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 .
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 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 3/26/2026 has been entered.
Status of Claims
Claims 1-3 and 5-20 remain pending, and are rejected.
Claim 4 has been cancelled.
Response to Arguments
Applicant’s arguments filed on 3/26/2026 with respect to the rejection under 35 U.S.C. 101 for claims directed to a judicial exception are not persuasive.
Applicant’s arguments filed on 3/26/2026 with respect to the rejection under 35 U.S.C. 101 for claims directed to a judicial exception are not persuasive.
Notably, on pages 16-17 of the Applicant’s Remarks, arguments are made that the present claims are directed to improvements to computer component or system performance based upon adjustments to parameters of a machine learning model. On page 18, it is argued that the claims recite training a machine learning model based on a specific set of operations, and also recites the execution of the trained machine learning model to represent items as embeddings in an embedding space. It is also argued that a very particular way of training the data is recited, including the obtaining of co-purchase data, establishing a first Gaussian distribution, parameterizing a mean and variance, and using those values to train the machine learning model. On page 19, the Applicant argues that the claims do not set forth any mathematical relationships, calculations, formulas, or equations using words or mathematical symbols. On page 20, it is further argued that the claims provide a technical improvement that trains a machine learning model, and uses the trained machine learning model to address technical problems in prior systems, including issues related to noisy data, including the lack of ground truth data to train machine learning models, citing specification paragraph [0003-0005] disclosing the problem of failing to account for the fact that co-purchased items are not necessarily complementary to each other and that co-viewed data are noisy. Ex Parte Carmody is cited on pages 20-21 as comparison showing that the reciting of an improvement in training models for use by a recommendation engine and citing Ex Parte Desjardins in its decision. Pages 21-22 argue that the claims also provide significantly more than any abstract idea including training a machine learning model using a specific set of operations and using the trained machine learning model to represent items as embeddings in an embedding space.
Examiner respectfully disagrees. The claims do not recite any particular technology behind the training or functioning of machine learning models, the parameters and set of operations are directed to the mathematical processes performed on the data to input to the machine learning model. The receiving of co-purchase data of a plurality of items, establishing a first Gaussian distribution based on the co-purchase data, parameterizing a mean of the Gaussian distribution, parameterizing a variance of the first Gaussian distribution are merely gathering of abstract data, and using statistical analysis (Gaussian distribution) to the data and the mean and variance. The claims then merely recite training the machine learning model based on these values as input without any recitation of the specific technical process of training machine learning models. The generating of item pairs, triplets with corresponding negative samples represent vectorizing data to perform further mathematical concepts on, such as minimizing the total loss function and mean vectors and matrices. As such, the claims do not recite any technical improvements to the training of machine learning models or any machine learning functionality; the claims merely reciting the mathematical manipulation of data to input into the machine learning model. Furthermore, the cited paragraphs [0003-0005] do not disclose any technical endeavors, but disclose the determining of identifying complementary relationships for item pairs, and methods of determining these relationships such as utilizing abstract data such as the co-purchase signal of two items, and determining data that is not quite correlated to the items being complementary. The noisy data does not represent any technical components, but merely represent data that does not necessarily represent items that are complementary to each other. The problem is dealing with determining the right data to input to generate a more accurate output of determining complementary items to recommend, which is a sales activity.
With regard to the comparison to Ex Parte Desjardins, the Appeals Review Panel determined that the specification identified improvements as to how the machine learning model itself operates, including training a machine learning model to learn new tasks while protecting knowledge about previous tasks to overcome the problem of “catastrophic forgetting” encountered in continuous learning systems, and were tantamount to how the machine learning model itself would function in operation and therefore not subsumed in the identified mathematical calculation. The present invention does not disclose such technical problems, as discussed above, but recite the inaccuracy from noisy data, which merely represents data of the abstract idea that does not represent the abstract relationship between items. While the Examiner appreciates the citation of Ex Parte Carmody, it is not precedential case law, and current USPTO policy is dictated by the body of precedential case law. Notably, the comparisons to Desjardins were addressed above.
In view of the above, the rejection under 35 U.S.C. 101 has been maintained below.
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 1-3 and 5-20 are rejected under 35 U.S.C. 101 because the claims are directed to a judicial exception without significantly more.
Step 1:
Claims 1-3 and 5-9 are directed to a system, which is an apparatus. Claims 10-15 are directed to a method, which is a process. Claims 16-20 are directed to a non-transitory computer readable medium, which is an article of manufacture. Therefore, claims 1-3 and 5-20 are directed to one of the four statutory categories of invention.
Step 2A (Prong 1):
Taking claim 1 as representative, claim 1 sets forth the following limitations reciting the abstract idea of scoring and ranking complementary items to recommend for an anchor item:
obtaining co-purchase data of a plurality of items;
establishing a first Gaussian distribution based on the co-purchase data;
parameterizing a mean of the first Gaussian distribution to represent co-purchases associated with a complementary relationship;
parameterizing a variance of the first Gaussian distribution to represent co-purchases associated with noise;
generating a plurality of positive item pairs based on a plurality of co-purchase item pairs from a plurality of transactions;
for each of the plurality of positive item pairs and each user of a plurality of users, generating a query-recommendation-user triplet and its corresponding negative samples;
generating initial item embeddings for each query-recommendation-user triplet and its corresponding negative samples as a Gaussian distribution with a random mean vector and a random covariance matrix
minimizing a total loss function based on the initial item embeddings;
represent each item of a set of items as an item embedding in an embedding space,
wherein the item embedding is a Gaussian distribution with a mean vector and a non-zero covariance matrix;
the mean vector of each item represent a location of the item in the embedding space with a maximum probability density;
the non-zero vector of each item represent a non-zero variation in a co-purchase behavior of the item;
determine, based on a selection by a user, an anchor item to be displayed to a user;
represent the anchor item as an anchor embedding in the embedding space, wherein the anchor embedding is a Gaussian distribution with an anchor mean vector and an anchor non-zero covariance matrix;
compute a complementarity score for each item of the set of items, based on a distance between the mean vector of the item and the anchor mean vector;
generate a ranking for the set of items based on their respective complementarity scores;
select a plurality of top items in the set of items based on the ranking as recommended complementary items for the anchor item;
transmit information about the recommended complementary items to the user to be displayed with the anchor item.
The recited limitations above set forth the process for scoring and ranking complementary items to recommend for an anchor item. These limitations amount to certain methods of organizing human activity, including commercial or legal transactions (e.g. agreements in the form of contracts, advertising, marketing or sales activities or behaviors, etc.). The claims are directed to determining an anchor item and complementary items to rank and display to a user (see specification [0003-0004] disclosing the problem of using co-purchase signals of items to determine complementary relationships and data that does not necessarily represent complementary items, and the inaccurate data leading to business metrics), which is an advertising and marketing activity. The limitations also amount to mathematical concepts, including mathematical formulas and calculation. The claims are directed to embeddings that are a Gaussian distribution with a mean vector and non-zero covariance matrix and computing a distance between the mean vectors, which are mathematical calculations.
Such concepts have been identified by the courts as abstract ideas (see: MPEP 2106.04(a)(2)).
Step 2A (Prong 2):
Examiner acknowledges that representative claim 1 recites additional elements, such as:
a processor;
a non-transitory computer-readable medium;
train a machine learning model;
training the machine learning model based on the parameterized mean of the first Gaussian distribution and the parameterized variance of the first Gaussian distribution to learn to represent each item of the plurality of items as a Gaussian embedding following the first Gaussian distribution;
execute the trained machine learning model;
learned by the machine learning model;
a user interface executed on a user device of the user;
Taken individually and as a whole, representative claim 1 does not integrate the recited judicial exception into a practical application of the exception. The additional elements do no more than generally link the use of a judicial exception to a particular technological environment or field of use.
Furthermore, this is also because the claim fails to (i) reflect an improvement in the functioning of a computer, or an improvement to other technology or technical field, (ii) implement a judicial exception with a particular machine, (iii) effect a transformation or reduction of a particular article to a different state or thing, or (iv) apply the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment.
While the claims recite non-transitory computer-readable storage medium, a processor, and a user interface executed on a user device, these elements are recited with a very high level of generality. Specification paragraphs [0032-0033] disclose generic components for the recommendation computing device, such as any generic processor, ASICs, state machines, digital circuitry, etc. The paragraphs also disclose that the item recommendation computing device can be a computer, a workstation, a laptop, a server, or any other suitable device. The user device is disclosed in paragraph [0034] as being a cellular phone, a smart phone, tablet, etc. As such, it is clear that any computing devices recited in the claims are generic computing devices that are merely applied to the abstract idea to provide a general link to a computing environment, but the focus of the claims are to the abstract idea of mathematical calculations to determine complementary items to display to a user. The claims also recite a machine learning model, it is recited with a very high level of generality, merely reciting that the machine learning model learns and trains with various data without any recitation of the technical underlying process of how that is being performed. Specification paragraph merely discloses that the item recommendation device may execute one or more models, including a machine learning model, a statistical model, etc. There is no disclosure of any specific machine learning model, or how any machine learning model functions. It is evident that any machine learning model, and any learning of training of the machine learning model, is by any generic machine learning technology, and is merely applied to the abstract idea to provide an output.
In view of the above, under Step 2A (Prong 2), representative claim 1 does not integrate the recited exception into a practical application (see: MPEP 2106.04(d)).
Step 2B:
Returning to representative claim 1, taken individually or as a whole, the additional elements of claim 1 do not provide an inventive concept (i.e. whether the additional elements amount to significantly more than the exception itself). As noted above, the additional elements recited in claim 1 are recited in a generic manner with a high level of generality and only serve to implement the abstract idea on a generic computing device. The claims result only in an improved abstract idea itself and do not reflect improvements to the functioning of a computer or another technology or technical field. As discussed above with respect to the integration of the abstract idea into a practical application, the additional elements used to perform the claimed process ultimately amount to no more than the mere instructions to apply the exception using a generic computer and/or no more than a general link to a technological environment.
Even when considered as an ordered combination, the additional elements of claim 1 do not add anything further than when they are considered individually.
In view of the above, claim 1 does not provide an inventive concept under step 2B, and is ineligible for patenting.
Regarding Claim 10 (method): Claim 10 recites at least substantially similar concepts and elements as recited in claim 1 such that similar analysis of the claims would be readily apparent to one of ordinary skill in the art. As such, claims 10 is rejected under at least similar rationale as provided above regarding claim 1.
Regarding Claim 16 (non-transitory computer readable medium): Claim 16 recites at least substantially similar concepts and elements as recited in claim 1 such that similar analysis of the claims would be readily apparent to one of ordinary skill in the art. As such, claims 16 is rejected under at least similar rationale as provided above regarding claim 1.
Dependent claims 2-3, 5-9, 11-15, and 17-20 recite further complexity to the judicial exception (abstract idea) of claim 1, such as by further defining the algorithm of scoring and ranking complementary items to recommend for an anchor item, and do not recite any further additional elements. Thus, each of claims 2-3, 5-9, 11-15, and 17-20 are held to recite a judicial exception under Step 2A (Prong 1) for at least similar reasons as discussed above.
Under prong 2 of step 2A, the additional elements of dependent claims 2-3, 5-9, 11-15, and 17-20 also do not integrate the abstract idea into a practical application, considered both individually or as a whole. More specifically, dependent claims 2-3, 5-9, 11-15, and 17-20 rely on at least similar elements as recited in claim 1. Further additional elements are also acknowledged; however, the additional elements of claims 2-3, 5-9, 11-15, and 17-20 are recited only at a high level of generality (i.e. as generic computing hardware) such that they amount to nothing more than the mere instructions to implement or apply the abstract idea on generic computing hardware (or, merely uses a computer as a tool to perform an abstract idea). Further, the additional elements do no more than generally link the use of a judicial exception to a particular technological environment or field of use (such as the Internet or computing networks).
Secondly, this is also because the claims fails to (i) reflect an improvement in the functioning of a computer, or an improvement to other technology or technical field, (ii) implement the judicial exception with, or use the judicial exception in conjunction with, a particular machine or manufacture that is integral to the claim, (iii) effect a transformation or reduction of a particular article to a different state or thing, or (iv) applies or uses the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment.
Taken individually and as a whole, dependent claims 2-3, 5-9, 11-15, and 17-20 do not integrate the recited judicial exception into a practical application of the exception under step 2A (prong 2).
Lastly, under step 2B, claims 2-3, 5-9, 11-15, and 17-20 also fail to result in “significantly more” than the abstract idea under step 2B. The dependent claims recite additional functions that describe the abstract idea and use the computing device to implement the abstract idea, while failing to provide an improvement to the functioning of a computer, another technology, or technical field. The dependent claims fail to confer eligibility under step 2B because the claims merely apply the exception on generic computing hardware and generally link the exception to a technological environment.
Even when viewed as an ordered combination (as a whole), the additional elements of the dependent claims do not add anything further than when they are considered individually.
Taken individually or as an ordered combination, the dependent claims simply convey the abstract idea itself applied on a generic computer and are held to be ineligible under Steps 2B for at least similar rationale as discussed above regarding claim 1. Thus, dependent claims 2-3, 5-9, 11-15, and 17-20 do not add “significantly more” to the abstract idea.
Subject Matter Free of the Prior Art
The following is a restatement of the reasons for indicating subject matter free of the prior art that was previously mailed on 8/6/2025.
Claim 1-3 and 5-20 are determined to have overcome the prior art of rejection and are free of the prior art, however, the claims remain rejected under 35 U.S.C. 101, as set forth above.
Claims 1-3 and 5-20 are found to overcome the prior art rejection for the reasons as set forth below.
Claim 1 recites the claimed features of:
for each of the plurality of positive item pairs and each user of a plurality of users, generating a query-recommendation-user triplet and its corresponding negative samples;
generate initial item embeddings for each query-recommendation-user triplet and its corresponding negative samples as a Gaussian distribution with a random mean vector and a random covariance matrix;
The closes prior art was found to be as follows:
Bhattacharjee (US 20190370879 A1) discloses [0041] – “generate item complementarity values. In the example shown, a Siamese network with two identical subnetworks consisting of k Fully-Connected hidden layers is employed for classification of items as complementary or not. In example embodiments, identical subnetworks are used, i.e., those that have the same configuration with the same parameters and weights. The purpose of using this network is to project these text embeddings into an N-dimensional vector space where embeddings for items that are complementary to each other are located closer in the space, and those that are not, are located further apart”.
Kysta (US 20210035188 A1) discloses [0135] – “a model is trained based on information about products previously purchased by each customer as stored in a customer purchase history database and product similarities as stored in the product similarity database 300. At this step S203 a model, for example, a machine learning model, is trained in customers' behaviours based on previous purchases made by customers and similar products of those orders, for example, products which are commonly purchased together. In this regard, previous customers' purchases are stored in a customer purchase history database. For example, the database may store information about specific orders placed by customers including a customer's name, address, products purchased, date and time of delivery etc. On the other hand, product similarities are stored in the product similarity database”.
Saville (US 20230019794 A1) discloses [0151] – “the complementarity model may be further be trained on the purchase history entries 271 (shown in FIG. 12), wherein items that a particular user purchases at the same time, or within a short time frame of each other, may be classified as “complementary” items. For example, as described in greater detail below in association with a shopping cart page shown generally at 780 in FIG. 35, a new instance of the purchase history entry 271 may be created each time a user purchases items or an item collection. The complementarity model may categorize item entries 171 identified in the item identifier field 274 of a single such purchase history entry 271 as “complementary” items. Alternatively, the complementarity model may also categorize different item entries 171 identified in respective item identifier fields 274 of a plurality of such purchase history entries 271 as “complementary” items, if the purchase history entries 271 also identify a same user in the user identifier fields 273, a same image (or palette) in the image identifier fields 275, and/or store times in the created fields 278 that are separated by a time gap below a purchase time gap threshold, to the complementarity model as records of the user purchasing “complementary” items”.
NPL Reference U (see PTO-892 Reference V mailed on 11/18/2024) discloses determining complementary products based on co-purchase scores and low substitutivity scores using a mixture model. Product features and user-product features generated from multiple time intervals are used to capture the time dynamic of users and product behaviors to determine context for complementary products.
It was found that no references alone or in combination, neither anticipates, reasonable teaches, nor renders obvious the below noted features of Applicant’s invention. The features of claim 1 in combination that overcome the prior art are:
for each of the plurality of positive item pairs and each user of a plurality of users, generating a query-recommendation-user triplet and its corresponding negative samples;
generate initial item embeddings for each query-recommendation-user triplet and its corresponding negative samples as a Gaussian distribution with a random mean vector and a random covariance matrix;
Therefore, none of the cited references disclose or render obvious each and every feature of the claimed invention and the claimed invention is determined to be free of the prior art. Although individually the claimed features could be taught, any combination of references would teach the claimed limitations using a piecemeal analysis, since references would only be combined and deemed obvious based on knowledge gleaned from the applicant's disclosure. Such a reconstruction is improper (i.e., hindsight reasoning). See In re McLaughlin, 443 F.2d 1392, 170 USPQ 209 (CCPA 1971). The examiner emphasizes that it is the interrelationship of the limitations that renders these claims free of the prior art/additional art.
Therefore, it is hereby asserted by the Examiner that, in light of the above, that claims 1-3 and 5-20 are free of prior art as the references do not anticipate the claims and do not render obvious any further modification of the references to a person of ordinary skill in art.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to TIMOTHY J KANG whose telephone number is (571)272-8069. The examiner can normally be reached Monday - Friday: 8:30am - 7:00pm EST.
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, Maria-Teresa Thein can be reached at 571-272-6764. 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.
/T.J.K./ Examiner, Art Unit 3689
/VICTORIA E. FRUNZI/ Primary Examiner, Art Unit 3689 5/27/2026
Prosecution Insights