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 .
This action is in response to the amendment filed 10/10/2025. Claims 1, 9 and 17 have been amended. 1-20 are pending and have been examined.
Claim Rejections - 35 USC § 101
35 U.S.C. § 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. § 101 because the claimed invention is directed
to an abstract idea without significantly more.
Regarding Claim 1:
Step 1: Claim 1 recites a method including steps which falls into the statutory category of a process.
Step 2A Prong 1: Claim 1 recites multiple mathematical processes, as explained below.
computing a regularization term using a predefined quantity of random samples and the Jacobian matrix of the DEQ, at an equilibrium point z*, the regularization term penalizing the spectral radius of the Jacobian matrix to stabilize fixed-point iterations; including the regularization term in an original loss function of the DEQ to form a regularized loss function that conditions both forward and backward fixed-point systems; computing a gradient of the regularized loss function with respect to model parameters of the DEQ; and using the gradient to update the model parameters wherein the regularization reduces computational cost of fixed-point solving. These limitations recite the abstract idea of mathematical calculations. (MPEP 2016.4(a)(2))
Step 2A Prong 2: Claim 1 does not integrate the abstract idea into a practical application, as the additional elements of A system for regularized training of a Deep Equilibrium Model (DEQ), and during both training and inference phases of the DEQ merely indicate a field of use or technological environment in which to apply a judicial exception. (MPEP 2106.05(h)) Accordingly, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The claim is directed to an abstract idea.
Step 2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. The additional elements of A system for regularized training of a Deep Equilibrium Model (DEQ), and during both training and inference phases of the DEQ do not amount to an inventive step for the reason set forth in Step 2A Prong 2. Thus, this claim is ineligible.
Regarding Claim 2, dependent upon Claim 1, recites the predefined quantity is defined for approximating a Frobenius norm of the DEQ, and further comprising regularizing the Jacobian matrix using the Frobenius norm of the Jacobian matrix, which describes a mathematical calculation. The claim does not include additional elements that are sufficient to amount to an integration of the abstract idea into practical application or significantly more than the judicial exception. Thus, the claim is ineligible.
Regarding Claim 3, dependent upon Claim 2, recites further comprising using a Hutchinson estimator to estimate the Frobenius norm, which describes a mathematical calculation. The claim does not include additional elements that are sufficient to amount to an integration of the abstract idea into practical application or significantly more than the judicial exception. Thus, the claim is ineligible.
Regarding Claim 4, dependent upon Claim 3, recites further comprising approximating the Hutchinson estimator using Monte-Carlo estimation, which further elaborates the mathematical calculation of the claim 3. The claim does not include additional elements that are sufficient to amount to an integration of the abstract idea into practical application or significantly more than the judicial exception. Thus, the claim is ineligible.
Regarding Claim 5, dependent upon Claim 2, recites further comprising applying an
upper bound on the spectral radius of the Jacobian matrix to estimate the Frobenius norm, which is directed to the abstract idea of mental process (including an observation, evaluation, judgement, opinion). The claim does not include additional elements that are sufficient to amount to an integration of the abstract idea into practical application or significantly more than the judicial exception. Thus, the claim is ineligible.
Regarding Claim 6, dependent upon Claim 1, recites wherein the regularization term is weighted in the regularized loss function according to a predefined coefficient, the coefficient controlling a relative importance of the regularization term in the regularized loss function, which is directed to the abstract idea of a mental process (including an observation, evaluation, judgement, opinion). The claim does not include additional elements that are sufficient to amount to an integration of the abstract idea into practical application or significantly more than the judicial exception. Thus, the claim is ineligible.
Regarding Claim 7, dependent upon Claim 1, recites an additional element of further comprising iteratively performing the operations of claim 1 for a plurality of training cycles.
The additional element of claim 7 is an insignificant extra-solution activity (performing repetitive calculations), further considered well-understood, routine and conventional (WURC) in particular field. (MPEP 2106.05(d)(II)(ii) and MPEP 2106.05(g))
Regarding Claim 8, dependent upon Claim 1, recites an additional element of further comprising utilizing the DEQ for sequence prediction, language modeling, computer vision tasks, image classification, and/or semantic segmentation, which neither integrate the abstract idea into a practical application, nor provide significantly more than the abstract idea itself. Specifically, the claims recite mere instructions to apply to an exception. (MPEP 2106.05(f))
Regarding Claim 9,
Step 1: Claim 9 recites a system falls into the statutory category of a machine.
Step 2A Prong 1: Claim 9 recites multiple mathematical concepts, as explained below.
to compute a regularization term using a predefined quantity of random samples and the Jacobian matrix of the DEQ at an equilibrium point z*, the regularization term penalizing the spectral radius of the Jacobian matrix to stabilize fixed-point iterations, include the regularization term in an original loss function of the DEQ to form a regularized loss function that conditions both forward and backward fixed- point systems, compute a gradient of the regularized loss function with respect to model parameters of the DEQ, and use the gradient to update the model parameters, wherein the regularization reduces computational cost of fixed-point solving.
The claim is directed to the abstract idea of mathematical calculations involving the Jacobian matrices, gradient computation, loss functions, and optimization processes. (MPEP 2016.4(a)(2))
Step 2A Prong 2: Claim 9 does not integrate the abstract idea into a practical application, as the additional elements of A system for regularized training of a Deep Equilibrium Model (DEQ), comprising: one or more computing devices programmed and during both training and inference phases of the DEQ. These limitations, namely, for regularized training of a Deep Equilibrium Model (DEQ), computing devices and during both training and inference phases of the DEQ, merely indicate a field of use or technological environment in which to apply a judicial exception. (MPEP 2106.05(h))
Accordingly, the additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The claim is directed to an abstract idea.
Step 2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements do not amount to an inventive step for the reason set forth in Step 2A Prong 2. Thus, this claim is ineligible.
Regarding Claim 10, dependent upon Claim 9, recites wherein the predefined quantity
is defined for approximating a Frobenius norm of the DEQ, to regularized the Jacobian matrix using the Frobenius norm of the Jacobian matrix, which describes a mathematical calculation. The claim does not include additional elements that are sufficient to amount to an integration of the abstract idea into practical application or significantly more than the judicial exception. Thus, the claim is ineligible.
Regarding Claim 11, dependent upon Claim 10, recites to use a Hutchinson estimator to estimate the Frobenius norm, which is a mathematical calculation. The claim does not include additional elements that are sufficient to amount to an integration of the abstract idea into practical application or significantly more than the judicial exception. Thus, the claim is ineligible.
Regarding Claim 12, dependent upon Claim 11, recites to approximate the Hutchinson estimator using Monte-Carlo estimation, which further elaborates the mathematical calculation of the claim 11. The claim does not include additional elements that are sufficient to amount to an integration of the abstract idea into practical application or significantly more than the judicial exception. Thus, the claim is ineligible.
Regarding Claim 13, dependent upon Claim 10, recites to apply an upper bound on the spectral radius of the Jacobian matrix to estimate the Frobenius norm, which is directed to the abstract idea of a mental process (including an observation, evaluation, judgement, opinion). The claim does not include additional elements that are sufficient to amount to an integration of the abstract idea into practical application or significantly more than the judicial exception. Thus, the claim is ineligible.
Regarding Claim 14, dependent upon Claim 9, recites wherein the regularization term is weighted in the regularized loss function according to a predefined coefficient, the coefficient being configured to control a relative importance of the regularization term in the regularized loss function, which is directed to the abstract idea of a mental process (including an observation, evaluation, judgement, opinion). The claim does not include additional elements that are sufficient to amount to an integration of the abstract idea into practical application or significantly more than the judicial exception. Thus, the claim is ineligible.
Regarding Claim 15, dependent upon Claim 9, recites an additional element of to iteratively perform the operations of claim 9 for a plurality of training cycles.
The additional element of claim 15 is an insignificant extra-solution activity (performing repetitive calculations), further considered well-understood, routine and conventional (WURC) in particular field. (MPEP 2106.05(d)(II)(ii) and MPEP 2106.05(g))
Regarding Claim 16, dependent upon Claim 9, recites an additional element of to utilize the DEQ for sequence prediction, language modeling, computer vision tasks, image classification, and/or semantic segmentation, which neither integrate the abstract idea into a practical application, nor provide significantly more than the abstract idea itself. Specifically, the claims recite mere instructions to apply to an exception. MPEP 2106.05(f)
Regarding Claim 17,
Step 1: Claim 17 recites a non-transitory computer-readable medium falls into the statutory category of a manufacture.
Step 2A Prong 1: The claim recited multiple mathematical processes as explained below:
compute a regularization term using a predefined quantity of random samples and the Jacobian matrix of the DEQ at an equilibrium point z*, the regularization term penalizing the spectral radius of the Jacobian matrix to stabilize fixed-point iterations, the predefined quantity being defined for approximating a Frobenius norm of the Jacobian matrix; include the regularization term in an original loss function of the DEQ to form a regularized loss function that conditions both forward and backward fixed-point systems to regularize the Jacobian matrix using the Frobenius norm, the regularization term being weighted in the regularized loss function according to a predefined coefficient, the coefficient being configured to control a relative importance of the regularization term in the regularized loss function; compute a gradient of the regularized loss function with respect to model parameters of the DEQ; and use the gradient to update the model parameters, wherein the regularization reduces computational cost of fixed-point solving.
The claim is directed to the abstract idea of mathematical calculations involving the Jacobian matrices, gradient computation, loss functions, and optimization processes. (MPEP 2016.4(a)(2))
Further, the claim recites the regularization term being weighted in the regularized loss function according to a predefined coefficient, the coefficient being configured to control a relative importance of the regularization term in the regularized loss function, which is directed to the abstract idea of a mental process (including an observation, evaluation, judgement, opinion) which can be performed in the human mind, or by a human using pen and paper. (MPEP 2106.04(a)(2))
Step 2A Prong 2: Claim 17 does not integrate the abstract idea into a practical application, as the additional elements of A non-transitory computer-readable medium comprising instructions for regularized training of a Deep Equilibrium Model (DEQ) that, when executed by one or more computing devices, cause the one or more computing device to perform operations and during both training and inference phases of the DEQ merely indicate a field of use or technological environment in which to apply a judicial exception. (MPEP 2106.05(h))
Accordingly, the additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The claim is directed to an abstract idea.
Step 2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements do not amount to an inventive step for the reason set forth in Step 2A Prong 2. Thus, this claim is ineligible.
Regarding Claim 18, dependent upon Claim 17, recites to use a Hutchinson estimator to estimate the Frobenius norm using Monte-Carlo estimation, which further elaborates the mathematical calculation of the claim 17. The claim does not include additional elements that are sufficient to amount to an integration of the abstract idea into practical application or significantly more than the judicial exception. Thus, the claim is ineligible.
Regarding Claim 19, dependent upon Claim 17, recites to apply an upper bound on the spectral radius of the Jacobian matrix to estimate the Frobenius norm, which is directed to the abstract idea of a mental process (including an observation, evaluation, judgement, opinion). The claim does not include additional elements that are sufficient to amount to an integration of the abstract idea into practical application or significantly more than the judicial exception. Thus, the claim is ineligible.
Regarding Claim 20, dependent upon Claim 17, recites an additional element of to utilize the DEQ for sequence prediction and/or language modeling, which neither integrate the abstract idea into a practical application, nor provide significantly more than the abstract idea itself. Specifically, the claim recites mere instructions to apply to an exception. (MPEP 2106.05(f))
Response to Arguments
Applicant’s arguments, see section III, pages 8-9, filed 10/10/2025, with respect to the 35 USC 103 rejection of claims 1-20 have been fully considered and are persuasive. The rejection of claims 1-20 under 35 USC 103 has been withdrawn.
Applicant's arguments filed 10/10/2025 regarding the rejection of claims 1-20 under 35 USC 101 have been fully considered but they are not persuasive.
Specifically, applicant argues in substance:
i) That the claim, as a whole, is not directed to an abstract idea. Instead, the claim is directed to a specific improvement in computer functionality, that of stabilizing and accelerating the fixed-point training and inference pipeline of DEQs. In particular, Applicant points to the specification, goes on to state that the claims integrate any mathematical tools into a practical application that improves the operation of a computer-implemented DEQ solver, akin to recognized eligibility for improvements to computer functionality (see, e.g., Enfish; McRO). See pages 7-8 of Applicant’s Remarks.
Examiner’s response:
Examiner respectfully disagrees. The features, as described in the specification, (e.g. "a regularization scheme for DEQ models…explicitly regularizes the Jacobian…to encourage simpler and stabler equilibrium networks being learned" where " to regularize training by instead adding a new component to the loss function … leading to better conditioned fixed-point systems," and further where "[t]his regularization may add only minimal computational cost, but significantly accelerates the fixed-point-solving convergence...". cited by Applicant) are simply improved mathematical functions described only as usable in training a DEQ model. The cited sections of the specification do not describe an improved computer function or another technology such as the process of learning or training of a DEQ model. Unlike the claims in Enfish and McRO, the improvement described in the specification as cited and asserted by applicant is only an improvement in an abstract idea.
ii) Applicant further notes that the amended claim recites concrete operations that change how the computer executes a DEQ. These are not field-of-use or result-only statements; they reconfigure the training inference process so the solver requires fewer iterations and fewer operations when run on general-purpose hardware. Thus, the claims recite a practical application of improved training inference.
In addition, Applicant asserts that the ordered combination, including penalizing the spectral radius of the Jacobian at z* to condition both fixed-point systems, is not well-understood, routine, or conventional in DEQ training. See page 7 of Applicant’s Remarks.
Examiner’s response:
Examiner respectfully disagrees. Contrary to Applicant’s assertion, the claim limitations as asserted merely amount to mathematical concepts, an abstract idea rather than reciting or reflecting an improved technology such as DEQ training process or improved computer function such as memory access in the Enfish case. Specifically, each of the claim limitations "computing a Jacobian-based regularization term at the equilibrium point z* using random samples, "to condition both forward and backward fixed-point systems," and (whereby the method) "reduces computational cost of fixed-point solving during both training and inference" only represents a mathematical concept or mental process, i.e. an abstract idea.
MPEP 2106.05(a) recites in part:
After the examiner has consulted the specification and determined that the disclosed invention improves technology, the claim must be evaluated to ensure the claim itself reflects the disclosed improvement in technology. Intellectual Ventures I LLC v. Symantec Corp., 838 F.3d 1307, 1316, 120 USPQ2d 1353, 1359 (Fed. Cir. 2016) (patent owner argued that the claimed email filtering system improved technology by shrinking the protection gap and mooting the volume problem, but the court disagreed because the claims themselves did not have any limitations that addressed these issues). That is, the claim must include the components or steps of the invention that provide the improvement described in the specification
…
It is important to note, the judicial exception alone cannot provide the improvement. The improvement can be provided by one or more additional elements. See the discussion of Diamond v. Diehr, 450 U.S. 175, 187 and 191-92, 209 USPQ 1, 10 (1981)) in subsection II, below. In addition, the improvement can be provided by the additional element(s) in combination with the recited judicial exception.
In this case, as noted in the rejection, the additional elements of generic computer elements such as processors and technologies such as training/learning of a DEQ model are merely recited as field of use/intended result limitations. The claims do not recite nor reflect an improvement in computer functions or the DEQ training, as required for integration of the judicial exception into a practical application. The rejection clearly sets forth the judicial exception (mathematical concepts) and provides a proper analysis of the additional elements, i.e. the training limitations, thus explaining why the claims are ineligible. Applicant has not pointed out why the mathematical concepts or mental processes cannot be considered to be abstract ideas or how the claims recite or reflect the purported improvement. Therefore, the rejection is proper and is maintained.
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 Kakali Chaki whose telephone number is (571)272-3719. The examiner can normally be reached on Mondays through Fridays from 10 am to 5 pm.
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, David Wiley can be reached on 571-272-4150. 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.
/KAKALI CHAKI/ Supervisory Patent Examiner, Art Unit 2122