Prosecution Insights
Last updated: November 20, 2026
Application No. 18/993,755

OBFUSCATING INFERENCE OPERATIONS OF A MACHINE LEARNING MODEL

Final Rejection §103
Filed
Jan 13, 2025
Priority
Jul 14, 2022 — nonprovisional of PCTUS2022037149
Examiner
LESNIEWSKI, VICTOR D
Art Unit
2493
Tech Center
2400 — Computer Networks
Assignee
Google LLC
OA Round
2 (Final)
59%
Grant Probability
Moderate
3-4
OA Rounds
1y 5m
Est. Remaining
99%
With Interview

Examiner Intelligence

Grants 59% of resolved cases
59%
Career Allowance Rate
291 granted / 496 resolved
+0.7% vs TC avg
Strong +55% interview lift
Without
With
+54.8%
Interview Lift
resolved cases with interview
Typical timeline
3y 3m
Avg Prosecution
20 currently pending
Career history
527
Total Applications
across all art units

Statute-Specific Performance

§101
8.9%
-31.1% vs TC avg
§103
57.9%
+17.9% vs TC avg
§102
17.0%
-23.0% vs TC avg
§112
13.0%
-27.0% vs TC avg
Black line = Tech Center average estimate • Based on career data from 496 resolved cases

Office Action

§103
DETAILED ACTION The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . The amendment filed 8/25/2026 has been placed of record in the file. Claims 1, 4, 12, 30, and 35 have been amended. The rejection of claim 35 under 35 U.S.C. 101 is withdrawn in view of the amendment. Claims 5 and 31 have been canceled. Claims 36 and 37 have been added. Claims 1-4, 6-12, 28-30, and 32-37 are now pending. The applicant’s arguments with respect to claims 1-4, 6-12, 28-30, and 32-37 have been considered but are moot in view of the following new grounds of rejection. The IDS filed 5/26/2026 and the IDS filed 8/25/2026 have been considered. Response to Amendment Claims have been amended to further define the structure of the hardware device. The amendment proves a change in scope to the independent claims as the independent claims now explicitly state a first processing element located apart from a second processing element, etc. However, none of the amended claims show a patentable distinction over the prior art as evidenced by the following new grounds of rejection. 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. Claims 1-4, 6-12, 28-30, and 32-37 are rejected under 35 U.S.C. 103 as being unpatentable over Yang et al. (U.S. Patent Application Publication Number 2023/0259638), hereinafter referred to as Yang, in view of Chen et al. (U.S. Patent Application Publication Number 2022/0114014) as listed on the Notice of References Cited dated 5/28/2026, hereinafter referred to as Chen. Yang disclosed techniques for obfuscated handling of confidential information. In an analogous art, Chen disclosed techniques for protecting AI model parameters. Both systems are directed toward the obfuscation of AI model parameters. Regarding claim 1, Yang discloses a method, comprising: receiving, by a hardware device, data representing a machine learning model comprising a plurality of model parameters for inference operations, wherein the hardware device comprises a set of computation units arranged in one or more processing elements (paragraph 67, neural network parameters loaded into ALUs and parameters used for inferencing); obtaining instructions for performing obfuscating operations configured to obfuscate one or more measurable characteristics of the machine learning model when the machine learning model is executed by the one or more processing elements (paragraph 41, obfuscating operations); causing a first portion of the set of computation units located in a first processing element of the one or more processing elements to perform the inference operations of the machine learning model (paragraph 47, real computations); and causing a second portion of the set of computation units located in a processing element of the one or more processing elements to perform the obfuscating operations concurrently with the first portion of the set of computation units performing the inference operations (paragraph 47, producing noise to obfuscate). Yang does not explicitly state that the one or more processing elements are a plurality of processing elements, the second portion of computation units located in a second processing element, wherein the first processing element is located apart from the second processing element. However, implementing a hardware device in such a fashion was well known in the art as evidenced by Chen. Since the inventions encompass the same field of endeavor, 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 system of Yang by adding the ability that the one or more processing elements are a plurality of processing elements, the second portion of computation units located in a second processing element, wherein the first processing element is located apart from the second processing element as provided by Chen (see paragraph 98, TEE provides obfuscation while non-TEE handles inference requests). One of ordinary skill in the art would have recognized the benefit that providing obfuscation in this way would assist in ensuring the privacy of AI model parameters (see Chen, paragraph 2). Regarding claim 2, the combination of Yang and Chen discloses wherein the machine learning model is a neural network, wherein the obfuscating operations are configured to obscure at least one of a number of network layers of the neural network, a number of nodes in a network layer of the neural network, a nodal operation for a node in a network layer of the neural network, or a weight value associated with a node in a network layer of the neural network (Yang, paragraph 67, parameters configure neurons or layers of neural network). Regarding claim 3, the combination of Yang and Chen discloses wherein the one or more measurable characteristics of the machine learning model comprises at least one of a power profile, an electromagnetic profile, or a time profile (Yang, paragraph 51, obfuscation changes SPA, DPA, etc.). Regarding claim 4, the combination of Yang and Chen discloses the second processing element is an additional processing element integrated within the hardware device to execute the obfuscating operations (Chen, paragraph 98, TEE). Regarding claim 6, the combination of Yang and Chen discloses wherein the obfuscating operations include an obfuscating nodal operation for a particular node in a network layer to be performed concurrently with a corresponding nodal operation for the particular node (Yang, paragraph 67, neurons of neural network, and paragraph 47, real computations and producing noise to obfuscate). Regarding claim 7, the combination of Yang and Chen discloses wherein obfuscating nodal operation specifies an activation function for the particular node that is different from an actual activation function of the particular node (Yang, paragraph 72, activations from neurons, and paragraph 47, real computations and producing noise to obfuscate). Regarding claim 8, the combination of Yang and Chen discloses wherein causing the second portion of the set of computation units to perform the obfuscating operations concurrently with the first portion of the set of computation units performing the inference operations comprises assigning the obfuscating operations to a dedicated processing element that performs the obfuscating operations (Yang, paragraph 76, dedicated computational resource). Regarding claim 9, the combination of Yang and Chen discloses wherein the dedicated processing element includes one or more processing elements or computation units that are additionally incorporated into a hardware device and are configured to perform substantially only corresponding obfuscating operations (Yang, paragraph 76, separate computational hardware that includes ALUs). Regarding claim 10, the combination of Yang and Chen discloses wherein causing the second portion of the set of computation units to perform the obfuscating operations concurrently with the first portion of the set of computation units performing the inference operations comprises: assigning the obfuscating operations to one or more processing elements that also perform inference operations for one or more machine learning models; and reassigning a subset of the inference operations from the one or more processing elements to other processing elements of the hardware device (Yang, paragraph 47, multiple stages perform computational operations but only one is real). Regarding claim 11, the combination of Yang and Chen discloses wherein the neural network is configured to perform human face recognition tasks for unlocking devices (Yang, paragraph 131, image recognition, where facial recognition for unlocking devices was common in the art). Regarding claim 12, Yang discloses a system comprising a hardware device and one or more storage devices storing instructions that when executed by the hardware device cause the hardware device to perform operations, the operations comprising: receiving, by the hardware device, data representing a machine learning model comprising a plurality of model parameters for inference operations, wherein the hardware device comprises a set of computation units arranged in one or more processing elements (paragraph 67, neural network parameters loaded into ALUs and parameters used for inferencing); obtaining instructions for performing obfuscating operations configured to obfuscate one or more measurable characteristics of the machine learning model when the machine learning model is executed by the one or more processing elements (paragraph 41, obfuscating operations); causing a first portion of the set of computation units located in a first processing element of the one or more processing elements to perform the inference operations of the machine learning model (paragraph 47, real computations); and causing a second portion of the set of computation units located in a processing element of the one or more processing elements to perform the obfuscating operations concurrently with the first portion of the set of computation units performing the inference operations (paragraph 47, producing noise to obfuscate). Yang does not explicitly state that the one or more processing elements are a plurality of processing elements, the second portion of computation units located in a second processing element, wherein the first processing element is located apart from the second processing element. However, implementing a hardware device in such a fashion was well known in the art as evidenced by Chen. Since the inventions encompass the same field of endeavor, 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 system of Yang by adding the ability that the one or more processing elements are a plurality of processing elements, the second portion of computation units located in a second processing element, wherein the first processing element is located apart from the second processing element as provided by Chen (see paragraph 98, TEE provides obfuscation while non-TEE handles inference requests). One of ordinary skill in the art would have recognized the benefit that providing obfuscation in this way would assist in ensuring the privacy of AI model parameters (see Chen, paragraph 2). Regarding claim 28, the combination of Yang and Chen discloses wherein the machine learning model is a neural network, wherein the obfuscating operations are configured to obscure at least one of a number of network layers of the neural network, a number of nodes in a network layer of the neural network, a nodal operation for a node in a network layer of the neural network, or a weight value associated with a node in a network layer of the neural network (Yang, paragraph 67, parameters configure neurons or layers of neural network). Regarding claim 29, the combination of Yang and Chen discloses wherein the one or more measurable characteristics of the machine learning model comprises at least one of a power profile, an electromagnetic profile, or a time profile (Yang, paragraph 51, obfuscation changes SPA, DPA, etc.). Regarding claim 30, the combination of Yang and Chen discloses wherein the second processing element is an additional processing element integrated within the hardware device to execute the obfuscating operations (Chen, paragraph 98, TEE). Regarding claim 32, the combination of Yang and Chen discloses wherein the obfuscating operations include an obfuscating nodal operation for a particular node in a network layer to be performed concurrently with a corresponding nodal operation for the particular node (Yang, paragraph 67, neurons of neural network, and paragraph 47, real computations and producing noise to obfuscate). Regarding claim 33, the combination of Yang and Chen discloses wherein obfuscating nodal operation specifies an activation function for the particular node that is different from an actual activation function of the particular node (Yang, paragraph 72, activations from neurons, and paragraph 47, real computations and producing noise to obfuscate). Regarding claim 34, the combination of Yang and Chen discloses wherein causing the second portion of the set of computation units to perform the obfuscating operations concurrently with the first portion of the set of computation units performing the inference operations comprises assigning the obfuscating operations to a dedicated processing element that performs the obfuscating operations (Yang, paragraph 76, dedicated computational resource). Regarding claim 35, Yang discloses one or more non-transitory computer-readable storage media storing instructions that when executed by one or more computers cause the one or more computers to perform operations comprising: receiving, by a hardware device, data representing a machine learning model comprising a plurality of model parameters for inference operations, wherein the hardware device comprises a set of computation units arranged in one or more processing elements (paragraph 67, neural network parameters loaded into ALUs and parameters used for inferencing); obtaining instructions for performing obfuscating operations configured to obfuscate one or more measurable characteristics of the machine learning model when the machine learning model is executed by the one or more processing elements (paragraph 41, obfuscating operations); causing a first portion of the set of computation units located in a first processing element of the one or more processing elements to perform the inference operations of the machine learning model (paragraph 47, real computations); and causing a second portion of the set of computation units located in a processing element of the one or more processing elements to perform the obfuscating operations concurrently with the first portion of the set of computation units performing the inference operations (paragraph 47, producing noise to obfuscate). Yang does not explicitly state that the one or more processing elements are a plurality of processing elements, the second portion of computation units located in a second processing element, wherein the first processing element is located apart from the second processing element. However, implementing a hardware device in such a fashion was well known in the art as evidenced by Chen. Since the inventions encompass the same field of endeavor, 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 system of Yang by adding the ability that the one or more processing elements are a plurality of processing elements, the second portion of computation units located in a second processing element, wherein the first processing element is located apart from the second processing element as provided by Chen (see paragraph 98, TEE provides obfuscation while non-TEE handles inference requests). One of ordinary skill in the art would have recognized the benefit that providing obfuscation in this way would assist in ensuring the privacy of AI model parameters (see Chen, paragraph 2). Regarding claim 36, the combination of Yang and Chen discloses wherein performing the obfuscating operations concurrently comprises executing the obfuscating operations during a first time period that at least partially overlaps with a second time period in which the inference operations are executed (Chen, paragraph 98, TEE and non-TEE operations overlap). Regarding claim 37, the combination of Yang and Chen discloses wherein performing the obfuscating operations comprises: executing an obfuscating activation function on the second portion of the set of computation units that is different from an actual activation function being executed by the first portion of the set of computation units (Yang, paragraph 47, producing noise to obfuscate). 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 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 Victor Lesniewski whose telephone number is (571)272-2812. The examiner can normally be reached Monday thru Friday, 9am to 5pm. 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, Carl Colin can be reached at 571-272-3862. 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. /Victor Lesniewski/Primary Examiner, Art Unit 2493
Read full office action

Prosecution Timeline

Jan 13, 2025
Application Filed
May 28, 2026
Non-Final Rejection mailed — §103
Aug 09, 2026
Interview Requested
Aug 20, 2026
Examiner Interview Summary
Aug 20, 2026
Applicant Interview (Telephonic)
Aug 25, 2026
Response Filed
Nov 09, 2026
Final Rejection mailed — §103 (current)

Precedent Cases

Applications granted by this same examiner with similar technology

Patent 12801739
PROVER DEVICE, VERIFIER DEVICE, AND REMOTE ATTESTATION SYSTEM
2y 6m to grant Granted Nov 17, 2026
Patent 12794925
APPARATUS AND METHOD FOR UPDATING RISK DETERMINATION BASED ON REAL-TIME CHANGES
4y 1m to grant Granted Nov 10, 2026
Patent 12794912
INTRUSION MONITORING SYSTEM, METHOD AND RELATED PRODUCTS
3y 7m to grant Granted Nov 10, 2026
Patent 12795367
USER EQUIPMENT PARAMETER UPDATE HEADER PROTECTION
2y 1m to grant Granted Nov 10, 2026
Patent 12750362
SECURE ELEMENT ARRAYS IN INTERNET-OF-THINGS SYSTEMS
3y 5m to grant Granted Sep 29, 2026
Study what changed to get past this examiner. Based on 5 most recent grants.

Strategy Recommendation AI-generated — please review before filing

Get a prosecution strategy drawn from examiner precedents, rejection analysis, and claim mapping.
Typically takes 5-10 seconds — AI-generated, attorney review required before filing

Prosecution Projections

3-4
Expected OA Rounds
59%
Grant Probability
99%
With Interview (+54.8%)
3y 3m (~1y 5m remaining)
Median Time to Grant
Moderate
PTA Risk
Based on 496 resolved cases by this examiner. Grant probability derived from career allowance rate.

Sign in with your work email

Enter your email to receive a magic link. No password needed.

Personal email addresses (Gmail, Yahoo, etc.) are not accepted.

Free tier: 3 strategy analyses per month