DETAILED ACTION
This is in response to the amendment filed on October 29th 2025.
Response to Arguments
Applicant’s arguments, see pg. 9, filed 10/29/25, with respect to the claim objections and 112 rejections have been fully considered and are persuasive. The claim objections and 112 rejections have been withdrawn.
Applicant’s arguments, see pg. 10, with respect to the 103 rejections have been fully considered and are persuasive. The 103 rejection of claims 3, 17 and their dependents has been withdrawn.
Applicant states new claims 26-27 are patentable because they depend from claim 12. Examiner agrees in part, see the detailed rejection below. However, based on the amendment to claim 12, a new ground of rejection is made in view of Karianakis et al. US 2021/0073563 A1.
Claim Objections
Claim 12 is objected to because of the following informalities: it recites “partitioning the data analytic model one or more first …”. It seems a word such as “into” is missing as this term was deleted in the prior amendment. Appropriate correction is required.
Claim Rejections - 35 USC § 103
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
Claim(s) 12 and 27 are rejected under 35 U.S.C. 103 as being unpatentable over Cheng et al. US 11,038,910 B1 in view of Dey et al. US 2020/0143254 A1 and Karianakis et al. US 2021/0073563 A1.
Regarding claim 12, Cheng discloses a method for supporting a home computing system (abstract) comprising:
identifying a data analytic model for the home computing system, wherein the home computing system supports at least one Internet of Things device (abstract, Fig. 2, col. 3 ln. 24-67);
partitioning the data analytic model (model is composed of sub-models – see Fig. 13 col. 13 ln. 26-43);
executing, by the home computing system and using input based on source data of the at least one IoT device, a first sub-model of the one or more first sub-models (Fig. 4 shows an embodiment where the models are executed locally at the home computing system – also see col. 4 ln. 48-66; input data comes from IoT devices – Figs. 2-4);
executing, by a public computing cloud … a second sub-model of the one or more second sub-models (backend “cloud” or IoT server executes the models – see Figs. 2-3 and col. 4 ln. 4-47).
Cheng does not explicitly disclose the data analytic model comprises a set of layers or sending, from the home computing system to a public computing cloud and based on the executing the first sub-model, output data from the first sub-model; and [executing the second model] using input based on the output data from the home computing system. But this is taught by Dey as a system that partitions layers of a convolution network between devices and uses the outputs from a first layer as inputs to a second layer (see paragraphs 6, 8 and Fig. 2B).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Cheng to partition the execution of layers as taught by Dey. Dey teaches that sending full input data over a network is costly (paragraph 5) and that by executing model layers on fog devices, network traffic is reduced (abstract).
The combination of Cheng and Dey does not explicitly disclose the set of layers of the model, consisting of layers of the data analytic model that have only fixed parameters, and one or more second subsets of the set of layers of the data analytic model, consisting of a second subset of layers, the set of layers of the data analytic model that have parameters subject to reinforcement learning. However, this is known in the art as transfer learning as taught by Karianakis as a model with frozen bottom layers and top layers that are re-trained or transferred (reinforcement learning – Figs. 6, 7, paragraphs 18 and 51).
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 combination of Cheng and Dey with the transfer learning technique taught by Karianakis wherein some layers are fixed/frozen and other layers use reinforcement learning. One of ordinary skill would understand the general benefits of transfer learning include reduced time and cost by using pretrained fixed layers and then updating a portion of layers to improve results.
Regarding claim 27, Cheng discloses performing, by the public computing cloud, reinforcement training of one or more of the parameters subject to reinforcement training (data is sent to backend system to update/retrain models – abstract; also see Figs. 6, 8-9, col. 5 ln. 54-58); and
sending, by the public computing cloud and to the home computing system, a reinforcement trained parameter of the one or more reinforcement trained parameters (home IoT devices execute updated models – Figs. 3-4, 8 and col. 8 ln. 4-16).
Allowable Subject Matter
Claims 3-9, 11, 13, 17-22 and 24-25 are allowed.
The following is an examiner’s statement of reasons for allowance: claims 3 and 17 incorporate subject matter previously indicated allowable. While many elements of the claim are taught by the art as explained in the previous Office Action, the amended claims when considered as a whole, now distinguish over the art. The prosecution history is clear so no additional reasons for allowance are necessary.
Any comments considered necessary by applicant must be submitted no later than the payment of the issue fee and, to avoid processing delays, should preferably accompany the issue fee. Such submissions should be clearly labeled “Comments on Statement of Reasons for Allowance.”
Claim 26 is objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims. Claim 26 contains subject matter similar to claims 3 and 17 so it is allowable for similar reasons.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
F. M. Campos de Oliveira and E. Borin, "Partitioning Convolutional Neural Networks for Inference on Constrained Internet-of-Things Devices," 2018 30th International Symposium on Computer Architecture and High Performance Computing (SBAC-PAD), Lyon, France, 2018, pp. 266-273; discloses partitioning a model into layers for execution on IoT devices (Section I) as well as performing a dynamic scheduling of devices (Section II).
Hyuk-Jin Jeong et al. "PerDNN: Offloading Deep Neural Network Computations to Pervasive Edge Servers," 2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS), Singapore, Singapore, 2020, pp. 1055-1066 discloses partitioning a model into sub-models/layers based on network traffic as explained in the previous Office Action.
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 nonprovisional extension fee (37 CFR 1.17(a)) 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 mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JASON D RECEK whose telephone number is (571)270-1975. The examiner can normally be reached Flex M-F 9-5.
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, Umar Cheema can be reached at 571-270-3037. 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.
/JASON D RECEK/Primary Examiner, Art Unit 2458