DETAILED ACTION
This action is in response to the original filing of 10-11-2023. Claims 1-20 are pending and have been considered below:
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 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 1-20 is/are rejected under 35 U.S.C. 103 as being
unpatentable over Erasing Concepts from Diffusion Models Rohit Gandikota et al. (“Gandikota”), 6-21-2023, pages 1-23 in view of Ullah et al. (“Ullah” 20230118785 A1), Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV) Been Kim et al. (“Kim”), 6-7-2018, pages 1-18 and LoRA: LOW-RANK ADAPTATION OF LARGE LAN GUAGE MODELS, Edward Hu et al. (“Hu”), 10-16-2021, pages 1-26.
Claim 1: Gandikota discloses a computer-implemented method comprising: receiving, by a machine learning service of a, a request to unlearn a concept from a pre-trained generative machine learning model, the request including an identification of the pre-trained generative machine learning model and a description of the concept (abstract, Figure 1); generating, using a language model, a set of negative prompts and a set of positive prompts, the negative prompts including the concept, the positive prompts not including concept (Page 5: Experiment-SD-NEG-PROMPT);
Gandikota may not explicitly disclose receiving, by a machine learning service of a cloud provider network, a request to unlearn. Ullah is provided because it also discloses a request to unlearn (Paragraphs 37-38; unlearn from data/concept) and provides cloud service (Paragraphs 35 and 37) which receives the request to unlearn a feature (Figure 13 and abstract and Paragraph 42).
Therefore it would be obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to use a known technique to improve a similar device and provide cloud service capabilities for processing in the system of Gandikota. One would have been motivated to provide the functionality as an additional method of accessing and computing data for a more expansive and efficient system.
Gandikota also may not explicitly disclose each feature of processing, with the pre-trained generative machine learning model, negative prompts and positive prompts (Gandikota: Page 5, SD-NEG-PROMPT; utilizes prompt capture) to generate associated activation volume maps, wherein an activation volume map for a given prompt includes outputs from one or more layers of the pre-trained generative model (Kim: Page 3 Section 2.2, Figure 8 ; map); identifying a set of activation conditions to differentiate activation volume maps associated with negative prompts from activation volume maps associated with positive prompts (Kim: Figure 1 Page 3, Section 3.2);
Kim as cited above discloses a map functionality to determine activation features showing how the model as trained will respond to inputs, further this allows the concept conditions for activation to be further defined.
Therefore it would have been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to use a known technique to improve a similar device and provide a map functionality process in the system of Gandikota. One would have been motivated to provide the functionality as an additional method for focusing on important features in order to optimize of the model.
Gandikota also may not explicitly disclose generating a model adapter to use a set of different model parameters when processing of a prompt by the pre-trained generative machine learning model satisfies the set of activation conditions. Hu is provided to disclose the prompt processing (Introduction, Figure 1, Page 2: LoRA advantages).
Hu as cited above discloses an adaptation functionality pertaining to activation conditions within the layers, using GPT-3 (prompt based function).
Therefore it would have been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to use a known technique to improve a similar device and provide the adaptation functionality in the system of Gandikota. One would have been motivated to provide the functionality as a simpler training method for reduced storage and processing requirements when optimizing the model.
Claim 2: Gandikota, Ullah, Kim and Hu disclose a computer-implemented method of claim 1, further comprising: executing the pre-trained generative machine learning model and the model adapter by a compute resource of the cloud provider network; receiving a user-submitted prompt; and processing the user-submitted prompt with the pre-trained generative machine learning model, wherein the processing includes: determining, with the model adapter, that the processing of the user-submitted prompt satisfies the set of activation conditions (Hu: Page 6; Prefix-embedding tuning (PreEmbed)/ Prefix-layer tuning (PreLayer) and Page 9 Prompt Engineering and Fine-Tuning/ Parameter-Efficient Adaptation; prompt engineering ensures that prompt perform activation); and using the set of different model parameters to process the user-submitted prompt (Hu: Page 9 Prompt Engineering and Fine-Tuning).
Claim 3: Gandikota, Ullah, Kim and Hu disclose a computer-implemented method of claim 1, wherein the set of different model parameters are based on at least one of zeroing out a corresponding set of pre-trained model weights, adding random noise to the corresponding set of pre-trained model weights, or adding a set of predetermined weights to the corresponding set of pre-trained model weights to bias the pre- trained generative machine learning model from the concept to another concept (Kim: Figure 8 and Page 7, Paragraph 2; noise added).
Claim 4 is similar in scope to claim 1 and therefore rejected under the same rationale.
Claim 5 is similar in scope to claim 2 and therefore rejected under the same rationale.
Claim 6: Gandikota, Ullah, Kim and Hu disclose a computer-implemented method of claim 5, wherein using the set of different model parameters to process the user-submitted prompt includes at least one of updating pre-trained model parameters with a set of update parameters (Hu: Page 2: LoRA advantages; trainable/updated matrices (parameters)) or redirecting an output from a layer of the pre- trained generative machine learning model through a different layer including the set of different model parameters.
Claim 7: Gandikota, Ullah, Kim and Hu disclose a computer-implemented method of claim 4, wherein the set of different model parameters are based on at least one of zeroing out a corresponding set of pre-trained model weights of the pre-trained generative machine learning model, adding random noise to the corresponding set of pre-trained model weights, or adding a set of predetermined weights to the corresponding set of pre-trained model weights to bias the pre-trained generative machine learning model from the concept to another concept (Kim: Figure 8 and Page 7, Paragraph 2; noise added).
Claim 8: Gandikota, Ullah, Kim and Hu disclose a computer-implemented method of claim 4, further comprising receiving an indication of a degree to which to unlearn the concept, wherein at least a portion of the set of different model parameters are based on the degree to which to unlearn the concept(Gandikota: Page 12: B.3, C and D; Erasure or Removal from diffusion can affect model parameters and determine a degree of unlearning).
Claim 9: Gandikota, Ullah, Kim and Hu disclose a computer-implemented method of claim 4, further comprising identifying, using a gradient-based activation mapping (Kim: Page 3, section 2.2 saliency maps are gradient based), a first region of a model activation volume sensitive to a first negative prompt and a second region of the model activation volume sensitive to a first positive negative prompt, and wherein the set of conditions are localized to activations within the first and second regions of the model activation volume (Kim: Page 3 Section 2.2 and 3.2, Figure 8; saliency maps importance).
Claim 10: Gandikota, Ullah, Kim and Hu disclose a computer-implemented method of claim 4, further comprising generating, using a language model, at least some of the negative prompts and at least some of the positive prompts, the negative prompts including the concept, the positive prompts not including concept (Gandikota: Page 5, SD-NEG-PROMPT).
Claim 11: Gandikota, Ullah, Kim and Hu disclose a computer-implemented method of claim 10, further comprising: receiving, from the language model, an indication of a related concept and an indication of a correlation between the concept to be unlearned and the related concept; determining that the indication of the correlation satisfies a threshold; generating, with the pre-trained generative machine learning model, an output based on a prompt that includes the related concept (Gandikota: Page 12, Section D and Figure C3; concepts and related concepts for erasure/unlearn); receiving a user-submitted classification of the output as a positive sample; and classifying the prompt that includes the related concept as a positive prompt(Gandikota: Page 12, Section D and Figure C3; concepts and related concepts are presented as “positive” prompts to erase and Page 14, Section E.2; looks at nudity/inappropriate and maps positive samples of any to erase).
Claim 12: Gandikota, Ullah, Kim and Hu disclose a computer-implemented method of claim 4 performed by a machine learning service of a cloud provider network (Ullah: Paragraph 42).
Claim 13: Gandikota, Ullah, Kim and Hu disclose a computer-implemented method of claim 4, wherein the description of the concept to be unlearned was received from an electronic device and further comprising (Gandikota: Page 3, Figure 2 and Paragraph 1; description provided): sending the model adapter to the electronic device, whereby the model adapter can be deployed with the pre-trained generative machine learning model to reduce a likelihood of the concept to be unlearned from appearing in an output of the pre- trained generative machine learning model (Gandikota: Page 2, Paragraph 3; ESD provides parameter adjustments (adapter) Kim: Page 2: customizations adapts concepts).
Claim 14 is similar in scope to claim 1 and therefore rejected under the same rationale.
System (Ullah: Figure 4 and Paragraph 42)
Regarding the multi-tenant provider (Ullah: Paragraph 44; cloud service to many users)
Regarding a second one or more computing devices to implement unlearning (Ullah: Paragraph 42 and 44; network of devices used in the unlearning process) Ullah also discloses a request to unlearn a concept/data (Paragraph 37-38)
Claim 15: Gandikota, Ullah, Kim and Hu disclose a system of claim 14, further comprising: a third one or more computing devices to execute the pre-trained generative machine learning model with the model adapter in the multi-tenant provider network, wherein an environment to execute the pre-trained generative machine learning model with the model adapter includes instructions to: receive a user-submitted prompt; and process the user-submitted prompt with the pre-trained generative machine learning model, wherein the instructions to process include instructions to: determine, with the model adapter, that the processing of the user-submitted prompt satisfies the set of conditions; and use the set of different model parameters to process the user-submitted prompt.
(Hu: Page 6; Prefix-embedding tuning (PreEmbed)/ Prefix-layer tuning (PreLayer) and Page 9 Prompt Engineering and Fine-Tuning/ Parameter-Efficient Adaptation; prompt engineering ensures that prompt perform activation).
Claim 16: Gandikota, Ullah, Kim and Hu disclose a system of claim 15, wherein the instructions to use the set of different model parameters to process the user-submitted prompt include at least one of instructions to update pre- trained model parameters with a set of update parameters (Hu: Page 2: LoRA advantages; trainable/updated matrices (parameters)) or instructions to redirect an output from a layer of the pre-trained generative machine learning model through a different layer including the set of different model parameters.
Claim 17: Gandikota, Ullah, Kim and Hu disclose a system of claim 14, wherein the set of different model parameters are based on at least one of zeroing out a corresponding set of pre-trained model weights of the pre-trained generative machine learning model, adding random noise to the corresponding set of pre-trained model weights, or adding a set of predetermined weights to the corresponding set of pre-trained model weights to bias the pre-trained generative machine learning model from the concept to another concept(Kim: Figure 8 and Page 7, paragraph 2; noise added).
Claim 18: Gandikota, Ullah, Kim and Hu disclose a system of claim 14, wherein the model unlearning training service includes further instructions that upon execution cause the model unlearning training service to receive an indication of a degree to which to unlearn the concept, wherein at least a portion of the set of different model parameters are based on the degree to which to unlearn the concept (Gandikota: Pages 11-13; provides inappropriate classes (Table C.1 and E.2.) and further defines individual classes and styles not to learn (i.e. objects/art styles)).
Claim 19: Gandikota, Ullah, Kim and Hu disclose a system of claim 14, wherein the model unlearning training service includes further instructions that upon execution cause the model unlearning training service to identify, using a gradient-based activation mapping (Kim: Page 3, section 2.2 saliency maps are gradient based), a first region of a model activation volume sensitive to a first negative prompt and a second region of the model activation volume sensitive to a first positive negative prompt, and wherein the set of conditions are localized to activations within the first and second regions of the model activation volume (Kim: Page 3 Section 2.2 and 3.2, Figure 8; saliency maps importance).
Claim 20: Gandikota, Ullah, Kim and Hu disclose a system of claim 14, wherein the model unlearning training service includes further instructions that upon execution cause the model unlearning training service to generate, using a language model, at least some of the negative prompts and at least some of the positive prompts, the negative prompts including the concept, the positive prompts not including concept (Gandikota: Page 11; negative guidance).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
20240070525 A1 0042
Applicant is required under 37 C.F.R. § 1.111(c) to consider these references fully when responding to this action.
It is noted that any citation to specific pages, columns, lines, or figures in the prior art references and any interpretation of the references should not be considered to be limiting in any way. A reference is relevant for all it contains and may be relied upon for all that it would have reasonably suggested to one having ordinary skill in the art. In re Heck, 699 F.2d 1331, 1332-33, 216 U.S.P.Q. 1038, 1039 (Fed. Cir. 1983) (quoting In re Lemelson, 397 F.2d 1006, 1009, 158 U.S.P.Q. 275, 277 (C.C.P.A. 1968)).
In the interests of compact prosecution, Applicant is invited to contact the examiner via electronic media pursuant to USPTO policy outlined MPEP § 502.03. All electronic communication must be authorized in writing. Applicant may wish to file an Internet Communications Authorization Form PTO/SB/439. Applicant may wish to request an interview using the Interview Practice website: http://www.uspto.gov/patent/laws-and-regulations/interview-practice.
Applicant is reminded Internet e-mail may not be used for communication for matters under 35 U.S.C. § 132 or which otherwise require a signature. A reply to an Office action may NOT be communicated by Applicant to the USPTO via Internet e-mail. If such a reply is submitted by Applicant via Internet e-mail, a paper copy will be placed in the appropriate patent application file with an indication that the reply is NOT ENTERED. See MPEP § 502.03(II).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SHERROD KEATON whose telephone number is 571-270-1697. The examiner can normally be reached 9:30am to 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 MICHELLE BECHTOLD can be reached at 571-431-0762. 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.
/SHERROD L KEATON/Primary Examiner, Art Unit 2148
5-27-2026
Prosecution Insights