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 .
Specification
The lengthy specification has not been checked to the extent necessary to determine the presence of all possible minor errors. Applicant’s cooperation is requested in correcting any errors of which applicant may become aware in the specification.
Drawings
The applicant’s submitted drawings appear to be acceptable for examination purposes. Applicant’s cooperation is requested in correcting any errors of which applicant may become aware in the drawings.
Information Disclosure Statement
As required by M.P.E.P. 609(c), the applicant's submission of the Information Disclosure Statement, dated 28 December 2023, 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 Objections
Claim 5 is objected to because of the following informalities: “wherein the Bayesian regularization utilize external” appears as though it should be “wherein the Bayesian regularization utilizes external” or similar. Appropriate correction is required.
Claim 17 is objected to because of the following informalities: “data in training dataset” appears as though it should be “data in a training dataset,” or “data in training datasets,” or similar. Appropriate correction is required.
Claims 18-19 depend upon claim 17, and thus include the aforementioned limitation(s).
Claim 20 is objected to because of the following informalities: “data in training dataset” appears as though it should be “data in a training dataset,” or “data in training datasets,” or similar. Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-20 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1 recites the limitation "the at least three unknown quantities” in line 6 (and further recitations of the same). There is insufficient antecedent basis for this limitation in the claim. Additionally, the intended scope of the claim is not clear because the relationship between the at least three unknown quantities and the MNAR data/outcomes is not clear. For the purposes of examination, the examiner has assumed that the unknown quantities are MNAR data.
Claims 2-16 depend upon claim 1, and thus include the aforementioned limitation(s).
Claim 5 also recites the limitation "the function” in line 2. There is insufficient antecedent basis for this limitation in the claim, as the only prior recitation (in claim 1) is to “a modified function” and thus it is not clear whether the function being referred to is the modified function or some relation to the modified function. For the purposes of examination, the examiner has assumed the former.
Claim 8 also recites “without using an assumption or a restriction about the distribution of missing values.” The intended scope of the claim is not clear because it is not clear what is meant by “an assumption or a restriction” in this context. The examiner has assumed, for the purposes of examination, that this intends to mean that no predefined constraint as to a distribution of missing values is used.
Claim 9 also recites the limitation “one of the estimation” in line 1. There is insufficient antecedent basis for this limitation in the claim, as the prior recitation only appears to include one estimation.
Claim 10 also recites the limitation “other of the estimation” in line 1. There is insufficient antecedent basis for this limitation in the claim, as the prior recitation only appears to include one estimation. Claim 10 also recites the limitation “the features in both the labeled data and unlabeled data” in line 2. There is insufficient antecedent basis for this limitation in the claim, as the prior recitation is only to features in labeled data.
Claim 17 recites the limitation "the at least three unknown quantities” in line 8 (and further recitations of the same). There is insufficient antecedent basis for this limitation in the claim. Additionally, the intended scope of the claim is not clear because the relationship between the at least three unknown quantities and the MNAR data/outcomes is not clear. For the purposes of examination, the examiner has assumed that the unknown quantities are MNAR data.
Claims 18-19 depend upon claim 17, and thus include the aforementioned limitation(s).
Claim 20 recites the limitation "the at least three unknown quantities” in line 7 (and further recitations of the same). There is insufficient antecedent basis for this limitation in the claim. Additionally, the intended scope of the claim is not clear because the relationship between the at least three unknown quantities and the MNAR data/outcomes is not clear. For the purposes of examination, the examiner has assumed that the unknown quantities are MNAR data.
Claim Rejections - 35 USC § 102
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.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claim(s) 1-5, 7-10, 14, 17, 18, and 20 is/are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Li et al. (Be Causal: De-biasing Social Network Confounding in Recommendation, May 2021, pgs. 1-11).
As per claim 1, Li teaches a method for accounting for missing not-at-random (MNAR) data in training datasets by performing Bayesian regularization [a method for accounting for MNAR data in recommendation training data for recommender systems (pg. 1, abstract; pg. 2, “The proposed approach” and fig. 2; etc.) using multiple regularization methods (pg. 5, sections 3.4-3.5; etc.)], the method comprising:
acquiring, by a processor and from at least one database, historical data of an organization, the historical data including the MNAR data, the MNAR data being data where outcomes are MNAR [the method receives past (historical) rating data for a number of users and items (pg. 3, section 3.1; pg. 6, section 4.1.2; etc.) that includes the MNAR ratings (outcomes) (pg. 4, section 3.3-3.3.2; etc.)];
performing, by the processor, estimation for two of the at least three unknown quantities based on the historical data [the deconfounder model learns two latent vectors to represent inherent factors of users and items, which are used to estimate ratings, including the MNAR ratings (pg. 4, fig. 4; pg. 5, sections 3.4-3.5; etc.); where the figure shows at least two estimations];
injecting, by the processor, quantitative information for remaining one of the at least three unknown quantities based on qualitative information regarding nature of missingness [the confounder and exposure models are used to quantify common bias factors (qualitative information) affecting the exposure and rating outcomes (pg. 4, fig. 4 and section 3.3; etc.); where the qualitative bias information is regarding the nature of the missingness (based on user exposure, as described in the same section)]; and
reassembling, by the processor, the estimation for two of the at least three unknown quantities and injected quantitative information for the remaining one of the at least three unknown quantities, to provide a modified function [the obtained inherent factors from the deconfounder model and the learned exposure information are combined to create the estimated combined ratings through a combined linear model (modified function) (pg. 4, fig. 4; pg. 5, section 3.5; etc.)].
As per claim 2, Li teaches wherein the estimation includes training machine learning models for the two unknown quantities [the social network confounder, exposure model, and deconfounder model used to estimate the rating quantities include training multiple machine learning models (pg. 4, fig. 4; pg. 5, section 3.5; etc.)].
As per claim 3, Li teaches wherein a Bayes risk minimizer is utilized in the performing of the estimation [optimization of the model(s) includes minimization of an error/loss function (pg. 5, sections 3.3-3.5; etc.); which is a Bayes risk minimizer].
As per claim 4, Li teaches wherein the Bayes risk minimizer is a mean squared error (MSE) [optimization of the model(s) includes minimization of an error/loss function including a squared l2 normal regularization term (pg. 5, sections 3.3-3.5; etc.)].
As per claim 5, Li teaches wherein the Bayesian regularization utilize external information about the function when a missingness indicator is 0 to construct the modified function [the aim is to estimate the ratings had all movies been seen by all users, i.e., estimate rating outcomes of 1, from potential rating outcome 0 (missingness indicator 0), utilizing exposure and inherent factors (external information) (pg. 3, section 3.2; etc.)].
As per claim 7, Li teaches wherein a value within a reference vicinity of one of the two unknown quantities being estimated when a missingness indicator is 0 is utilized to construct the modified function [the aim is to estimate the ratings had all movies been seen by all users, i.e., estimate rating outcomes of 1, from potential rating outcome 0 (missingness indicator 0), utilizing exposure and inherent factors (external information) (pg. 3, section 3.2; etc.) where the inherent factors are learned for the user and item by the deconfounder model (pg. 5, section 3.4, etc.), where the latent vectors representing the inherent factors of the user/item from interactions are the reference vicinity for the missing rating (missingness indicator 0)].
As per claim 8, Li teaches wherein the MNAR data is addressed without using an assumption or a restriction about the distribution of missing values [the distribution of missing values is learned (pgs. 4-5, sections 3.3-3.5)].
As per claim 9, Li teaches wherein one of the estimation is constructed by regressing an outcome on features in labeled data [a logistic regression network is used (pg. 5, section 3.3.2) to estimate the exposure and rating values from the rating (labeled) and exposure (unlabeled) spaces (pg. 2, fig. 2, etc.)].
As per claim 10, Li teaches wherein other of the estimation is constructed by regressing a missingness indicator on the features in both the labeled data and unlabeled data [a logistic regression network is used (pg. 5, section 3.3.2) to estimate the exposure and rating values from the rating (labeled) and exposure (unlabeled) spaces (pg. 2, fig. 2, etc.) to estimate rating outcomes of 1, from potential rating outcome 0 (missingness indicator 0), utilizing exposure and inherent factors (pg. 3, section 3.2; etc.)].
As per claim 14, Li teaches wherein the reassembling is performed using a linear combination [the obtained inherent factors from the deconfounder model and the learned exposure information are combined to create the estimated combined ratings through a combined linear model (pg. 4, fig. 4; pg. 5, section 3.5; etc.)].
As per claim 17, see the rejection of claim 1, above, wherein Li also teaches a system for accounting for missing not-at-random (MNAR) data in training dataset by performing Bayesian regularization, the system comprising:
a memory; and
a processor,
wherein the system is configured to perform [the method] [the system can be implemented on a server with at least one GPU executing instructions from associated memories (pg. 6, section 4.1.4; etc.)].
As per claim 18, see the rejection of claim 2, above.
As per claim 20, see the rejection of claim 1, above, wherein Li also teaches a non-transitory computer readable storage medium that stores a computer program for accounting for missing not-at-random (MNAR) data in training dataset by performing Bayesian regularization, the computer program, when executed by a processor, causing a system to perform a plurality of processes comprising: [the method] [the system can be implemented on a server with at least one GPU executing instructions from associated memories (pg. 6, section 4.1.4; etc.)].
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) 13, 15, and 16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Li et al. (Be Causal: De-biasing Social Network Confounding in Recommendation, May 2021, pgs. 1-11) in view of Linero et al. (Bayesian Approaches for Missing Not at Random Outcome Data: The Role of Identifying Restrictions, May 2018, pgs. 1-25 – cited in an IDS).
As per claim 13, Li teaches the method according to claim 1, as described above.
While Li teaches reassembling quantitative data for ratings including qualitative assessment from prior data (see above), it has not been relied upon for teaching wherein the reassembling incorporates priors over the MNAR outcomes.
Linero teaches wherein the reassembling incorporates priors over the MNAR outcomes [the Bayesian approach allows uncertainty about the unidentified components of the model to be encoded in an informative prior, and for expert knowledge to be formally incorporated into the analysis by eliciting informative priors on sensitivity parameters, including independent priors for each treatment (outcome) (pg. 2, final paragraph; pgs. 17-18, section 5; etc.)].
Li and Linero are analogous art, as they are within the same field of endeavor, namely utilizing models to account for MNAR data.
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to incorporate priors over the MNAR outcomes in reassembling the outcomes/estimates, as taught by Linero, in the reassembling of the outcomes/estimates in the system taught by Li.
Linero provides motivation as [An advantage of the Bayesian approach is that it allows for uncertainty about the unidentified components of the model to be encoded in an informative prior, allowing the analyst to incorporate subject-matter expertise formally into the analysis and the Bayesian framework allows for expert knowledge to be formally incorporated into the analysis by eliciting informative priors on sensitivity parameters. This also allows for simultaneous inference about functionals of the full-data distribution, rather than just a specifically chosen functional such as the mean; for example it is possible to make inferences about means and quantiles simultaneously (pgs. 2-3, section 1; etc.) and allows incorporating uncertainty. An advantage of this approach is that it combines all restrictions under consideration to achieve a single, final, inference (pg. 13, section 3.3; etc.) and avoids repeating computations (pg. 15, section 4.1; etc.)].
As per claim 15, Li/Linero teaches wherein the injecting of the quantitative information includes injecting the qualitative information in a form of a prior over E[m0(x)] at each point x [the Bayesian approach allows uncertainty about the unidentified components of the model to be encoded in an informative prior, and for expert knowledge to be formally incorporated into the analysis by eliciting informative priors on sensitivity parameters, including independent priors for each treatment (outcome) (Linero: pg. 2, final paragraph; pgs. 17-18, section 5; etc.) at each point of the distribution (Linero: pg. 15, section 4; etc.), where the prior for each sample is the form E[m0(x)]].
Examiner’s Note: the reasoning and motivation for the combination is the same as that provided, above, in the rejection of claim 13.
As per claim 16, Li/Linero teaches wherein the quantitative information for remaining one of the at least three unknown quantities is injected by a domain expert [the Bayesian approach allows uncertainty about the unidentified components of the model to be encoded in an informative prior, and for expert knowledge to be formally incorporated into the analysis by eliciting informative priors on sensitivity parameters, including independent priors for each treatment (outcome) (Linero: pg. 2, final paragraph; pgs. 17-18, section 5; etc.)].
Examiner’s Note: the reasoning and motivation for the combination is the same as that provided, above, in the rejection of claim 13.
Claim(s) 6 is/are rejected under 35 U.S.C. 103 as being unpatentable over Li et al. (Be Causal: De-biasing Social Network Confounding in Recommendation, May 2021, pgs. 1-11) in view of Sharifi (US 2021/0082577).
As per claim 6, Li teaches the method according to claim 1, as described above.
While Li also teaches a Bayes risk minimizer including minimizing a loss/error function (see above), it has not been relied upon for teaching wherein the Bayes risk minimizer is 0-1 loss.
Sharifi teaches wherein the Bayes risk minimizer is 0-1 loss [a prediction model can be generated by minimizing a 0-1 loss function for accuracy and a L0 norm for sparsity, subject to the prediction criteria (paras. 0050-52, etc.)].
Li and Sharifi are analogous art, as they are within the same field of endeavor, namely optimizing machine learning models for estimating outcomes for users (which can include treatment outcomes).
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to utilize multiple risk minimizer functions, including 0-1 loss, in optimizing the prediction model, as taught by Sharifi, for the loss function minimization in the optimization of the prediction model in the system taught by Li.
Sharifi provides motivation as [minimizing the 0-1 loss function improves the accuracy of the prediction model, while remaining within the other constraints imposed (paras. 0050-52, etc.)].
Claim(s) 11, 12, and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Li et al. (Be Causal: De-biasing Social Network Confounding in Recommendation, May 2021, pgs. 1-11) in view of well-known practices in the art.
As per claim 11, Li teaches the method according to claim 1, as described above.
While Li teaches the estimation being performed using at least a machine learning model (see above), it has not been relied upon for teaching wherein the estimation of the two of the at least three unknown quantities is performed by training two separate machine learning models.
However, the examiner takes official notice that utilizing multiple machine learning models (e.g., an ensemble of machine learning models) is old and well known within the art for improving the accuracy of outputs. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to perform the estimation of the two of the at least three unknown quantities in the system/method taught by Li by training two separate machine learning models (e.g., separate machine learning models may be trained to produce the latent vectors for the user and item features separately), to achieve the predictable result of improving the accuracy/performance of the model outputs. Additionally, it has been held that mere duplication of the essential working parts of a device (in this case, duplicating a machine learning model to use two machine learning models), involves only routine skill in the art. St. Regis Paper Co. v. Bemis Co., 193 USPQ 8.
As per claim 12, Li teaches the method according to claim 1, as described above.
While Li teaches the estimation being performed using at least a machine learning model (see above), it has not been relied upon for teaching wherein a first of the two machine learning models is trained on both labeled data and unlabeled data, and a second of the two machine learning models is trained based on the labeled data only.
However, the examiner takes official notice that utilizing different combinations of labeled and unlabeled training data in a semi-supervised training routine to train different models of an ensemble is old and well-known within the art. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to utilize semi-supervised training for the machine learning model(s) of the system taught by Li including wherein a first of the two machine learning models is trained on both labeled data and unlabeled data, and a second of the two machine learning models is trained based on the labeled data only, to achieve the predictable result of improving the accuracy/performance of the model for the combination of unlabeled and labeled training data used by the system taught by Li (see, e.g., Li: fig. 2)
As per claim 19, see the rejection of claim 11, 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.
Tseng et al. (Regularized approach for data missing not at random, 2017, pgs. 134-150 – cited in an IDS) – discloses a system/method utilizing LASSO and ridge-regularized selection models to estimate the MNAR data.
Busha (US 2023/0134223) – discloses a system/method for predicting missing category data for cookies, utilizing a semi-supervised training for an ensemble of models.
Xiong (US 2020/0234199) – discloses a system/method for semi-supervised training of an ensemble of models for estimating missing variable data.
Choi (US 2016/0217385) – discloses a system/method for estimating MNAR training data for a recommendation system.
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