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 .
Response to Arguments
According to applicants’ arguments filed on 08/20/2026, regarding the rejection under 112(b) have been fully considered and are persuasive. Accordingly, the previously stated rejection of claims 1-19 and 21 under 112(b) is withdrawn.
Applicants’ arguments with respect to 103 rejection for claims 1,16 and 21 have been fully considered but they are not persuasive.
Regarding the asserted absence of separate owner-specific attestation procedures, the primary reference Grancharov in para:0058-0059 discloses model owner authorization using remote attestation. Fig.7 and Para:0014-0015 teaches attestation between the model release tool and model provisioning server, with storage of the model following successful attestation.Para:0123-0124 teaches verification of the receiving platforms integrity before sensitive user data is provisioned. Thus, Grancharov provides teachings relevant to both owner’s respective attestation conditions.
Regarding erasing, the rejection relies on the combined teachings of Grancharov and Zhai. Grancharov in Para0107 teaches an orchestrator that schedules the creation and destruction of ML-TEES and manages their lifecycle. Zhai teaches a delete/erase operation that clears all the secret data deleting a predetermined portion or all secret data within a secure execution space in response to user instruction (see, Page.5, para.12). As such it would have been obvious to apply this operation to the model, raw data, and results remaining in Grancharov’ s TEE after output, then destroy the TEE, to reduce the risk of later disclosure.
Although Zhai’s example responds to a security risk, the proposed combination uses its erasure mechanism when processing is complete. This is the proposed modification; neither reference is relied upon individually to teach the entire sequence.
Claim Rejections - 35 USC § 103
7. 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.
8. Claim(s) 1-11,16-19 and 21 are rejected under 35 U.S.C. 103 as being unpatentable over Grancharov (US Pub.No.2022/0114249) in view of Zhai (WO 2018/0149110).
9. Regarding claims 1,16 and 21 Grancharov teaches a computer-implemented method, a computer system and a computer-readable storage medium the method comprising: receiving, at a data processing pipeline comprising a trusted execution environment (TEE): a data-processing function from a data-processing function owner, a raw-data set from a data owner (Para:0035-0036, Para:0047-0048 and Para:0050 teaches an ML model supplied by a model owner and proprietary data supplied by a data owner to cloud computing infrastructure for processing with an ML-TEE.
Figs.1, 5, element.130 and Para:0047-0048 teaches data is protected while being provisioned to the cloud provider and while being used by an ML model in a TEE provided by the cloud provider for the secure execution of the ML model, hereafter referred to as an ML-TEE. The ML-TEE provides security guarantees for protecting data that is securely provisioned into it or computed inside of it. Para:0050 teaches the ML model owner is able to securely transmit the ML model 102 over network link 118 to cloud computing provider 130. Data 110 is proprietary to an associated data owner and may be stored in computing resources 112 that are local to, owned by, leased by, or controlled by the data owner. Similarly to the ML model owner, the data owner is able to transmit data 110 securely over network link 122 to the cloud computing provider 130 as well as receive results produced by the ML application after it is executed);
generating, in the data processing pipeline, analysis results, based on the raw-data set, by using the data-processing function; providing the analysis results to an output (Para:0048 Para:0050 and Para:0065 teaches executing the ML model within the ML-TEE and returning the results produced by the ML application to the data owner);
wherein: the raw data set is provided by the data owner in response to satisfaction of first-user trustworthiness-criteria verified by the data owner using a first-user remote-attestation-procedure (Para:0123-0124 teaches validating the security state of a hardware backed TEE before transmitting sensitive user data. A client or user verifies the integrity the integrity of a remote computing platform before provisioning secrets.Para:0127 teaches remote attestation-based provisioning of the data providers encryption key, followed by transmission of encrypted user data to an authorized ML-TEE container for processing);
and the data-processing function is provided by the data-processing function owner in response to satisfaction of second-user trustworthiness- criteria verified by the data-processing function owner using a second-user remote attestation procedure (Fig.7, Para: 0058-0059 and Para: 0114-0115 teaches remote attestation between the model release tool in the model owner’s network and the ML volume provisioning server. And once the remote attestation 704 is successfully completed the ML volume provisioning server 232 may store 706 the ML model 204 in its TEE secure enclave).
Grancharov in para0107 already disclosed an orchestrator that schedules the creation and destruction of ML-TEES and manages their lifecycle. However, Grancharov does not itself specify this complete post output cleanup sequence.
Zhai teaches deleting a predetermined portion or all secret data within a secure execution space in response to a third operation instruction associated with an identified security risk (see, Page.5, para.12)
Therefore, it would have been obvious to incorporate Zhai’s secret data erasure operation into Grancharov’ s TEE teardown process after the results are output. The modified process would clear the confidential model, raw data and retained analysis results during teardown to reduce the risk of this later disclosure.
10. Regarding claims 2, 17 Grancharov in view of Zhai teaches the method, the computer system, further comprising establishing the trusted execution environment in a volatile memory device (Grancharov: Fig.2, Para:0051-0052 teaches DRAM for program and data storage during execution. Para:0048 teaches execution of the model using data within an ML-TEE.
Zhai: Fig. 10, Pages.30-31 teaches computing device 700 usable as a server supporting a user application and a secure execution space, the computer device including processor 701 and memory 703. The memory 703 may include RAM or another dynamic storage device, and the processor 701 executes executable codes stored in memory).
11. Regarding claim 3 Grancharov teaches the method, further comprising establishing the trusted execution environment according to criteria defined by at least one of the data owner and the data-processing function owner (Para:0061, Para:0064 teaches model provisioning initiated by the ML model owner by passing a deployment configuration to an orchestrator. The orchestrator defines the attributes of the ML-TEEs in which the model will execute. Para:0107 teaches an orchestrator schedules creation of ML-TEEs).
12. Regarding claim 4 Zhai teaches the method, wherein erasure of the trusted execution environment results in permanent erasure of the data-processing function, the raw-data set and the analysis results from the data- processing pipeline (Page.5, para.12 teaches deleting a predetermined portion or all secret data within a secure execution space in response to a third operation instruction associated with an identified security risk) .
13. Regarding claim 5 Grancharov teaches the method, wherein the data processing pipeline comprises a number of trusted execution environments, each trusted execution environment including a unique trusted execution environment identity (Fig.5 and Para:0109 teaches ML-TEE hosts 546 containing multiple ML-TEE containers 540, each providing an isolated execution environment for ML images and volume, as such each TEE has unique TEE identity).
14. Regarding claim 6 Grancharov, Zhai teaches the method, wherein each trusted execution environment is erased after the analysis results associated with that trusted execution environment are provided to the output (Grancharov: Para:0050 teaches providing results produced by the ML application to the data owner. Para:00107 teaches an orchestrator schedules creation and destruction of ML-TEE and manages their lifecycle.
Zhai: Page.5, para.12 teaches deleting a predetermined portion or all secret data within a secure execution space in response to a third operation instruction associated with an identified security risk) .
15. Regarding claim 7 Grancharov in view of Zhai teaches the method, further comprising: receiving, at the data processing pipeline a training-data set from the data owner, the training-data set suitable for training the data-processing function (Grancharov : Para:0040, teaches training an ML model using training data, and the model is owned by one party and data is owned by another party. Para:0050 teaches receiving data supplied by a data owner. and Para:0057-0058 teaches a dynamic ML model being trained during a model execution stage) and erasing the training data set from the data processing pipeline (Zhai: Page.5, para:12 teaches deleting a predetermined portion or all secret data within a secure execution space in response to an operation instruction), wherein the training-data set is provided by the data owner in response to satisfaction of the first-user trustworthiness-criteria determined by the trusted execution environment using the first-user remote-attestation- procedure (Grancharov : Para: 0058-0059, Para: 0115 and Para:012-0124 teaches validating the security state of the TEE before sensitive user data is transmitted and verifying the integrity of the remote platform before provisioning secrets).
16. Regarding claim 8 Grancharov teaches the method , wherein the training-data set comprises a synthetic-data sample generated by the data owner, wherein the synthetic-data sample is: statistically representative of the raw-data set; and distinct from the raw-data set (Para: 0038-0040 teaches training an ML model using training data, the data comprises of synthetic-data i.e., artificially generated data)
17. Regarding claim 9 Grancharov teaches the method, comprising: receiving a synthetic-data generator at the data processing pipeline; processing, in a first-trusted-execution environment of the trusted execution environment, the training-data set, using the synthetic-data generator, to generate a synthetic-data sample, wherein: the training-data set comprises a raw-data sample, wherein the raw-data sample is statistically representative of the raw-data set; and the synthetic-data sample is: statistically representative of the raw-data set; distinct from the raw-data sample and the raw-data set; and suitable for training the data-processing function (Para:0038-0043 teaches ML model using training data, the data comprises of synthetic-data i.e., artificially generated data); wherein the training-data set is provided by the data owner in response to satisfaction of a first-user first-trust-criterion of the first-user trustworthiness-criteria determined by the first-trusted-execution environment using a first-user first-remote-attestation-protocol of the first- user remote-attestation-procedure (Para: 0095, Para:0098, Para: 0115 and Para:0123-0124 teaches verifying the receiving TEE before sensitive information is provisioned. Applying that verification to the TEE receiving the source training data would protect the data owner’s confidential information before processing begins).
18. Regarding claim 10 Grancharov teaches the method, wherein the synthetic-data sample is generated in the first-trusted-execution environment, using the synthetic-data generator, from a plurality of training-data sets, including the training-data set (Para:0038-0043 teaches ML model using training data, the data comprises of synthetic-data i.e., artificially generated data), each respective training- data set provided by a respective data owner in response to satisfaction of a respective user-first-trust-criterion of respective user-trustworthiness-criteria, each respective satisfaction determined by the first-trusted-execution environment using a respective user-first-remote-attestation-protocol of a respective user-remote-attestation-procedure (Para: 0095, Para:0098, Para: 0115 and Para:0123-0124 teaches verifying the receiving TEE before sensitive information is provisioned. Applying that verification to the TEE receiving the source training data would protect the data owner’s confidential information before processing begins. Para:0127 teaches remote attestation-based provisioning associated with a data provider).
19. Regarding claim 10 Grancharov teaches the method, wherein the data-processing function is an untrained-data-processing function provided by the data-processing-function owner in response to satisfaction of a second-user first-trust-criterion of the second-user trustworthiness-criteria, determined by the first-trusted-execution environment using a second-user first-remote- attestation-protocol of the second-user remote-attestation-procedure (Para:0040 teaches training may use large datasets from multiple sources and that the data may belong to multiple parties. Para:0123-0127 teaches verifying the receiving platform before provisioning sensitive user data, including attestation-based provisioning associated with the data provider).
20. Regarding claim 11 Grancharov teaches the method, wherein the data-processing function is an untrained-data-processing function provided by the data-processing-function owner in response to satisfaction of a second-user first-trust-criterion of the second-user trustworthiness-criteria, determined by the first-trusted-execution environment using a second-user first-remote- attestation-protocol of the second-user remote-attestation-procedure (Para:0057 teaches a model generation output ready for deployment into a training environment and distinguishes that output from a trained model ready for production . This suggests providing the data processing function in an untrained state. Fig.7 and Para:0144-0155 teaches remote attestation between model owner’s release tool and the model provisioning server, followed by storage of the model in the server’s TEE upon successful attestation).
21. Regarding claim 18 Grancharov, in view of Zhai teaches the method computer system according to claim 17, wherein the trusted execution environment is established exclusively within the volatile memory (Grancharov: Para:0052 teaches volatile memory, including DRAM, for program and data storage during execution. Zhai : Page. 31, Para.3)
22. Regarding claim 19 Grancharov, in view of Zhai teaches the method computer system according to claim 17 or claim 18, wherein the volatile memory device comprises a random-access memory (RAM) (Grancharov: Para:0052 teaches DRAM, a type of random-access memory, for program and data storage during execution. Para:0095 and Para:0097-0098 teaches application structures and components residing in protected ML-TEE memory.
Zhai: Page. 31, Para.3).
23. Claims 12-15 are rejected under 35 U.S.C. 103 as being unpatentable over Grancharov (US Pub.No.2022/0114249) in view of Zhai (WO 2018/0149110) as applied to claims 9, 11 above and further in view of Swanson (US Pub.No.2017/0177457).
24. Regarding claim 12 Grancharov in view of Zhai teaches the method, comprising: using the first-trusted-execution environment to determine: a trained-data-processing function based on the untrained-data-processing function and the synthetic-data sample; and the analysis results, based on the raw-data set, by using the trained-data- processing function (Grancharov : Para:0057 teaches model training, including online training during model execution), wherein the raw-data set is provided by the data owner in response to satisfaction of a second first-user first-trust-criterion of the first-user trustworthiness-criteria determined by the first-trusted- execution environment using a second first-user first-remote-attestation- protocol of the first-user remote-attestation-procedure (Grancharov: Para:0064 teaches execution of a model within an ML-TEE. Para:0123-0124 teaches verifying the receiving platform before provisioning sensitive user data);
Grancharov, in view of Zhai teaches all the above claimed limitations but fails to teach erasing the trained-data-processing function and the synthetic-data sample from the data processing pipeline.
Swanson teaches erasing the trained-data-processing function and the synthetic-data sample from the data processing pipeline (Fig.5, block 508, Para:0048, Para:0159 and Para:0163 teaches security cleanup that securely erases memory associated with a secure enclave, including cleanup before enclave teardown).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date, to incorporate Swanson’s secure memory cleanup into the Grancharov-Zhai combination after analysis to prevent unauthorized recovery of the training function and synthetic data sample remaining in the TEE’s memory.
25. Regarding claim 13 Grancharov teaches the method, comprising: providing the synthetic-data sample to the data-processing-function owner; in response to satisfaction of a second-user second-trust-criterion of the second-user trustworthiness-criteria, determined by a second-trusted-execution environment, of the trusted execution environment, using a second-user second-remote- attestation-protocol of the second-user remote-attestation-procedure, receiving at the data processing pipeline the data-processing function; wherein the data-processing function is a trained-data-processing function, trained using the synthetic-data-sample by the data-processing-function owner; using the second-trusted-execution environment to determine the analysis results, based on the raw-data set, by using the trained-data-processing function; wherein the raw-data set is provided by the data owner in response to satisfaction of a first-user second-trust-criterion of the first-user trustworthiness-criteria determined by the second-trusted-execution environment using a first-user second-remote-attestation-protocol of the first-user remote-attestation-procedure (Grancharov: Para:0057 teaches offline model training and deployment of a training model for production use. Fig.7, Para:0114-0115 teaches model owner provisioning through remote attestation, followed by storage of the model upon successful attestation. Para:0064 teaches execution of a model within an ML-TEE. Para:0123-0124 teaches verifying the receiving platform before provisioning sensitive user data);
Grancharov, in view of Zhai teaches all the above claimed limitations but fails to expressly teach erasing the synthetic-data sample from the data processing pipeline.
Swanson teaches erasing the synthetic-data sample from the data processing pipeline (Fig.5, block 508, Para:0048, Para:0159 and Para:0163 teaches security cleanup that securely erases memory associated with a secure enclave, including cleanup before enclave teardown).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date, to incorporate Swanson’s secure memory cleanup into the Grancharov-Zhai combination after analysis to prevent unauthorized recovery of synthetic data sample remaining in the TEE’s memory.
26. Regarding claim 14 Grancharov teaches the method, comprising: receiving the synthetic-data sample at a second-trusted-execution environment of the trusted execution environment; in response to satisfaction of a second-user second-trust-criterion of the second-user trustworthiness-criteria, determined by the second-trusted-execution environment, using a second-user second-remote-attestation-protocol of the second-user remote-attestation-procedure, receiving at the data processing pipeline the data- processing function, wherein the data-processing function is an untrained-data- processing function; determining a trained-data-processing function, using the second-trusted-execution environment, based on the synthetic-data sample and the untrained-data- processing function(Grancharov: Para:0057 teaches models ready for deployment into a training environment and online model training. Para:0109 teaches multiple isolated ML-TEE containers. Fig.7, Para:0114-0115 teaches model owner provisioning through remote attestation, followed by storage of the model upon successful attestation. Para:0064 teaches execution of a model within an ML-TEE. Para:0123-0124 teaches verifying the receiving platform before provisioning sensitive user data), providing the trained-data-processing function to the data-processing-function owner (Para:0050 teaches returning updated model parameters to the model owner, as such providing the trained function to its owner);
Grancharov, in view of Zhai teaches all the above claimed limitations but fails to teach erasing the trained-data-processing function and the synthetic-data sample from the data processing pipeline.
Swanson teaches erasing the trained-data-processing function and the synthetic-data sample from the data processing pipeline (Fig.5, block 508, Para:0048, Para:0159 and Para:0163 teaches security cleanup that securely erases memory associated with a secure enclave, including cleanup before enclave teardown).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date, to incorporate Swanson’s secure memory cleanup into the Grancharov-Zhai combination after analysis to prevent unauthorized recovery of the training function and synthetic data sample remaining in the TEE’s memory.
27. Regarding claim 15 Grancharov teaches the method, comprising: in response to satisfaction of a second-user third-trust-criterion of the second-user trustworthiness-criteria, determined by a third-trusted-execution environment, of the trusted execution environment, using a second-user third-remote-attestation- protocol of the second-user remote-attestation-procedure, receiving at the third- trusted-execution environment the trained-data-processing function from the data- processing-function owner; in response to satisfaction of a first-user third-trust-criterion of the first-user trustworthiness-criteria, determined by the third-trusted-execution environment using a first-user third-remote-attestation protocol of the first-user remote- attestation-procedure, receiving at the third-trusted-execution environment the raw-data set; determining, by the third-trusted-execution environment, the analysis results from the raw-data set and the trained-data-processing function (Grancharov: Fig.5, Para:0107 and Para:0109 teaches multiple isolated ML-TEE containers and an orchestrator managing their creation and destruction. Fig.7, Para:0114-0115 teaches model owner provisioning through remote attestation, followed by storage of the model upon successful attestation. Para:0123-0124 teaches verifying before sensitive user data provisioning. Para:0064 teaches execution of a model within an ML-TEE).
Grancharov in view of Zhai teaches all the above claimed limitations but fails to teach erasing the trained-data-processing function from the data processing pipeline.
Swanson teaches erasing the trained-data-processing function from the data processing pipeline (Fig.5, block 508, Para:0048, Para:0159 and Para:0163 teaches security cleanup that securely erases memory associated with a secure enclave, including cleanup before enclave teardown).
Therefore, it would have been obvious to a person of ordinary skill in the art, before the effective filing date, to incorporate Swanson’s secure memory cleanup into the Grancharov-Zhai combination after analysis to prevent unauthorized recovery of the training function remaining in the TEE’s memory.
Conclusion
THIS ACTION IS MADE FINAL. 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 DEREENA T CATTUNGAL whose telephone number is (571)270-0506. The examiner can normally be reached Mon-Fri : 7:30 AM-5 PM 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, Lynn Feild can be reached at 571-272-2092. 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.
/DEREENA T CATTUNGAL/Primary Examiner, Art Unit 2431