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 Amendment
This Office Action is in response to applicant’s communication filed 31 March 2026, in response to the Office Action mailed 18 March 2026. The applicant’s remarks and any amendments to the claims or specification have been considered, with the results that follow.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 31 March 2026 has been entered.
Information Disclosure Statement
As required by M.P.E.P. 609(c), the applicant's submission of the Information Disclosure Statement, dated 27 April 2026, is acknowledged by the examiner and the cited references have been considered in the examination of the claims now pending. As required by M.P.E.P 609 C(2), a copy of the PTOL-1449 initialed and dated by the examiner is attached to the instant office action.
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.
Claim(s) 1-20 is/are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The claim(s) recite(s) mathematical concepts and/or mental processes. This judicial exception is not integrated into a practical application and does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception, as described below.
Step 1 for all claims:
Under the first part of the analysis, claims 1-7 recite a method, claims 8-14 recite a manufacture, and claims 15-20 recite a device. Accordingly, these claims fall within the four statutory categories of invention and the analysis proceeds to Step 2A, prongs 1 and 2, and Step 2B, as described below.
As per claim 1:
Under step 2A, prong 1, the claim recites an abstract idea including the following mathematical concept elements:
obtaining… a predictive distribution pθ(y| x̄) of an outcome y by using the incomplete set of covariates x̄ and a parameter θ, the parameter θ being unknown – this is describing a mathematical formula/calculation and its parameters.
wherein a learning of the parameter θ includes performing a maximization by maximizing a stochastically approximated conditional evidence lower bound (CELBO) – maximizing a value is a mathematical calculation/operation (see, e.g., claim 6 for the formula of the stochastically approximated conditional evidence lower bound being maximized).
If a claim, under the broadest reasonable interpretation covers a mathematical relationship between variables or numbers, a numerical formula or equation, or a mathematical calculation, it will be considered as falling within the “mathematical concepts” grouping of abstract ideas. If a claim, under the broadest reasonable interpretation covers concepts that can be performed in the human mind, or by a human using a pen and paper, including observation, evaluation, judgment, or opinion, it will be considered as falling within the “mental processes” grouping of abstract ideas. Additionally, performing mathematical calculations using a formula that could be practically performed in the human mind may be considered to fall within both the mathematical concepts grouping and the mental process grouping. See MPEP § 2106.04(a)(2).
Accordingly, at step 2A, prong one, the claim is directed to an abstract idea.
Under step 2A, prong two, the judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements of:
A computer-implemented method for learning with incomplete data in which some of entries are missing, comprising: – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
acquiring an incomplete set of covariates x̄ including incomplete features x̃ and an incomplete pattern m indicating missing entries of the incomplete set of covariates x̃ – this is recited at a high level of generality and amounts to insignificant extra-solution activity as data gathering/storage that is limited to a particular type of data, generally linking the judicial exception to a particular technological environment or field of use. See MPEP § 2106.05(g) and (h), and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
and obtaining, by a hardware processor – this amounts to mere instructions to apply the exception using a generic computer component, recited at a high level of generality. See MPEP § 2106.05(f).
having a mask vector – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
and where the stochastically approximated conditional evidence lower bound includes a surrogate parameterized density ratio between densities configured to keep a gradient of the stochastically approximated conditional evidence lower bound below a threshold during the maximization – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
Accordingly, at step 2A, prong two, these additional elements do not integrate the abstract idea into a practical application for the claim as a whole, because it does not impose any meaningful limits on practicing the abstract idea. See MPEP § 2106.04(d).
Under step 2B, the claims do not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the claim recites the additional elements of:
A computer-implemented method for learning with incomplete data in which some of entries are missing, comprising: – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
acquiring an incomplete set of covariates x̄ including incomplete features x̃ and an incomplete pattern m indicating missing entries of the incomplete set of covariates x̃ – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data). The courts have also found limitations directed to obtaining and storing information electronically, recited at a high level of generality, to be well-understood, routine, and conventional. See MPEP § 2106.05(d)(II) “receiving or transmitting data over a network,” "electronic record keeping,” and "storing and retrieving information in memory.”
and obtaining, by a hardware processor – this amounts to mere instructions to apply the exception using a generic computer component, recited at a high level of generality. See MPEP § 2106.05(f).
having a mask vector – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
and where the stochastically approximated conditional evidence lower bound includes a surrogate parameterized density ratio between densities configured to keep a gradient of the stochastically approximated conditional evidence lower bound below a threshold during the maximization – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
Accordingly, at step 2B, these additional elements, both individually and in combination, do not amount to significantly more than the judicial exception. See MPEP § 2106.05.
Therefore, the claim is not eligible subject matter under 35 U.S.C. 101.
As per claim 2:
Under step 2A, prong two, the judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements of:
wherein the incomplete set of covariates x̄ represent patient measurements taken from hardware based patient-interactive medical devices – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
Accordingly, at step 2A, prong two, these additional elements do not integrate the abstract idea into a practical application for the claim as a whole, because it does not impose any meaningful limits on practicing the abstract idea. See MPEP § 2106.04(d).
Under step 2B, the claims do not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the claim recites the additional elements of:
wherein the incomplete set of covariates x̄ represent patient measurements taken from hardware based patient-interactive medical devices – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
Accordingly, at step 2B, these additional elements, both individually and in combination, do not amount to significantly more than the judicial exception. See MPEP § 2106.05.
Therefore, the claim is not eligible subject matter under 35 U.S.C. 101.
As per claim 3:
The claim recites the following additional mental process elements:
further comprising limiting a number of covariates per patient in the incomplete set of covariates x̄ – a data scientist can determine the number of covariates per patient to be used (limiting the number of covariates per patient).
Accordingly, at step 2A, prong one, the claim is directed to an abstract idea.
The claim does not include any additional elements, under step 2A prong two, or step 2B, except those listed above in prior claim(s). Accordingly, at step 2A, prong two, the claim as a whole does not integrate the judicial exception into a practical application. See MPEP § 2106.04(d). Furthermore, at step 2B, the claim elements both individually and in combination do not amount to significantly more than the judicial exception. See MPEP § 2106.05.
Therefore, the claim is not eligible subject matter under 35 U.S.C. 101.
As per claim 4:
Under step 2A, prong two, the judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements of:
wherein the outcome y is a prediction time of an adverse medical event requiring medical intervention – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
Accordingly, at step 2A, prong two, these additional elements do not integrate the abstract idea into a practical application for the claim as a whole, because it does not impose any meaningful limits on practicing the abstract idea. See MPEP § 2106.04(d).
Under step 2B, the claims do not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the claim recites the additional elements of:
wherein the outcome y is a prediction time of an adverse medical event requiring medical intervention – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
Accordingly, at step 2B, these additional elements, both individually and in combination, do not amount to significantly more than the judicial exception. See MPEP § 2106.05.
Therefore, the claim is not eligible subject matter under 35 U.S.C. 101.
As per claim 5:
The claim recites the following additional mathematical concept elements:
wherein a computation of the predictive distribution pθ(y| x̄) is performed by maximizing an objective function ℒ(θ) := ln pθ(y| x̄) = - ln pθ(x̃ | m) + ln pθ(y, x̃ | m), and the objective function ℒ(θ) is bounded with a difference between an evidence upper bound ℒEUBO and an evidence lower bound ℒELBO, where ln pθ(x̃ | m) ≤ ℒEUBO, ln pθ(x̃ | m) ≥ ℒELBO , and wherein m is a mask vector indicating missing entries of x̃ – this is describing a mathematical operation using a recited mathematical formula, as well as describing the elements of the formula used.
Accordingly, at step 2A, prong one, the claim is directed to an abstract idea.
The claim does not include any additional elements, under step 2A prong two, or step 2B, except those listed above in prior claim(s). Accordingly, at step 2A, prong two, the claim as a whole does not integrate the judicial exception into a practical application. See MPEP § 2106.04(d). Furthermore, at step 2B, the claim elements both individually and in combination do not amount to significantly more than the judicial exception. See MPEP § 2106.05.
Therefore, the claim is not eligible subject matter under 35 U.S.C. 101.
As per claim 6:
The claim recites the following additional mathematical concept elements:
wherein the stochastically approximated conditional evidence lower bound is stochastically approximated with
PNG
media_image1.png
46
542
media_image1.png
Greyscale
wherein z is a latent variable of a variational autoencoder, qϕ(z|y, x̄) = q(z|ϕ(y, x̄)) is a conditional density function defined by a neural network ϕ to be trained together with the parameter θ, qψ(z|x̄) = q(z|ψ(x̄)) is a conditional density function defined by a neural network ψ to be trained together with the parameter θ, ξ is a surrogate network, α is a fixed real number greater than 1, m is a mask vector indicating missing entries of x̃, and zϕ and zψ are random variables drawn from q(z|y, x̃, m, ϕ) and q(z|x̃, m, ψ), respectively – this is describing a mathematical operation using a recited mathematical formula, as well as describing the elements of the formula used.
Accordingly, at step 2A, prong one, the claim is directed to an abstract idea.
The claim does not include any additional elements, under step 2A prong two, or step 2B, except those listed above in prior claim(s). Accordingly, at step 2A, prong two, the claim as a whole does not integrate the judicial exception into a practical application. See MPEP § 2106.04(d). Furthermore, at step 2B, the claim elements both individually and in combination do not amount to significantly more than the judicial exception. See MPEP § 2106.05.
Therefore, the claim is not eligible subject matter under 35 U.S.C. 101.
As per claim 7:
The claim recites the following additional mathematical concept elements:
wherein for the portion of the second term of
PNG
media_image2.png
57
287
media_image2.png
Greyscale
a density ratio
PNG
media_image3.png
36
235
media_image3.png
Greyscale
variables (θ, ϕ, ψ, ξ) are changed to (θ’, ϕ’, ψ’, ξ’) so that a resultant density ratio wθ’,ψ’,ξ’(x̃, z|m) stabilizes the maximization of the stochastically approximated CELBO – this is describing a mathematical operation using a recited mathematical formula, as well as describing the elements of the formula used.
Accordingly, at step 2A, prong one, the claim is directed to an abstract idea.
The claim does not include any additional elements, under step 2A prong two, or step 2B, except those listed above in prior claim(s). Accordingly, at step 2A, prong two, the claim as a whole does not integrate the judicial exception into a practical application. See MPEP § 2106.04(d). Furthermore, at step 2B, the claim elements both individually and in combination do not amount to significantly more than the judicial exception. See MPEP § 2106.05.
Therefore, the claim is not eligible subject matter under 35 U.S.C. 101.
As per claim 8:
See the rejection of claim 1, above, wherein under step 2A, prong two, the judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements of:
A computer program product for learning with incomplete data in which some the entries are missing – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to [perform the method] – this amounts to mere instructions to apply the exception using a generic computer component, recited at a high level of generality. See MPEP § 2106.05(f).
Accordingly, at step 2A, prong two, these additional elements do not integrate the abstract idea into a practical application for the claim as a whole, because it does not impose any meaningful limits on practicing the abstract idea. See MPEP § 2106.04(d).
Under step 2B, the claims do not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the claim recites the additional elements of:
A computer program product for learning with incomplete data in which some the entries are missing – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to [perform the method] – this amounts to mere instructions to apply the exception using a generic computer component, recited at a high level of generality. See MPEP § 2106.05(f).
Accordingly, at step 2B, these additional elements, both individually and in combination, do not amount to significantly more than the judicial exception. See MPEP § 2106.05.
Therefore, the claim is not eligible subject matter under 35 U.S.C. 101.
As per claim 9, see the rejection of claim 2, above.
As per claim 10, see the rejection of claim 3, above.
As per claim 11, see the rejection of claim 4, above.
As per claim 12, see the rejection of claim 5, above.
As per claim 13, see the rejection of claim 6, above.
As per claim 14, see the rejection of claim 7, above.
As per claim 15:
See the rejection of claim 1, above, wherein under step 2A, prong two, the judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements of:
A computer processing system for learning with incomplete data in which some entries are missing, comprising: – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
a memory device for storing program code – this amounts to mere instructions to apply the exception using a generic computer component, recited at a high level of generality. See MPEP § 2106.05(f).
and a hardware processor operatively coupled to the memory device for running the program code to: [perform the method] – this amounts to mere instructions to apply the exception using a generic computer component, recited at a high level of generality. See MPEP § 2106.05(f).
Accordingly, at step 2A, prong two, these additional elements do not integrate the abstract idea into a practical application for the claim as a whole, because it does not impose any meaningful limits on practicing the abstract idea. See MPEP § 2106.04(d).
Under step 2B, the claims do not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the claim recites the additional elements of:
A computer processing system for learning with incomplete data in which some entries are missing, comprising: – this amounts to generally linking the use of the judicial exception to a particular technological environment or field of use by limiting it to a particular data source or type. See MPEP § 2106.05(h) and Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data).
a memory device for storing program code – this amounts to mere instructions to apply the exception using a generic computer component, recited at a high level of generality. See MPEP § 2106.05(f).
and a hardware processor operatively coupled to the memory device for running the program code to: [perform the method] – this amounts to mere instructions to apply the exception using a generic computer component, recited at a high level of generality. See MPEP § 2106.05(f).
Accordingly, at step 2B, these additional elements, both individually and in combination, do not amount to significantly more than the judicial exception. See MPEP § 2106.05.
Therefore, the claim is not eligible subject matter under 35 U.S.C. 101.
As per claim 16, see the rejection of claim 2, above.
As per claim 17, see the rejection of claim 3, above.
As per claim 18, see the rejection of claim 4, above.
As per claim 19, see the rejection of claim 5, above.
As per claim 20, see the rejection of claim 6, above.
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 (i.e., changing from AIA to pre-AIA ) 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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 1-3, 5, 7-10, 12, 14-17, and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Jebara (Discriminative, Generative and Imitative Learning, Feb 2002, pgs. 1-212 – cited in an IDS), in view of Collier et al. (VAEs in the Presence of Missing Data, March 2021, pgs. 1-9), and further in view of Huszár (Variational Inference using Implicit Distributions, Feb 2017, pgs. 1-11).
As per claim 1, Jebara teaches a computer-implemented method for learning with incomplete data in which some of entries are missing, comprising: acquiring an incomplete set of covariates x̄ including incomplete features x̃ and an incomplete pattern m indicating missing entries of the incomplete set of covariates x̃ [the system uses multiple models to infer missing/incomplete input data for classification, etc. (pg. 95, section 5.4; see also: pg. 29, section 2.1.2; pg. 30, section 2.2.1; pg. 48, section 3.4; pg. 58, section 3.9.2; pg. 79, section 4.3; pg. 81, section 4.3.1; pg. 104, section 5.7 and fig. 5.6; pg. 191, section 9.2; etc.) which can include missing/incomplete features and labels (covariates x̄ and features x̃) (pg. 69, section 4; see also: pg. 39, section 2.5; pg. 65, section 3.10.1; etc.)]; and obtaining, by a hardware processor [the system can be implemented in a wearable computer system (pg. 168, section 8.2; etc.); which requires a hardware processor obtaining data and processing instructions/operations], a predictive distribution pθ(y| x̄) of an outcome y by using the incomplete set of covariates x̄ and a parameter θ, the parameter θ being unknown [mixtures of exponential family (e-family) models are used to produce a prediction distribution p(X|θ) (pg. 89, section 4.1; pg. 112, section 6.1; etc.) and the mixture is done by introducing a latent variable which is denoted as m here. Thus, we have an incomplete data representation and since m is unobserved, multiple e-family models are mixed (pg. 90, section 5.1.1; pg. 112, section 6.1; etc.); where the parameter is unobserved/unknown], wherein a learning of the parameter θ includes performing a maximization by maximizing a stochastically approximated conditional evidence lower bound [The CEM (Conditional Expectation Maximization) algorithm mirrors the EM algorithm in its approach to maximizing joint likelihood. EM iterates by lower bounding and then maximizing the joint loglikelihood. CEM iterates by lower bounding and then maximizing the conditional log-likelihood. Due to the guarantees behind both the Jensen and reverse-Jensen bounds, CEM converges monotonically to a local maximum of conditional likelihood (pg. 103, section 5.7; see also: pgs. 88-89, section 5; pgs. 93-94, section 5.3; pgs. 97-99, section 5.5.1; pg. 100, fig. 5.5; pgs. 114-115, section 6.2.1; pg. 118, fig. 6.4; pg. 127, section 6.5; etc.); where the conditional likelihood is the stochastically approximated conditional evidence lower bound, maximized using the Jensen (lower) and reverse-Jensen (upper) bounds], and where the stochastically approximated conditional evidence lower bound includes a density ratio to keep a gradient of the stochastically approximated conditional evidence lower bound below a threshold during the maximization [the models are trained using a probability density over each datum (density ratio) to constrain the gradient of the upper and lower bounds (pg. 127, section 6.5; etc.)].
While Jebara teaches maximizing the objective function and the conditional likelihood (conditional evidence lower bound) using log likelihood functions (see above), it has not been relied upon for teaching the stochastically approximated CELBO having a mask vector, and where the stochastically approximated CELBO includes a surrogate parameterized density ratio between densities configured to keep a gradient of the stochastically approximated CELBO below a threshold during the maximization.
Collier teaches the stochastically approximated CELBO having a mask vector [a binary mask vector m(i) is used to indicate observed vs. missing data values (pg. 2, section 3; etc.) and the generative model is conditioned on the mask (pg. 3, section 3, etc.); for the missing entries, and conditioning the functions in Jebara, above].
Jebara and Collier are analogous art, as they are within the same field of endeavor, namely predictive modelling, including using latent and generative models, on data sets including incomplete/missing data.
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to use a mask vector indicating the missing data entries, as taught by Collier, for the missing data entries in the acquired data in the system/method taught by Jebara.
Collier provides motivation as [providing the missing mask, m, to the models can improve performance and disambiguates the input data (pg. 4, section 4; etc.)].
Huszár teaches where the stochastically approximated CELBO includes a surrogate parameterized density ratio between densities configured to keep a gradient of the stochastically approximated CELBO below a threshold during the maximization [re-parameterization can be used to construct a low variance estimator to ELBO by formulating it in terms of density ratios (pgs. 3-4, section 2; pgs. 4-5, section 3.1; etc.), which constrains the gradients (pg. 4, sections 2.1-2.2; etc.)].
Jebara and Huszár are analogous art, as they are within the same field of endeavor, namely utilizing variational inference and generative models.
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to utilize re-parameterization to construct a low variance estimator of the ELBO, formulated in terms of density ratios, as taught by Huszár, to keep the gradient of the stochastically approximated CELBO below a threshold during the maximization of the system taught by Jebara/Collier.
Huszár provides motivation as [formulating the ELBO in terms of density ratios can be useful when ELBO must be approximated (pg. 3, section 2; etc.) and is relevant in many machine learning applications, such as dealing with covariate shift or domain adaptation (pg. 4, section 3; etc.)].
As per claim 2, Jebara/Collier/Huszár teaches wherein the incomplete set of covariates x̄ represent patient measurements taken from hardware based patient-interactive medical devices [the input data features (covariates) can represent numerical attributes from patients’ tumors (Jebara: pg. 62, section 3.9.3; etc.); where numerical attributes of patients’ tumors represent patient measurements taken from hardware based patient-interactive medical devices].
As per claim 3, Jebara/Collier/Huszár teaches further comprising limiting a number of covariates per patient in the incomplete set of covariates x̄ [one data set used includes classification from 9 numerical attributes from patients’ tumors (Jebara: pg. 62, section 3.9.2; etc.); and training sets may be limited to set numbers of samples/features (covariates) (Jebara: pg. 78, section 4.2.2; etc.) or set ranges of variation/values/domain (Jebara: pg. 135, section 7.1; etc.); where 9 is a limit of the number of covariates per patient].
As per claim 5, Jebara/Collier/Huszár teaches wherein a computation of the predictive distribution pθ(y| x̄) is performed by maximizing an objective function ℒ(θ) := ln pθ(y| x̄) = - ln pθ(x̃ | m) + ln pθ(y, x̃ | m), and the objective function ℒ(θ) is bounded with a difference between an evidence upper bound ℒEUBO and an evidence lower bound ℒELBO, where ln pθ(x̃ | m) ≤ ℒEUBO, ln pθ(x̃ | m) ≥ ℒELBO [mixtures of exponential family (e-family) models are used to produce a prediction distribution p(X|θ) (Jebara: pg. 89, section 4.1; pg. 112, section 6.1; etc.); by maximizing an objective function, bounded by the Jensen (lower) and reverse-Jensen (upper) bounds (Jebara: pgs. 124-125, section 6.4; where the objective functions there show the conditional log likelihoods (using natural log/ln functions) used and bounded by the upper and lower bounds), and the conditioning on m is addressed by Collier, below], and wherein m is a mask vector indicating missing entries of x̃ [a binary mask vector m(i) is used to indicate observed vs. missing data values (Collier: pg. 2, section 3; etc.) and the generative model is conditioned on the mask (Collier: pg. 3, section 3, etc.); for the missing entries, and conditioning the functions in Jebara, above].
As per claim 7, Jebara/Collier/Huszár teaches wherein for the portion of the second term of
PNG
media_image2.png
57
287
media_image2.png
Greyscale
the density ratio
PNG
media_image3.png
36
235
media_image3.png
Greyscale
variables (θ, ϕ, ψ, ξ) are changed to (θ’, ϕ’, ψ’, ξ’) so that the resultant density ratio wθ’,ψ’,ξ’(x̃, z|m) stabilizes the maximization of the stochastically approximated CELBO [the models are trained using a probability density over each datum (density ratio) to constrain the gradient of the upper and lower bounds (Jebara: pg. 127, section 6.5; etc.); which illustrates he density ratio, and training changes the parameters to the new set of ‘ parameters].
As per claim 8, see the rejection of claim 1, above, wherein Jebara/Collier/Huszár also teaches a computer program product for learning with incomplete data in which some of the entries are missing, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to [perform the method] [the system can be implemented in a wearable computer system (pg. 168, section 8.2; etc.); which requires a processor executing stored instructions].
As per claim 9, see the rejection of claim 2, above.
As per claim 10, see the rejection of claim 3, above.
As per claim 12, see the rejection of claim 5, above.
As per claim 14, see the rejection of claim 7, above.
As per claim 15, see the rejection of claim 1, above, wherein Jebara/Collier/Huszár also teaches a computer processing system for learning with incomplete data in which some of entries are missing, comprising: a memory device for storing program code; and a hardware processor operatively coupled to the memory device for running the program code to: [perform the method] [the system can be implemented in a wearable computer system (Jebara: pg. 168, section 8.2; etc.); which requires a processor executing stored instructions].
As per claim 16, see the rejection of claim 2, above.
As per claim 17, see the rejection of claim 3, above.
As per claim 19, see the rejection of claim 5, above.
Claim(s) 4, 11, and 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Jebara (Discriminative, Generative and Imitative Learning, Feb 2002, pgs. 1-212 – cited in an IDS), in view of Collier et al. (VAEs in the Presence of Missing Data, March 2021, pgs. 1-9), further in view of Huszár (Variational Inference using Implicit Distributions, Feb 2017, pgs. 1-11), and further in view of Saria (US 2020/0005941).
As per claim 4, Jebara/Collier/Huszár teaches the computer-implemented method of claim 1, as described above.
While Jebara/Collier/Huszár teaches that the prediction is medical information for a patient (see above), it has not been relied upon for teaching wherein the outcome y is a prediction time of an adverse medical event requiring medical intervention.
Saria teaches wherein the outcome y is a prediction time of an adverse medical event requiring medical intervention [the models are used for predicting adverse medical events from irregularly sampled multivariate time series data (para. 0004, etc.)].
Jebara/Collier/Huszár and Saria are analogous art, as they are within the same field of endeavor, namely making medical predictions for patients from incomplete data.
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to predict adverse medical events (times) from the incomplete data, as taught by Saria, in the system predicting medical classifications from incomplete data taught by Jebara/Collier/Huszár.
Because both Jebara/Collier/Huszár and Saria teach systems using multiple models to predict medical outcomes based upon incomplete/missing data from patients, it would have been obvious to one of ordinary skill in the art to predict adverse medical events (times) from the incomplete data, as taught by Saria, in the system predicting medical classifications from incomplete data taught by Jebara, to achieve the predictable result of providing more detail for any necessary/useful medical interventions (see, e.g., Saria: paras. 0005-6, etc.).
As per claim 11, see the rejection of claim 4, above.
As per claim 18, see the rejection of claim 4, above.
Allowable Subject Matter
Examiner’s Note: regarding claims 6, 13, and 20, while the cited art teaches various systems/methods for approximating the conditional evidence lower bound (see above), none of the cited art appear to teach, either alone or in combination, the claimed functions used in the manner claimed.
Response to Arguments
The rejections under 35 U.S.C. 112 have been withdrawn due to the amendments filed.
Applicant's arguments filed 31 March 2026, regarding the rejections under 35 U.S.C. 101, have been fully considered but they are not persuasive.
Applicant argues that the specification provides a technical explanation that sufficiently describes to a person of ordinary skill in the art that the claims provide for an improvement by “preventing generative models from being intractable.” Applicant argues that this improvement is provided by “using a generative model using a stochastically approximated conditional evidence lower bound to predict missing values,” “[including] a mask vector m to indicate the missing entries,” “using surrogate parameterization to prevent the generative model from being intractable,” “surrogate parameterization of a density ratio can be used to stabilize the generative model when the generative model is maximizing the conditional evidence lower bound and prevents the generative model from becoming intractable,” and in other words, “generative models struggle with missing values and tend to go haywire.”
However, applicant appears to be describing an improvement to the identified mathematical concepts. Therefore, (assuming that the invention provides these advantages) this amounts to an improvement to an abstract idea rather than to a computer or technology. See MPEP 2106.05(a). It appears that any benefits to the computer itself are based solely on the use of an improvement to the abstract idea(s), using generic computer components to apply the abstract idea(s). Additionally, to find a valid improvement to a computer or technology the specification must disclose the improvement and the claim must include the necessary components to realize the improvement. MPEP 2106.05(d)(1).
Applicant further argues that the surrogate parameterized density ratio and the mask vectors are considered additional elements and “thus, the improvements cannot be directed towards mathematical concepts as described in the Office Action.”
However, selecting improved parameters/data types for the mathematical functions is still an improvement to the mathematical concepts. Therefore, this amounts to an improvement to an abstract idea rather than to a computer or technology. See MPEP 2106.05(a).
Applicant’s arguments, see the remarks, filed 31 March 2026, with respect to the rejection(s) of claim(s) 1-3, 5, 7-10, 12, 14-17, and 19 under 35 U.S.C. 103 have been fully considered and are persuasive in view of the amendments made to the independent claim. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection is made in view of Huszár, which has been relied upon for teaching where the stochastically approximated CELBO includes a surrogate parameterized density ratio between densities configured to keep a gradient of the stochastically approximated CELBO below a threshold during the maximization, in combination with Jebara and Collier (see above).
Conclusion
The following is a summary of the treatment and status of all claims in the application as recommended by M.P.E.P. 707.07(i): claims 1-20 are rejected.
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Zhang et al. (Advances in Variational Inference, Aug 2019, pgs. 2008-2026) – discusses various systems/methods utilizing variational inference.
Itkina et al. (Evidential Sparsification of Multimodal Latent Spaces in Conditional Variational Autoencoders, Jan 2021, pgs. 1-21) – discloses sparsifying a latent space of a conditional VAE, including maximizing a standard conditional evidence lower bound during training of the model.
Bakshi (US 2021/0295963) – discloses a system/method for providing real time data of a patient, including predicting upcoming adverse medical events.
Menon et al. (Linking losses for density ratio and class-probability estimation, 2016, pgs. 1-10) – discloses density ration estimation (DRE) and class-probability estimation (CPE) using existing losses from one to the other.
Azoury et al. (Relative Loss Bounds for On-line Density Estimation with the Exponential Family of Distributions, 2013, pgs. 31-40) – discloses online density estimation with a parameterized density from the exponential family, which chooses the parameter, based on the training examples, to bound the total loss of the online algorithm over the total loss of the offline algorithm.
Poole et al. (On Variational Bounds of Mutual Information, May 2019, pgs. 1-14) – discloses training a log density ratio estimator to maximize a lower bound on the Jensen-Shannon (JS) divergence, and use the density ratio estimate.
Shi et al. (Implicit Variational Inference with Kernel Density Ratio Fitting, 2017, pgs. 1-9) – discloses an implicit variational inference approach with kernel density ratio fitting.
The examiner requests, in response to this Office action, that support be shown for language added to any original claims on amendment and any new claims. That is, indicate support for newly added claim language by specifically pointing to page(s) and line number(s) in the specification and/or drawing figure(s). This will assist the examiner in prosecuting the application.
When responding to this office action, Applicant is advised to clearly point out the patentable novelty which he or she thinks the claims present, in view of the state of the art disclosed by the references cited or the objections made. He or she must also show how the amendments avoid such references or objections. See 37 CFR 1.111(c).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to GEORGE GIROUX whose telephone number is (571)272-9769. The examiner can normally be reached M-F 10am-6pm.
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, Omar Fernandez Rivas can be reached at 571-272-2589. 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.
/GEORGE GIROUX/Primary Examiner, Art Unit 2128
Prosecution Insights