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 .
DETAILED ACTION
This non-final office action is responsive to the U.S. patent application no. 18/724,145 filed on June 25, 2024.
Claims 1-18 are pending.
Claims 1-18 are rejected.
Priority
The application claims priority under 35 U.S.C. 365(a) to the international application no. PCT/CN2021/142842 filed on December 30, 2021.
Information Disclosure Statement
The information disclosure statements (IDS) submitted on August 12, 2024, June 27, 2025 and October 22, 2025 are compliant with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statements have been considered by the examiner.
Claim Objections
Claim 2 is objected to because it recites “the second input including a base model that is to be fine-tuned in the fine-tune to generate the fine-tuned model.” The underlined portion appears to have a minor grammatical error.
Claim 8 has the same issue as claim 2.
Claim 3 is objected to because it recites in the first clause “the second input including a fine-tuning computer-executable instructions” that appears to have a minor grammatical error. Appropriate correction is required.
Claim 9 has the same issue a claim 3.
Claim 6 and claim 16 appear to be identical, therefore one should be cancelled.
Claim Rejections - 35 USC § 102
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)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 1-5 and 7-12 are rejected under 35 U.S.C. 102(a)(2) as being unpatentable over Li et al. (U.S. 2021/018266).
Regarding claim 1, Li disclosed a computing system comprising:
one or more processors; and
one or more computer-readable media having thereon computer-executable instructions that are structured such that, if executed by the one or more processors (Li, [0050], “Referring to FIG. 1, a distributed training system 100 is shown in accordance with an embodiment of the invention. The system 100 includes a master device 110 and a slave device 120. The master device 110 and the slave device 120 include computing devices, i.e. devices with at least one processor and a memory wherein the at least one processor is configured to execute computer program code loaded into the memory to perform one or more functions”), the computing system would be configured to fine-tune a machine learning model in a protected environment such that a tune initiator system that instruct that the fine-tuning occur does not having visibility of a resulting fine-tuned model or at least some information used in the fine-tuning (Li, [0023], “a method of training a neural network model includes: generating, at a master device, first configuration data for the neural network model based on a first version of the neural network model; sending, from the master device, the first configuration data to a slave device, the slave device being remote from the master device and having access to a first data source that is inaccessible by the master device;” see [0028, 0029, 0066, 0084, 0086, 0088] for additional disclosure in this aspect), by being configured to perform the following:
receive first input into the protected environment from the tune initiator system via a first channel that is visible to the tune initiator system, the first input including training data (Li, Figs. 2, 3 and 8, and [0018], “receiving, at a slave device, first configuration data for the neural network model from a master device, the master device being remote from the slave device” And [0088], “ The master device 710 is configured to transmit first configuration data 780 for distribution to the plurality of slave devices 722, 724, and 726 via one or more communication networks 750.” Li’s salve device is an embodiment of the protected environment in the claim, and Li’s master device performs the function of the tune initiator system in the claim and Li’s first configuration data is an example of the first input);
access second input over a second channel that is not visible to the tune initiator system (Li, [0018], “training, at the slave device, the second version of the neural network model using data from a first data source, the first data source being inaccessible by the master device.” Li’s “data from a first data source” anticipates the “second input” in the claim);
use the first input and the second input to fine-tune a machine learning model to thereby form a fine-tuned machine learning model (Li, [0018], “instantiating, at the slave device, a second version of the neural network model using the first configuration data; training, at the slave device, the second version of the neural network model using data from a first data source, the first data source being inaccessible by the master device”); and
store the fine-tuned machine learning model in the protected environment such that the fine-tuned machine learning model is available for the tune initiator system to provide input data to and receive output data from, but such that the fine-tuned machine learning model cannot be directly accessed by the tune initiator system (Li, [0018, 0068], “Following training on data from the slave data source 230, the slave device 220 stores an updated set of parameters in the storage device 272 and use these to generate a second configuration data (CD2) 290 that is sent to the master device 210.” Li’s “second version of the neural network model” anticipates the ”fine-tuned machine learning model” in the claim and Li’s second configuration data CD2 anticipates the “output data” from the fine-tuned machine learning model).
Claim 7 lists substantially the same elements as claim 1, but in method form rather than system form. Therefore, the rejection rationale for claim 1 applies equally as well to claim 7.
Regarding claims 2 and 8, Li disclosed the subject matter in accordance with Claims 1 and 7, respectively.
Li further disclosed that the second input including a base model that is to be fine-tuned in the fine-tune to generate the fine-tuned model (Li disclosed in [0073] that “The first version of the neural network model 360 is instantiated for use as a teacher model. The parameters of the first version of the neural network model 360 are fixed ” and “An output of the first version of the neural network model 360 is communicated to the second version of the neural network model 370 and is used to train the second version of the neural network model 370.”).
Regarding claims 3 and 9, Li disclosed the subject matter in accordance with Claims 2 and 8, respectively.
Li further disclosed the second input including a fine-tuning computer-executable instructions, the computing system further being configured to perform the following: execute the fine-tuning computer-executable instructions by one or more processors of a computing system to cause the computing system to use the base model accessed over the second channel and the training data received over the first channel to form the fine-tuned machine learning model (Li disclosed in [0071] that “as indicated by the dotted line, one or more of hyperparameters and parameters from a first configuration data (CD.sub.1) 380 may also be used to instantiate the second version of the neural network model 370.” and in [0073] that “An output of the first version of the neural network model 360 is communicated to the second version of the neural network model 370 and is used to train the second version of the neural network model 370. As such, the teacher model is used to train the student model.”).
Regarding claims 4 and 10, Li disclosed the subject matter in accordance with Claims 2 and 8, respectively.
Li further disclosed that the first input including fine-tuning computer-executable instructions, the computing system further configured to perform the following: execute the fine-tuning computer-executable instructions by one or more processors of a computing system to cause the computing system to use the base model accessed over the second channel and the training data received over the first channel to form the fine-tuned machine learning model (Li disclosed in [0072] that “the slave device 320 is configured to use first configuration data 380 that originates from a master device to instantiate a first version of the neural network model 360 within the slave device 320.” Li additionally disclosed in [0016] that “the first configuration data includes hyperparameters for the neural network model and parameters for the first version of the neural network model….”).
Regarding claims 5 and 11, Li disclosed the subject matter in accordance with Claims 1 and 7, respectively.
Li further disclosed that the first input including a base model that is to be fine-tuned in the fine-tune to generate the fine-tuned model (Li, Figs. 2, 3 and 8, and [0018], “receiving, at a slave device, first configuration data for the neural network model from a master device, the master device being remote from the slave device” said neural network model from a master device anticipates the base model in the claim).
Regarding claim 12, Li disclosed the method in accordance with Claim 11,
Li further disclosed that the second input received over the second channel including a fine-tuning computer-executable instructions, the method further comprising:
executing the fine-tuning computer-executable instructions by one or more processors of a computing system to cause the computing system to use the base model accessed over the first channel and the training data received over the first channel (Li, [0073], “Data from the first data source 330 is input to both the first version and second version of the neural network model 360 and 370, respectively. The data may be one or more of labelled and unlabeled data. An output of the first version of the neural network model 360 is communicated to the second version of the neural network model 370 and is used to train the second version of the neural network model 370. As such, the teacher model is used to train the student model.”).
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 of this title, 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 6, 13 and 16 are rejected under 35 U.S.C. 103 as obvious over Li et al. (U.S. 2021/0182661).
Regarding claims 6, 13 and 16, Li disclosed the subject matter in accordance with claims 1 and 7 respectively.
Li further disclosed wherein using the first input and the second input to fine-tune a machine learning model is performed using a plurality of containers (Li, [0014], “the executed binary executable is configured to load the first configuration data and instantiate a second version of the neural network model independently of the master device. The executed binary executable may be configured to output the second configuration data and to control transmission to the master device. A binary executable may provide an efficient container that may be distributed to the slave devices (e.g. either by non-transitory media or transmitted data).”).
Li might not have explicitly disclosed that
a first subset of the containers containing code provided by the tune initiator system, a second subset of the containers containing code not provided by the external network entity and which prevents the first subset of containers from accessing the Internet.
However, given Li’s disclosure in [0014] that “the slave device includes at least one processor to execute a binary executable stored in memory… A binary executable may provide an efficient container that may be distributed to the slave devices” regarding the use of “container”, one of ordinary skill in the art would have found the subject matter in the instant claim obvious before the effective filing date of the claimed invention as it is well-known that “containers” provide security boundaries for binary executables.
Claims 14-15 and 17-18 are rejected under 35 U.S.C. 103 as obvious over Li et al. (U.S. 2021/0182661) in view of Uehara (U.S. 2023/0004785)
Regarding claims 14 and 17, Li disclosed the subject matter in accordance with Claims 13 and 16.
Li might not have explicitly disclosed but Uehara disclosed that the first subset of containers operating within an overlay network operating within a virtual network (Uehara, [0068], “The integration server 30 may be on a computer network on which the development entity 80 of the AI model has access rights, and a form of the server may be a physical server, a virtual server, or the like”).
One of ordinary skill in the art would have been motivated to combine Li and Uehara because both references disclosed methods and systems for training/fine-tuning private slave/local models on client devices using a master model on a central server (Li, Fig. 1 and Abstract; Uehara, Fig. 1 and Abstract). Therefore, before the effective filing date of the claimed invention, it would have been obvious for one of ordinary skill in the art to read Li’s disclosure in light of Uehara’s teaching of using virtual servers to run the integration server to realize that the master device in Li could be implemented using virtual servers as was already common-place at that time.
Regarding claims 15 and 18, Li and Uehara disclosed the subject matter in accordance with Claims 14 and 17.
Li might not have explicitly disclosed but Uehara disclosed that the code in the second subset of containers communicating outside of the virtual network using private endpoints (Uehara, [0054], “ …a terminal 20, which is installed on a network in each medical institution of a plurality of medical institutions, and an integration server 30. The terminal 20 refers to a computing resource existing in a network in which data in a medical institution can be safely accessed, … the terminal 20 in each medical institution may be a physical machine or a virtual machine, …”).
The motivation for combining Li and Uehara is substantially the same as that provided in the rejection rationale for claims 14 and 17 above.
Relevant Prior Art
Bhowmick et al. (US 2021/0166157) is directed to a method and system for private federated learning with protection against reconstruction, which provides privacy protection to local data and models at client devices.
Chu et al. (US 2021/0073678) disclosed a method and a system for machine learning using data obtained from fragmented and isolated data owners to train a machine learning model while preserving the privacy, confidentiality, and security of the data.
Sav et al. (US 2023/0325529) disclosed a method and a system for privacy-preserving distributed training of a global neural network model on distributed datasets (DS1 to DSn).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SHIRLEY X ZHANG whose telephone number is (571)270-5012. The examiner can normally be reached 8:30am - 5:00pm.
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, Joon H Hwang can be reached at 571-272-4036. 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.
/SHIRLEY X ZHANG/Primary Examiner, Art Unit 2447