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
Applicant’s arguments, filed 02/26/2026, with respect to claims 1-18 have been fully considered and are persuasive. The rejections of claims 1-18 has been withdrawn.
Applicant argues the claims are directed to a practical application because of the improvement recited by the current application’s paragraphs [0044]-[0046].
MPEP 2106.4(a)(2) Abstract Idea Groupings I. Mathematical concepts: The Court’s rationale for identifying these "mathematical concepts" as judicial exceptions is that a ‘‘mathematical formula as such is not accorded the protection of our patent laws,’’ Diehr, 450 U.S. at 191, 209 USPQ at 15 (citing Benson, 409 U.S. 63, 175 USPQ 673), and thus ‘‘the discovery of [a mathematical formula] cannot support a patent unless there is some other inventive concept in its application.’’ Flook, 437 U.S. at 594, 198 USPQ at 199.
As shown above, a claim employing a mathematical algorithm does direct the claims to an abstract idea unless “there is some other inventive concept in its application”. Currently, the claims state a first and second cost function, a mathematical algorithm.
To identify the inventive concept, in view of the claims, we look at the updated subject matter analysis provided by Ex Parte Desjardins (December 5, 2025), herein after referred to as Desjardins, and 2024 Guidance Update on Patent Subject Matter Eligibility, Including on Artificial Intelligence published July 17, 2024 (89 FR 58128) (AI-SME Update). Desjardins, page 1 “2)”, states “improvements in computational performance, learning, storage, data set and structures, for example, can constitute patent-eligible technological advancements under the Alice framework.” In order to identify the improvement as being an eligible improvement in technology “ when the specification provides sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing a technological improvement and the claims reflect the disclosed improvement. The specification need not explicitly set forth the improvement, so long as the specification describes the invention such that the improvement would be apparent to one of ordinary skill in the art.” (Desjardin page 3)(MPEP2106.05(a)).
Therefore, establishing whether the claimed mathematical concepts are directed to an abstract idea is tied to whether the inventive concept is claimed or reflected by the claims. Applicant’s argument provides paragraphs [0044]-[0046] as the improvement to the technology. Applicant states the improvement as “…fast[er] convergence and results in more accurate model than a one-step training process involving a single cost function.” The claims state to use a first and second cost function for training the model. The use of the two costs functions is explicitly described by the specification to produce a particular improvement of faster convergence. Therefore, the claims are deemed to reflect the improvement to the technology and provide a practical application. The 101 rejection is withdrawn.
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, 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.
Claim 21 are rejected under 35 U.S.C. 103 as being unpatentable over Middlebrooks et al. (US 2016/0313651 A1) in view of Lee (US 2021/0125075 A1).
With respect to Claim 21, Middlebrooks’651 shows a computer program product comprising a non-transitory computer readable medium having instructions therein, the instructions, when executed by a computer system (paragraphs [0073] and [0075]), configured to cause the computer system to at least:
train a machine learning model based on a first cost function to obtain a first set of model parameter values of the machine learning model (Paragraphs [0042], [0046]-[0047] and [0049] to train the model 513 with process parameters 511, layout/design parameters 512; paragraph [0130] wherein the layout/design parameters/variables formulate a function to define the cost function), wherein the machine learning model is configured to predict values of a physical characteristic associated with a physical substrate and for use in adjusting a patterning process (Figures 5-6 and paragraphs [0033], [0037], [0044] and [0050] a classification model 513 to predict defects 514 in an etch image, a pattern transferred to a layer of the substrate by etching using the resist thereon as a mask, and a correction step 515 to adjust the parameters to reduce or eliminate defects. Paragraphs [0042], [0046]-[0047] and [0049] to train the model 513 with process parameters 511, layout/design parameters 512, and data 516) and wherein the first cost function represents a physical characteristic of the patterning process (Paragraph [0130] to minimize the cost function of the design variables for low k photolithography (paragraph [0122] the fine-tuned design layout parameters to resemble the physical planned/intended pattern characteristics)) or an error caused by a modeling process; and
train the machine learning model based on the first set of parameter values of the machine learning model (Paragraphs [0042], [0046]-[0047] and [0049] to train the model 513 with process parameters 511, layout/design parameters 512; paragraph [0130] wherein the layout/design parameters/variables formulate a function to define the cost function) and [ ].
Middlebrooks’651 does not show on a second cost function to obtain a second set of model parameter values of the machine learning model, wherein the second cost function represents a physical characteristic of the patterning process or an error caused by a modeling process, different from the first cost function.
Lee’075 shows processor may train a classification model using a second cost function during a second period (paragraph [0179]), wherein the second cost function represents an error caused by a modeling process (Figure 11 and paragraphs [0196]-[0197] that an error is determined by applying the second cost function), different from the first cost function (Paragraphs [0171] the equations of the first and second cost function are different and respectively regard 1) modifying plurality of parameters of the classification model based on a difference between a label and inference value (similar to Middlebrook) and 2) a probability of inferring if real/simulated data is applied).
At the time of the invention, it would have been obvious to one of ordinary skill in the art before the effective filling date of the claim invention to modify Middlebrooks’651 to include a second cost function to obtain a second set of model parameter values of the machine learning model, wherein the second cost function represents a physical characteristic of the patterning process or an error caused by a modeling process, different from the first cost function method taught by Lee’075. The suggestion/motivation for doing so would have been to improve the system’s ability to be able to benefits of training the model on additional data gaining the ability to predict a classification result for data outside of the original training data (paragraph [0180]).
Claim 22 are rejected under 35 U.S.C. 103 as being unpatentable over Middlebrooks et al. (US 2016/0313651 A1) in view of Lee (US 2021/0125075 A1) further in view of Hashimoto et al. (US 2018/0157167 A1).
With respect to Claim 22, Middlebrooks’651 and Lee’167 do not specifically show the computer program product of claim 21, wherein the physical characteristic comprises a difference between a reference image of a pattern as produced in the patterning process and an image of the pattern as produced in the patterning process generated via the machine learning model, or wherein the physical characteristic comprises a difference between measured values of a parameter associated with the pattern and predicted values of the parameter associated with the pattern generated via the machine learning model, or wherein the error caused by a modeling process comprises grid dependency.
Hashimoto’167 shows wherein the error caused by a modeling process comprises a grid dependency (paragraph [0030] describes a cost function related to an error between the prediction shape and hypothetical target shape is calculated in terms of the positional shift between an edge of the prediction shape and edge of the hypothetical target shape (describing grid dependency)).
At the time of the invention, it would have been obvious to one of ordinary skill in the art before the effective filling date of the claim invention to modify Middlebrooks’651 and Lee’075 to include the error caused by a modeling process comprises a grid dependency method taught by Hashimoto’167. The suggestion/motivation for doing so would have been to improve highly accurately patterns (paragraph [0045]).
Allowable Subject Matter
Claims 1-18 are allowed.
The following is an examiner’s statement of reasons for allowance: none of the references either singularly or in combination teach or fairly suggest a method of training a machine learning model configured to predict values of a physical characteristic associated with a substrate and for use in adjusting a patterning process, the method comprising: obtaining a reference image associated with a desired pattern to be printed on the substrate; determining a set of model parameter values of the machine learning model such that a first cost function is reduced from a value of the first cost function obtained by using initial model parameter values, wherein the first cost function represents a difference between the reference image and an image generated via the machine learning model; and training, by using the set of model parameter values, the machine learning model based on a combination of the first cost function and a second cost function, wherein the second cost function represents a difference between measured values and predicted values of the physical characteristic associated with the desired pattern, wherein the predicted values are predicted via the machine learning model.
Midlebrooks’651 shows in figures 5-6 and paragraphs [0033], [0037], [0044] and [0050] a classification model 513 to predict defects 514 in an etch image, a pattern transferred to a layer of the substrate by etching using the resist thereon as a mask, and a correction step 515 to adjust the parameters to reduce or eliminate defects. Paragraphs [0042], [0046]-[0047] and [0049] describes to train the model 513 with process parameters 511, layout/design parameters 512, and data 516. Paragraph [0067] describes that the model may use pattern layout data. Paragraphs [0122]-[0127] relates the patterns in design layout to patterns intended to print on the substrate. Paragraph [0130] describes to minimize the cost function of the design variables for low k photolithography (paragraph [0122] the fine-tuned design layout parameters to resemble the planned/intended pattern. Paragraphs [0042], [0046]-[0047] and [0049] describes to train the model 513 with process parameters 511, layout/design parameters 512; paragraph [0130] wherein the layout/design parameters/variables formulate a function to define the cost function. Midlebrooks’651 do not include all the detailed combined limitations included in the claim including a method of training a machine learning model configured to predict values of a physical characteristic associated with a substrate and for use in adjusting a patterning process, the method comprising: obtaining a reference image associated with a desired pattern to be printed on the substrate; determining a set of model parameter values of the machine learning model such that a first cost function is reduced from a value of the first cost function obtained by using initial model parameter values, wherein the first cost function represents a difference between the reference image and an image generated via the machine learning model; and training, by using the set of model parameter values, the machine learning model based on a combination of the first cost function and a second cost function, wherein the second cost function represents a difference between measured values and predicted values of the physical characteristic associated with the desired pattern, wherein the predicted values are predicted via the machine learning model, therefore this claim is allowable.
Lee’075 shows a processor may train a classification model using a second cost function during a second period in paragraph [0179]. Lee’075 do not include all the detailed combined limitations included in the claim including a method of training a machine learning model configured to predict values of a physical characteristic associated with a substrate and for use in adjusting a patterning process, the method comprising: obtaining a reference image associated with a desired pattern to be printed on the substrate; determining a set of model parameter values of the machine learning model such that a first cost function is reduced from a value of the first cost function obtained by using initial model parameter values, wherein the first cost function represents a difference between the reference image and an image generated via the machine learning model; and training, by using the set of model parameter values, the machine learning model based on a combination of the first cost function and a second cost function, wherein the second cost function represents a difference between measured values and predicted values of the physical characteristic associated with the desired pattern, wherein the predicted values are predicted via the machine learning model, therefore this claim is allowable.
Hashimoto’167 shows in paragraph [0030] a cost function related to an error between the prediction shape and hypothetical target shape is calculated in terms of the positional shift between an edge of the prediction shape and edge of the hypothetical target shape (describing grid dependency). Hashimoto’167 do not include all the detailed combined limitations included in the claim including a method of training a machine learning model configured to predict values of a physical characteristic associated with a substrate and for use in adjusting a patterning process, the method comprising: obtaining a reference image associated with a desired pattern to be printed on the substrate; determining a set of model parameter values of the machine learning model such that a first cost function is reduced from a value of the first cost function obtained by using initial model parameter values, wherein the first cost function represents a difference between the reference image and an image generated via the machine learning model; and training, by using the set of model parameter values, the machine learning model based on a combination of the first cost function and a second cost function, wherein the second cost function represents a difference between measured values and predicted values of the physical characteristic associated with the desired pattern, wherein the predicted values are predicted via the machine learning model, therefore this claim is allowable.
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.”
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 IRIANA CRUZ whose telephone number is (571)270-3246. The examiner can normally be reached 10-6.
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, Akwasi M. Sarpong can be reached at (571) 270-3438. 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.
/IRIANA CRUZ/Primary Examiner, Art Unit 2681