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 .
Examiner’s Note
Regarding 35 USC § 112(f) and 35 USC § 112, have been withdrawn in light of the instant amendments to the claims.
Response to Argument
Applicant's arguments filed 02/25/2026 ("Arguments/Remarks") have been fully considered but they are not persuasive.
In Remarks/Arguments (pg. 7) Applicant contends: “First, regarding Step 2A, Prong One, Applicant asserts that the amended claims are directed to an improvement in computer functionality…”
Regarding the above argument, the Examiner respectfully disagrees with Applicant’s assertion that the amended claims are directed to an improvement in computer functionality. The amended limitations do not integrate the judicial exception into practical application. The recited steps of predicting a time period for training an AI model used on user behavior and heuristic data and scheduling the training accordingly, merely use information to determine when a process should occur. Although the Applicant asserted that the heuristic data includes technical parameters such as the number of operations in each layer and execution time, these are simply data used in the prediction and scheduling process and do not recite any improvement to the operation of the process or memory. The amended claim do not modify how the training is performed, improve the AI model itself or provide a technological improvement to the functioning of a computer. Instead, the claims invoke generic computer components to collect data, analyze the data, and schedule execution based on the analysis. Any alleged benefits, such as reducing battery consumption or avoidance of thermal throttling, are merely the inherent result of choosing a more convenient time to execute a computation task. Accordingly, the claims are directed to abstract idea and the additional elements amount to no more than applying the abstract idea using generic computer components and therefore do not overcome the rejection under 35 USC § 101.
In Remarks/Arguments (pg. 8) Applicant contends: “Second, regarding Step 2B, Applicant asserts that the amended claims include an inventive concept. Even if the claims were viewed as directed to an abstract idea (which Applicant denies), the claims recite significantly more than the alleged abstract idea itself.…”
Regarding the above argument, the Examiner respectfully disagrees with Applicant’s assertion that the amended claims are significantly more than the abstract idea. The recited heuristic data, including the number of operations in each layer of the AI model and execution time, together with the user behavioral data, are merely inputs used by the claimed prediction and scheduling algorithm. The claims, as amended, lacks sufficient technical details to support a conclusion that they recite a technological improvement. In particular, the claims do not adequately describe how the heuristic data is processed with the user behavioral data to achieve the stated improvement. Accordingly, the additional elements amounts to no more than well-understood, routine and conventional computer functions of collecting, analyzing and using data, and therefore do not provide significantly more than the abstract idea under step 2B.
Applicant’s arguments (pgs. 9 – 13) with respect to amended claim(s) have been considered but are moot, because arguments/remarks are directed to amended claim limitations that were not previously examined by the examiner. The rejections are noted in the current office action to address amended claim limitations.
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 1-2, 4-5, 7-13, 15-16 and 18 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1: “Is the claim directed to a process, machine, manufacture or composition of matter?”
Yes. Claims 1-2, 4-5, 7-11 are directed to a process (method as claimed). Claims 12-13, 15-16 and 18 are directed to a machine (system as claimed).
Step 2:
Regarding claim 1, it recites:
An on-device training method to train an artificial intelligence (AI) model on a device, the on-device training method comprising:
receiving a training request for initiating on-device training from one or more applications based on dataset for the on-device training of the AI model being obtained through the one or more applications, wherein the training request includes heuristic data regarding the AI model;
determining whether at least one policy of a plurality of predefined policies regarding state of at least one component included in the device is satisfied, based on the training request; and
training, based on the at least one policy being satisfied, the AI model using data associated with the one or more applications.
in response to determining that the at least one policy is not satisfied, predicting a time period for training the AI model based on a user behavior and the heuristic data; and scheduling the training of the AI model in the predicted time period, wherein the heuristic data includes a number of operations in each layer of the AI model and time taken to execute the number of operations in each layer of the AI model.
Step 2A(I):
The abstract idea is recited in the following limitations:
determining whether at least one policy of a plurality of predefined policies regarding state of at least one component included in the device is satisfied, based on the training request.
This is interpreted as performance in the mind with the aid of pen and paper. A user can review data received on current device status and determine whether to begin a process.
in response to determining that the at least one policy is not satisfied, predicting a time period for training the AI model based on a user behavior and the heuristic data; and scheduling the training of the AI model in the predicted time period
This is interpreted as performance in the mind with the aid of pen and paper. A user can evaluate whether a condition is satisfied, analyzing information (i.e.: user behavior and heuristic data), predicting and appropriate time period based on the information and deciding when to perform an activity.
Step 2A(II):
The judicial exceptions as recited are not integrated into a practical application. In particular, claim 1 only recites “a device” that “trains” implying a computer to perform the steps listed above. The implied computer of the limitations are recited at a high level of granularity such that it amounts to no more than mere instructions to apply the exception using a generic computing component.
receiving a training request for initiating on-device training from one or more applications based on dataset for the on-device training of the AI model being obtained through the one or more applications; (Mere Data Gathering- receiving a request for data and corresponding data set)
wherein the training request includes heuristic data regarding the AI model; (simply links the judicial exception to a field of use and/or technology environment, MPEP 2106.05(h))
training, based on the at least one policy being satisfied, the AI model using data associated with the one or more applications. (Insignificant Application- Performing a task after determining to perform the task- In re Brown, 645 Fed. App'x 1014, 1016-1017 (Fed. Cir. 2016) (non-precedential)).
wherein the heuristic data includes a number of operations in each layer of the AI model and time taken to execute the number of operations in each layer of the AI model. (simply links the judicial exception to a field of use and/or technology environment, MPEP 2106.05(h))
Step 2B:
The claim limitations reciting the abstract idea do not include additional elements that are sufficient to amount to significantly more than the judicial exception. A discussed above with respect to integration of the abstract idea into a practical application, the additional element of implying a computer to perform the recited steps amount to no more than mere instructions to apply the exception using a generic computer component.
Regarding limitation (III), recites mere application of the abstract idea or mere instructions to implement an abstract idea on a computer are deemed insufficient to transform the judicial exception to a patentable invention because the limitations generally apply the use of a generic computer and/or process with the judicial exception, see MPEP 2106.05(f).
Regarding limitation (I), additional elements considered extra/post solution activity, as analyzed above, are activity that are well-understood routine and conventional, specifically: the courts have recognized the computer functions as well‐understood, routine, and conventional functions.
Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); TL| Communications LLC v. AV Auto. LLC, 823 F.3d 607, 610, 118 USPQ2d 1744, 1745 (Fed. Cir. 2016) (using a telephone for image transmission); OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network). See MPEP 2106.05(d)(II).
Regarding limitation (II and IV), additional elements are deemed insufficient to transform the judicial exception to a patentable invention to a patentable invention because they generally link the judicial exception to the technology environment, see MPEP 2106.05(h).
Regarding claim 2, fail to resolve the deficiencies identified above by integrating the judicial exception into a practical application, or introducing significantly more than the judicial exception. it recites the abstract ideas of:
triggering a wait period, if the at least one policy is not satisfied (Interpreted as performance in the mind with the aid of pen and paper. A user can begin a process when a policy is satisfied); and
scheduling, after expiry of the wait period, the training of the AI model based on redetermination process of determining each of whether the on-device training has been triggered and whether the at least one policy is satisfied (Interpreted as performance in the mind with the aid of pen and paper. A user can schedule an event).
Claim 13, recite similar subject matter as claim 2, so is rejected under the same rationale.
Regarding claim 4, fail to resolve the deficiencies identified above by integrating the judicial exception into a practical application, or introducing significantly more than the judicial exception. The claim recites:
wherein the plurality of predefined policies include temperature of the device, state of battery charge of the device, sleep mode status of the device, priority of current tasks running on the device, and availability of device resources.
The recitation in the additional limitation simply links the judicial exception to a field of use and/or technology environment, see MPEP 2106.05(h).
Limitations directed to field of use cannot integrate a judicial exception into a practical application at Step 2A or provide an inventive concept in Step 2B
Claim 15, recite similar subject matter as claim 4, so is rejected under the same rationale.
Regarding claim 5, fail to resolve the deficiencies identified above by integrating the judicial exception into a practical application, or introducing significantly more than the judicial exception. The claim recites:
calculating loss function and gradients of loss with respect to weights for each layer of the AI model, based on the data associated with the one or more applications; (Interpreted as abstract idea: mathematical concept, it involves performing a mathematical calculation on data to determine values (i.e.: loss function and gradient));
constructing at least one graph based on the loss function and gradients of loss; (Interpreted as abstract idea: mental process, it involves organizing and presenting the calculated information in a graphical representation.)
and training the AI model based on the at least one graph.
Deemed insufficient to transform the judicial exception to a patentable invention because the limitation is directed to mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea and are considered to adding the words “apply it” (or an equivalent) with the judicial exception, See MPEP 2106.05(f).
Limitations directed to using the computer as a tool for implementing an abstract idea cannot integrate a judicial exception into a practical application at Step 2A or provide an inventive concept in Step 2B.
Claim 16, recite similar subject matter as claim 5, so is rejected under the same rationale.
Regarding claim 7, fail to resolve the deficiencies identified above by integrating the judicial exception into a practical application, or introducing significantly more than the judicial exception. The claim recites:
wherein the AI model is a backpropagation model.
The recitation in the additional limitation simply links the judicial exception to a field of use and/or technology environment, see MPEP 2106.05(h).
Limitations directed to field of use cannot integrate a judicial exception into a practical application at Step 2A or provide an inventive concept in Step 2B.
Claim 18, recite similar subject matter as claim 7, so is rejected under the same rationale.
Regarding claim 8, fail to resolve the deficiencies identified above by integrating the judicial exception into a practical application, or introducing significantly more than the judicial exception. The claim recites:
wherein the one or more applications are AI based applications.
The recitation in the additional limitation simply links the judicial exception to a field of use and/or technology environment, see MPEP 2106.05(h).
Limitations directed to field of use cannot integrate a judicial exception into a practical application at Step 2A or provide an inventive concept in Step 2B.
Regarding claim 9, fail to resolve the deficiencies identified above by integrating the judicial exception into a practical application, or introducing significantly more than the judicial exception. The claim recites:
wherein the at least one policy includes a dataset being sufficient to train the AI model associated with the one or more applications.
The recitation in the additional limitation simply links the judicial exception to a field of use and/or technology environment, see MPEP 2106.05(h).
Limitations directed to field of use cannot integrate a judicial exception into a practical application at Step 2A or provide an inventive concept in Step 2B.
Regarding claim 10, fail to resolve the deficiencies identified above by integrating the judicial exception into a practical application, or introducing significantly more than the judicial exception. The claim recites:
wherein determination of the sufficient dataset is configurable by the one or more applications.
The recitation in the additional limitation simply links the judicial exception to a field of use and/or technology environment, see MPEP 2106.05(h).
Limitations directed to field of use cannot integrate a judicial exception into a practical application at Step 2A or provide an inventive concept in Step 2B.
Regarding claim 11, fail to resolve the deficiencies identified above by integrating the judicial exception into a practical application, or introducing significantly more than the judicial exception. The claim recites:
wherein the sufficient dataset differs according to different applications.
The recitation in the additional limitation simply links the judicial exception to a field of use and/or technology environment, see MPEP 2106.05(h).
Limitations directed to field of use cannot integrate a judicial exception into a practical application at Step 2A or provide an inventive concept in Step 2B.
Regarding claim 12, it recites:
The rest of the limitations are analogous to claim 1, so are rejected under similar rationale.
An on-device training system to train an artificial intelligence (AI) model on a device, the system comprising: at least one processor including processing circuitry, at least one memory storing instructions that, when executed by the at least one processor individually or collectively, cause the electronic device to
Deemed insufficient to transform the judicial exception to a patentable invention because the limitation is directed to mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea and are considered to adding the words “apply it” (or an equivalent) with the judicial exception, See MPEP 2106.05(f).
Limitations directed to using the computer as a tool for implementing an abstract idea cannot integrate a judicial exception into a practical application at Step 2A or provide an inventive concept in Step 2B.
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.
Claim(s) 1-2, 4, 8, 12-13, and 15 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Sanketi United States Patent Application Publication US 2019/0050746 in view of Cho, United States Patent Application Publication: US12443876B2 and Daniels, United States Patent Application Publication: US20210397486A1.
Regarding claim 1, Sanketi discloses an on-device training method to train an artificial intelligence (AI) model on a device (Sanketi, fig 1, on-device structure), the on-device training method comprising:
receiving a training request for initiating on-device training from one or more applications based on dataset for the on-device training of the AI model being obtained through the one or more applications (Sanketi, para [0043, 100], communication to initiate training via training API represents a training request; Sanketi, fig 1, on device machine learning platform 122 accessed by applications 120a-120n);
determining whether at least one policy of a plurality of predefined policies regarding state of at least one component included in the device is satisfied, based on the training request (Sanketi, para [0076], training plan for on-device training via training API); and
training, based on the at least one policy being satisfied, the AI model using data associated with the one or more applications (Sanketi, para [0043, 76], training can be performed based on training plan).
in response to determining that the at least one policy is not satisfied, predicting a time period for training the AI model (Sanketi, para [0165] In some implementations, as the training phase may be long (e.g., minutes), the training process can suspend and resume based on changing device conditions.)
scheduling the training of the AI model in the predicted time period, (Sanketi, para [0043] … For example, the training can be performed in the background at scheduled times and/or when the device is idle.)
Sanketi does not teach:
wherein the training request includes heuristic data regarding the AI model;
wherein the heuristic data includes a number of operations in each layer of the AI model and time taken to execute the number of operations in each layer of the AI model.
… based on a user behavior and the heuristic data;
Cho discloses:
wherein the training request includes heuristic data regarding the AI model (Cho, col. 2 line [55 – 57], In autotuning frameworks, these processes find an optimal low-level implementation using sequential operations consisting of four stages before finding the optimal one.)
wherein the heuristic data includes a number of operations in each layer of the AI model (Cho, col. 3 line [58 – 62], For example, the scheduler may achieve this in three ways: 1) leveraging fine-grained domain knowledge (e.g., the cost of each stage in terms of execution time, update model interval, and queue status) with a Shortest Job First (SJF) scheduling policy to mitigate long queuing delay,).
and time taken to execute the number of operations in each layer of the AI model. (Cho, col. 2 line [36 – 39] For example, several ML and AI computations are matrix operations (e.g., adding and multiplication) and the speed of these operations, in the aggregate, effect the overall computation time of the system. )
Before the time of the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art to have modified the on-device learning to implement the stateless task scheduling and remote key value store techniques. The motivation for doing so would have been to reduce queuing delay and improve resource utilization (Cho, Abstract).
Sanketi in view of Cho do not teach:
… based on a user behavior and the heuristic data;
Daniels discloses:
… based on a user behavior and the heuristic data; (Daniels, para [0162] At block 340, process flow 300 determines to monitor a second entity's user activity for an indication to suspend the event. The indication of suspending the event may be detected based on determining that the second entity has approved the suspension of the event…)
Before the time of the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art to have modified the on-device learning to incorporate the event suspension and resource synchronization techniques. The motivation for doing so would have been to coordinate updates across multiple computing devices. (Daniels, Abstract).
Regarding claim 2, Sanketi in view of Cho and Daniels discloses the method of claim 1. Sanketi additionally discloses further comprising: triggering a wait period, if the at least one policy is not satisfied; and scheduling, after expiry of the wait period, the training of the AI model based on redetermination process of determining each of whether the on-device training has been triggered and whether the at least one policy is satisfied (Sanketi, para [0116], Training will only be scheduled if device conditions allow, for instance, waits for device to be in both idle and charging).
Regarding claim 4, Sanketi in view of Cho and Daniels discloses the method of claim 1. Sanketi additionally discloses wherein the plurality of predefined policies include temperature of the device, state of battery charge of the device, sleep mode status of the device, priority of current tasks running on the device, and availability of device resources (Sanketi, para [0116], determines if charging).
Regarding claim 8, Sanketi in view of Cho and Daniels discloses the method of claim 1. Sanketi additionally discloses wherein the one or more applications are AI based applications (Sanketi, para [0027], perform machine learning management operations which enable performance of on-device machine learning functions on behalf of one or more locally-stored applications).
Claims 12-13 and 15 recite substantially similar limitations to claims 1-2 and 4, respectively, and are thus similarly rejected.
Claim(s) 5 and 16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sanketi in view of Cho, Daniels and in further view of Xie United States Patent Application Publication US 20180302498.
Regarding claim 5, Sanketi in view of Cho and Daniels discloses the method of claim 1. Sanketi in view of Cho and Daniels do not disclose the additional steps of the present claim.
Xie discloses wherein training the AI model comprises: calculating loss function and gradients of loss with respect to weights for each layer of the AI model, based on the data associated with the one or more applications; constructing at least one graph based on the loss function and gradients of loss; and training the AI model based on the at least one graph (Xie, para [0023], expression graph includes loss function and gradient; Xie, para [0025], updates model based on expression graph).
Before the time of the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art to have modified the type of learning based on the teaching of Xie. The motivation for doing so would have been increased flexibility of resource adjustment (Xie, para [0003]).
Claims 16 recites substantially similar limitations to claims 5, and is thus similarly rejected.
Claim(s) 7 and 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sanketi in view of Cho, Daniels and in further view of Ravi United States Patent Application Publication US20200125956.
Regarding claim 7, Sanketi in view of Cho and Daniels discloses the method of claim 1. Sanketi in view of Cho and Daniels do not disclose the additional steps of the present claim.
Ravi discloses wherein the AI model is a backpropagation model (Ravi, para [0110], the model (e.g., network) learns to optimize the weights and activations in the quantized space using gradients computed via backpropagation.).
Before the time of the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art to have modified the machine learning to include computation via backpropagation. The motivation for doing so would have been to be more effective that just applying post training (Ravi, para [0110]).
Claims 18 recites substantially similar limitations to claims 7, and is thus similarly rejected.
Claim(s) 9-11 are rejected under 35 U.S.C. 103 as being unpatentable over Sanketi in view of Cho, Daniels and in further view of Das United States Patent US 10,700,992.
Regarding claim 9, Sanketi in view of Cho and Daniels discloses the method of claim 1. Sanketi in view of Cho and Daniels do not disclose the additional limitations of the present claim.
Das discloses wherein at least one policy includes a dataset being sufficient to train the AI model associated with the one or more applications (Das, col 7, rows 36-54, with regards to fig 5, element 502, plurality of parameters represents data set).
Before the time of the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art to have modified the policies according to Das. The motivation for doing so would have been to less human intervention with additional inputs (Das, col 1, rows 36-42).
Regarding claim 10, Sanketi in view of Cho, Daniels and Das discloses the method of claim 9. Das additionally discloses wherein determination of the sufficient dataset is configurable by the one or more applications (Das, col 8, rows 26-33, “predefined threshold” represents “sufficient”; Das, col 10, rows 6-17, configurable range).
Before the time of the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art to have modified the policies according to Das. The motivation for doing so would have been to less human intervention with additional inputs (Das, col 1, rows 36-42).
Regarding claim 11, Sanketi in view of Cho, Daniels and Das discloses the method of claim 9. Das additionally discloses wherein the sufficient dataset differs according to different applications (Das, col 10, rows 6-17, configurable range for circumstance).
Before the time of the effective filing date of the claimed invention, it would have been obvious to one of ordinary skill in the art to have modified the policies according to Das. The motivation for doing so would have been to less human intervention with additional inputs (Das, col 1, rows 36-42).
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any extension fee pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner
should be directed to MATIYAS T MARU whose telephone number is (571)270-0902 or via email: matiyas.maru@uspto.gov. The examiner can normally be reached Monday - Friday (8:00am - 4:00pm) EST.
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, Michelle Bechtold can be reached on (571)431-0762. The fax phone number for the organization were 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.
/M.T.M./Examiner, Art Unit 2148
/MICHELLE T BECHTOLD/Supervisory Patent Examiner, Art Unit 2148
Prosecution Insights