim is ineligible.
Regarding claims 8 – 9:
Claim 8 recites analogous limitations to claim 1 (respectively) and therefore they are rejected on the same grounds as claims 1.
Claim 9 recites analogous limitations to claim 3 (respectively) and therefore they are rejected on the same grounds as claims 3.
Regarding Claim 10:
Step 1 – Is the claim to a process, machine, manufacture, or composition of matter?
Yes
Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites the abstract ideas of:
The method of claim 8, further comprising causing a computer device to estimate the relative contribution based, at least in part, on determining similarities between first gradients that are generated using one portion of the neural network information and second gradients that are generated using other portions of the neural network information. This limitation is directed to the abstract idea of a mental process as estimating the contribution, and determining similarities is(are) analogous to evaluation and judgment, which can be performed by the human mind, or by a human using pen and paper (see MPEP 2106.04(a)(2) III. C.). Also, this limitation is directed to the abstract idea of a mathematical concepts, as determining similarities and generating gradients is(are) analogous to a mathematical calculation (see MPEP 2106.04(a)(2) I. C.).
Regarding Claim 11:
Step 1 – Is the claim to a process, machine, manufacture, or composition of matter?
Yes
Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Claim 11 does not recite abstract ideas other than the ones recited at claim 8.
Step 2A – Prong 2 – Does the claim recite additional elements that integrate the judicial exception into a practical application? – No, there are no additional elements that integrate the judicial exception into a practical application.
The method of claim 8, further comprising receiving at least one of an estimate of the relative contribution of the neural network training data, one or more performance metrics of the neural network, or neural network training information from a plurality of computing devices. This limitation recites as an insignificant extra solution activity, as receiving contribution, performance or training information from devices, under BRI, is mere data gathering per MPEP 2106.05(g)(3).
Step 2B – Does the claim recite any additional elements that amount to significantly more than the judicial exception? – No, there are no additional elements that amount to significantly more than the judicial exception.
The method of claim 8, further comprising receiving at least one of an estimate of the relative contribution of the neural network training data, one or more performance metrics of the neural network, or neural network training information from a plurality of computing devices. This limitation is directed to receiving data, which the courts have recognized as well-understood, routine, conventional activity when they are claimed at a high level of generality or as insignificant extra-solution activity [see MPEP 2106.05(d) II. i] and therefore fails to amount to significantly more than the judicial exception.
Step 2A Prong Two and Step 2B:
Thus, the judicial exception is not integrated into a practical application (see MPEP 2106.04(d) I.), failing step 2A prong 2. The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception under step 2B. The claim is ineligible.
Regarding claims 12:
Claim 12 recites analogous limitations to claim 7 (respectively) and therefore they are rejected on the same grounds as claims 7.
Regarding Claim 13:
Step 1 – Is the claim to a process, machine, manufacture, or composition of matter?
Yes
Step 2A – Prong 1 – Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, the claim recites the abstract ideas of:
The method of claim 8, wherein the performance metrics of the neural network are based, at least in part, on comparing an output of the neural network using a portion of neural network training data with ground truth of the portion. This limitation is directed to the abstract idea of a mental process as comparing an output with ground truth analogous to evaluation and judgment, which can be performed by the human mind, or by a human using pen and paper (see MPEP 2106.04(a)(2) III. C.).
Regarding claims 14- 20
Claims 14-20 recites analogous limitations to claims 1-7 (respectively) and therefore they are rejected on the same grounds as claims 1-7.
Claim Rejections - 35 USC § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claim(s) 1 – 3, 5 – 9, 11 – 16, and 18 -20 are rejected under 35 U.S.C. 102(a) (1) as being anticipated by Giuseppi (NPL, An Adaptive Model Averaging Procedure for Federated Learning (AdaFed) dated on 12/06/2022, by Giuseppi et al - hereinafter Giuseppi).
Referring to Claim 1, Giuseppi teaches:
A processor comprising: one or more circuits to cause neural network training information to be aggregated based, at least in part, on a relative contribution of the neural network training data to one or more performance metrics of the neural network. See Giuseppi at [Page 540 III. (1)]:” Weighted Model Averaging (lines 4-9): When the server carries-out the model averaging procedure and it gathers the clients’ models, a preliminary evaluation of all the models is conducted on a dataset available to the server. According to the models’ performances, a different weight is given to each model for the model averaging, so that better performing models are given more focus.” Examiner interprets the server gathering the work of the client’s model and weighting more focus on better performing models as equivalent as the computer device causing network aggregating training information based on the contribution and performance.
Referring to Claim 2, Giuseppi teaches:
compute an importance value based at least in part, on the contribution of the neural network training data and one or more performance metrics of the neural network; See Giuseppi at [Page 541, under the figure Algorithm 1]: “The performance can be evaluated, generally speaking, with any indicator that captures the quality of the solution proposed by the model for the given task (e.g., accuracy for classification tasks).” Examiner interprets evaluating the performance based on the accuracy for the tasks as equivalent as computing an importance value based on the contribution.
perform weighted averaging of neural network training information using the importance value. See Giuseppi at [Page 542, Para 3]: “The same idea is used in the Weighted Model Average step. Note that, in principle, the weight 𝑝𝑖 of the i-th client can be computed via the performance of its model over the server test set with a different metric than the one used for the Adaptive Loss step, so instead of the F1-score one may utilise for example the model accuracy or the diagnostic odds ratio, depending on the specific use case.” Examiner interprets computing weight pi via the performance in the weighted model average step as equivalent as performing weighted averaging.
Referring to Claim 3, Giuseppi teaches:
at least one portion of the neural network training data is to be used to generate at least one portion of the neural network training information. See Giuseppi at [page 539, Introduction]:” Specifically, in FL the server’s model is updated at every communication round by averaging the models of the federated clients, trained on their locally available data, while AdaFed proposes a two-step procedure to improve both the model averaging and the local training processes by i) dynamically weighting the client models contributions to the federation based on their performance, and ii) adapting the federated loss function depending the global model performance at each communication round.” Examiner using local data to train the client’s model and updating the server’s model as equivalent as using at least one portion of training data to generating the neural network training information.
Referring to Claim 5, Giuseppi teaches:
the training information is generated by a plurality of computing devices. See Giuseppi at [ Page 539, Abstract]:” FL envisages that distributed clients cooperate to learn a model without any data exchange, in favor of a model averaging procedure that is coordinated by a server.” Examiner interprets distributed clients as equivalent as a plurality of computing devices.
Referring to Claim 6, Giuseppi teaches:
the neural network training information comprises weights of a plurality of neural networks. See Giuseppi at [Page 539, Introduction, Para 3]:” Specifically, in FL the server’s model is updated at every communication round by averaging the models of the federated clients, trained on their locally available data, while AdaFed proposes a two-step procedure to improve both the model averaging and the local training processes by i) dynamically weighting the client models contributions to the federation based on their performance, and ii) adapting the federated loss function depending the global model performance at each communication round.” Examiner interprets the dynamically weighting the client models as equivalent as the weights of a plurality of networks.
Referring to Claim 7, Giuseppi teaches:
the one or more circuits are further to send, to a plurality of computing devices, aggregated neural network training information.” See Giuseppi at [Page 540]:” The concept of FL was introduced in 2016 by the authors of [7] and, in its original formulation [6], [7], it was presented as a solution to specifically address the collaboration among a group of smartphones by means of an iterative model-averaging procedure. According to which, local model updates, independently computed by the smartphones, were gathered and averaged by a centralised server that then propagated the updated global model in the network, as depicted in Fig. 1.” Examiner interprets the server propagating the updated global model as equivalent as sending aggregated network training information.
Referring to Claim 8, the claim is rejected on the same basis as claim 1, mutatis mutandis, since they are analogous claims.
Referring to Claim 9, the claim is rejected on the same basis as claim 3, mutatis mutandis, since they are analogous claims.
Referring to Claim 11, the claim is rejected on the same basis as claim 4, mutatis mutandis, since they are analogous claims.
Referring to Claim 12, the claim is rejected on the same basis as claim 7, mutatis mutandis, since they are analogous claims.
Referring to Claim 13, Giuseppi teaches:
the performance metrics of the neural network are based, at least in part, on comparing an output of the neural network using a portion of neural network training data with ground truth of the portion. See Giuseppi at [Page 3, A. Application:]:” Multi-class classification is a typical example found in computer vision and in general Machine Learning applications. The most common loss function for this kind of problems is the so-called categorical cross-entropy:
PNG
media_image1.png
134
638
media_image1.png
Greyscale
where 𝑥𝑚 ∈ 𝑋 and 𝑦𝑚 ∈ 𝑌 are the m-th data sample and label, respectively, in the dataset (𝑋, 𝑌) , M and C are respectively the number of data samples and classes, ycm and ycm denote the c-th component of the vectors ym and 𝑦̂m and are respectively the true and predicted labels for the sample m regarding class c. Note that 𝑦̂cm is typically produced, for single-label problems, by a deep neural network with a softmax output activation function and can be interpreted as the probability of correctness for the given label.” Examiner interprets, in the cross-entropy function (1), dataset y as equivalent as truth ground data and dataset 𝑦̂ as equivalent as the output of the network. It is well-known in the art that the categorical cross-entropy function is a standard mathematical tool used to measure the dissimilarity or ‘distance’ between two probability distributions. In the context of machine learning, this junction inherently performs the step of comparing the model’s output (predicted distribution) with the ground truth (actual distribution) to quantify the error or performance. Thus, Giuseppi teaches the limitation.
Referring to Claims 14- 16, these claim(s) is/are rejected on the same basis as claim 1 - 3, mutatis mutandis, since they are analogous claims.
Referring to Claims 18 - 20, these claim(s) is/are rejected on the same basis as claim 5 - 7, mutatis mutandis, since they are analogous claims.
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.
The factual inquiries 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 non-obviousness.
Claim(s) 4, 10 and 17 is/are rejected under 35 U.S.C. 103 as being unpatentable over Giuseppi in view of Wu (NPL, Fast-Convergent Federated Learning with Adaptive Weighting, dated on 04/06/2021, by Wu et al - hereinafter Wu).
Referring to Claim 4, Giuseppi teaches the device product of claim 1, however, it fails to teach wherein the contribution of the neural network training data is estimated based, at least in part, on first gradients that are generated using one portion of the training information and second gradients that are generated using other portions of the training information.
Wu teaches, in an analogous system:
the contribution of the neural network training data is estimated based, at least in part, on first gradients that are generated using one portion of the training information and second gradients that are generated using other portions of the training information. See Wu at [Page 5, Para 6]:” we measure the contribution of participating nodes based on the correlation between local gradient and global gradient. Particularly, we quantify the contribution of each node at each global round based on angle Θi (t), that is defined as:” Examiner interprets measuring the contribution based on correlation between local gradient and global gradient as equivalent as estimating the contribution on the first gradients and second gradients.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Giuseppi with the above teachings of Wu by aggregating training information based on a relative contribution of the neural network training data, as taught by Giuseppi, wherein the contribution of the neural network training data is estimated based, at least in part, on first gradients that are generated using one portion of the training information and second gradients that are generated using other portions of the training information, as taught by Wu. The modification would have been obvious because one of ordinary skill in the art would be motivated to provide a technical solution and theoretical basis for using gradient correlation to evaluate data contribution and optimize training efficiency, as suggested by Wu. See at [Abstract]: “Its superiority over the commonly adopted Federated Averaging (FedAvg) is verified both theoretically and experimentally. With extensive experiments performed in Pytorch and PySyft, we show that FL training with FedAdp can reduce the number of communication rounds by up to 54.1% on MNIST dataset and up to 45.4% on FashionMNIST dataset, as compared to FedAvg algorithm.”
Referring to Claims 10 and Claim 17, these claims are rejected on the same basis as claim 4, mutatis mutandis, since they are analogous claims.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JIAYUE MA whose telephone number is (571)272-9658. The examiner can normally be reached between 9 am to 5 pm.
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 Yi can be reached at (571) 270-7519. 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.
/Jiayue Ma/
Examiner, Art Unit 2126
/DAVID YI/Supervisory Patent Examiner, Art Unit 2126
Prosecution Insights