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 .
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 2/23/2026 has been entered. Claims 1-19 and 21 are pending. Claim 20 is cancelled.
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.
Step 1: The claims 1-10 and 21 are a method and claims 11-19 are a computer program product. Thus, each independent claim, on its face, is directed to one of the statutory categories of 35 U.S.C. §101. However, the claims 1-19 and 21 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 2A Prong 1: The independent claims (1 and 11) recite:
retrieving prior orders received from users of the computer system, each order including one or more items;
generating, from the prior orders, a tensor having more than two dimensions, each element of the tensor corresponding to a user, an item, and an additional item and having a value indicating a connection between the item and the additional item in the one or more prior orders;
generating an item embedding and a user embedding by applying a tensor decomposition model to the tensor, wherein the tensor decomposition model is trained to generate item embeddings and user embeddings that occupy a common latent space based on tensors input to the tensor decomposition model;
receiving, via a user interface of a client device, a request from a user for an order, requesting one or more items to be included in the order;
generating an order embedding for the order from item embeddings for the one or more items included in the order;
identifying a set of candidate items identified by the order;
generating a score for each candidate item of the set , the score for a candidate item of the set based on the order embedding for the order and an item embedding of the candidate item
selecting one or more candidate items based on the generated scores;
and transmitting an interface to the client device of the user including information describing the selected one or more candidate items
These limitations, except for the italicized portions, under their broadest reasonable interpretations, recite certain methods of organizing human activity for managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules or instructions) as well as commercial or legal interactions (including agreements in the form of contracts; legal obligations; advertising, marketing or sales activities or behaviors; business relations). The claimed invention recites steps for receiving retrieving previous user data related to orders, generating a tensor based on a set of data related to a user and items, receiving a request for an order, generating an order embedding for items in the order, identifying a set of candidate items and selecting candidate items based on the scores. The steps under its broadest reasonable interpretation specifically fall under sales activities. The Examiner notes that although the claim limitations are summarized, the analysis regarding subject matter eligibility considers the entirety of the claim and all of the claim elements individually, as a whole, and in ordered combination.
These limitations, except for the italicized portions, under their broadest reasonable interpretations, recite mathematical formulas and relationships. Specifically, the steps of generating, from the prior orders, a tensor having more than two dimensions, each element of the tensor corresponding to a user, an item, and an additional item and having a value indicating a connection between the item and the additional item in the one or more prior orders; generating a score for each candidate item of the set , the score for a candidate item of the set based on the order embedding for the order and an item embedding of the candidate item recite generating data sets based on mathematical relationships. The Examiner notes that although the claim limitations are summarized, the analysis regarding subject matter eligibility considers the entirety of the claim and all of the claim elements individually, as a whole, and in ordered combination.
Prong 2: This judicial exception is not integrated into a practical application. In particular, the claims recite the additional elements of
A method, performed at a computer system comprising a processor and a computer-readable medium, comprising (claim 1)
A computer program product comprising a non-transitory computer readable storage medium having instructions encoded thereon that, when executed by a processor, cause the processor to: (claim 11)
the computer system
wherein the tensor decomposition model is trained
receiving, via a user interface of a client device,
and transmitting an interface to the client device of the user including information describing the selected one or more candidate items
The additional elements of A computer program product comprising a non-transitory computer readable storage medium having instructions encoded thereon that, when executed by a processor, cause the processor to: (claim 11); the computer system; receiving, via a user interface of a client device, wherein the tensor decomposition model is trained are recited at a high-level of generality (i.e., as a generic processor performing a generic computer function of processing data) such that it amounts no more than mere instructions to apply the exception using a generic computer component. These limitations do not impose any meaningful limits on practicing the abstract idea, and therefore do not integrate the abstract idea into a practical application – MPEP 2106.05(f).
The additional elements of “and transmitting an interface to the client device of the user including information describing the selected one or more candidate items” are merely adding insignificant extra-solution activity to the judicial exception by providing data in the form of transmitting resulting the information (i.e. data output) - see MPEP 2106.05(g).
Accordingly, these additional elements when considered individually or as a whole do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The independent claims are directed to an abstract idea.
Step 2B: The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed with respect to Step 2A Prong two, the additional elements in the claims amount to no more than mere instructions to apply the judicial exception using a generic computer component.
The step of “and transmitting an interface to the client device of the user including information describing the selected one or more candidate items” of Step 2A has been re-evaluated in Step 2B and determined to be well-understood, routine, conventional activity in the field. The receiving or transmitting data over a network, e.g., using the Internet to gather data has been determined to be recognized as well understood, routine and conventional function when it is claimed in a merely generic manner. The claimed limitation merely transmits to a user device the result of the selected candidate items- MPEP 2106.05(d)(II)
For these reasons, there is no inventive concept and the claims are not patent eligible.
Dependent claims 2-10, 21 and 12-19 when analyzed as a whole, are held to be patent ineligible under 35 U.S.C. §101 because the additional recited limitations fail to establish that the claims are not directed to the same abstract idea of Independent Claims 1 and 11 without significantly more.
Claim 2 recites the method of claim 1, wherein generating the order embedding for the order from item embeddings for the one or more items included in the order comprises: generating the order embedding as an average of the item embeddings for the one or more items included in the order. The claim merely further limits the abstract idea without adding significantly more to integrate the abstract idea into a practical application.
Claim 3 recites the method of claim 2, wherein generating the score for each candidate item of the set, the score for the candidate item of the set based on the order embedding for the order and the item embedding of the candidate item comprises: determining a product of the order embedding and the user embedding for the user; determining the score for the candidate item of the set as a measure of similarity between the determined product and the item embedding for the candidate item of the set. The claim merely further limits the abstract idea without adding significantly more to integrate the abstract idea into a practical application.
Claim 4 recites the method of claim 3, wherein the measure of similarity between the determined product and the item embedding for the candidate item of the set comprises a dot product between the determined product and the item embedding for the candidate item of the set. The claim merely further limits the abstract idea without adding significantly more to integrate the abstract idea into a practical application.
Claim 5 recites the method of claim 1, wherein generating the order embedding for the order from item embeddings for the one or more items included in the order comprises: generating the order embedding as a product of an average of the item embeddings for the one or more items included in the order and the user embedding. The claim merely further limits the abstract idea without adding significantly more to integrate the abstract idea into a practical application.
Claim 6 recites the method of claim 5, wherein generating the score for each candidate item of the set, the score for the candidate item of the set based on the order embedding for the order and the item embedding of the candidate item comprises: determining the score for a candidate item of the set as a measure of similarity between the order embedding and the item embedding for the candidate item of the set. The claim merely further limits the abstract idea without adding significantly more to integrate the abstract idea into a practical application.
Claim 7 recites the method of claim 6, wherein the measure of similarity between the order embedding and the item embedding for the candidate item of the set comprises a dot product between the order embedding and the item embedding for the candidate item of the set. The claim merely further limits the abstract idea without adding significantly more to integrate the abstract idea into a practical application.
Claim 8 recites the method of claim 1, wherein selecting one or more candidate items based on the generated scores comprises: ranking the candidate items of the set based on the generated scores; and selecting one or more candidate items of the set having at least a threshold position in the ranking. The claim merely further limits the abstract idea without adding significantly more to integrate the abstract idea into a practical application.
Claim 9 recites the method of claim 1, wherein selecting one or more candidate items based on the generated scores comprises: selecting one or more candidate items of the set having at least a threshold score. The claim merely further limits the abstract idea without adding significantly more to integrate the abstract idea into a practical application.
Claim 10 recites the method of claim 1, wherein identifying the set of candidate items by the order comprises: identifying items offered by the order having item embeddings within a threshold distance from the order embedding. The claim merely further limits the abstract idea without adding significantly more to integrate the abstract idea into a practical application.
Claim 21 recites wherein each element of the tensor further specifies a number of times the user has included the item and the additional item in orders fulfilled by the computer system. The claim merely further limits the abstract idea without adding significantly more to integrate the abstract idea into a practical application.
The above analysis also applies to claims 12-19 and 21 as they recite the same claimed subject matters. For these reasons claims 1-19 and 21 are rejected under 35 USC 101.
Subject Matter Free of Prior Art
Claims 1-19 and 21 recite subject matter free of prior art, however remain rejected under 35 USC 101.
Taking amended claim 1 as a representative claim, the claims as amended are found to overcome the prior art rejection for the reasons set forth below.
Claim 1 now recites the additional claimed features of generating, from the prior orders, a tensor having more than two dimensions, each element of the tensor corresponding to a user, an item, and an additional item and having a value indicating a connection between the item and the additional item in the one or more prior orders; generating an item embedding and a user embedding by applying a trained tensor decomposition model to the tensor, wherein the tensor decomposition model is trained to generate item embeddings and user embeddings that occupy a common latent space based on tensors input to the tensor decomposition model
The closest prior art was found to be as follows:
Biswas (US 20220351021) discloses producing embeddings from decomposed matrices in order to product recommendations to a user based on a scoring (see 0056 and 0067 and previously cited portions in the prior art rejection in the previous office action).
Yalla (US 11645559) discloses the collection of user transaction data that includes item IDs, user IDs and item information such as categories (Figure 3). The personalization service uses lookup embedding vectors to transform vectors into tensors. The tensors are transformed to recommendations (see Figure 5). However, the reference does not disclose generating, from the prior orders, a tensor having more than two dimensions, each element of the tensor corresponding to a user, an item, and an additional item and having a value indicating a connection between the item and the additional item in the one or more prior orders; generating an item embedding and a user embedding by applying a trained tensor decomposition model to the tensor, wherein the tensor decomposition model is trained to generate item embeddings and user embeddings that occupy a common latent space based on tensors input to the tensor decomposition model.
“Tensor Methods and Recommender Systems” discloses tensor-based recommender models (abstract). The reference discloses a tensor of order 3 that is then reduced using decomposition modeling to creating a representative embedding (pages 10-11). Examples of using triplets in the tensor modeling are disclosed (pages 7, 16, and 17), however the reference recites these concepts at a high level of application and does not specifically disclose generating, from the prior orders, a tensor having more than two dimensions, each element of the tensor corresponding to a user, an item, and an additional item and having a value indicating a connection between the item and the additional item in the one or more prior orders; generating an item embedding and a user embedding by applying a trained tensor decomposition model to the tensor, wherein the tensor decomposition model is trained to generate item embeddings and user embeddings that occupy a common latent space based on tensors input to the tensor decomposition model.
“MULTILAYER TENSOR FACTORIZATION WITH APPLICATIONS TO RECOMMENDER SYSTEMS” discloses the tensor, also called multidimensional array, is well-recognized as a powerful tool to represent complex and unstructured data [53]. It is applied in many areas such as signal processing, neuroimaging, and psychometrics [e.g., 32, 27, 26]. In recommender systems, the tensor shows its flexibility to accommodate contextual information, and is also regarded as one of the most effective tools for developing context-aware recommender systems [CARS; 2, 3]. In addition to user and item information from traditional recommender systems, tensor-based recommender systems also take the effect of contextual variables into account, such as time, location, users’ companions, stores’ promotion strategies, other relevant variables, or any combinations thereof. Hence, CARS are capable of utilizing more information and provide more accurate recommendations. The reference also discloses the application of Tucker decomposition (see page 2). High-order tensors are generally disclosed corresponding to user, item or a contextual variable, and each element of the tensor represents a utility, such as a rating or sales volume (see page 3). However, the reference recites these concepts at a high level of application and does not specifically disclose generating, from the prior orders, a tensor having more than two dimensions, each element of the tensor corresponding to a user, an item, and an additional item and having a value indicating a connection between the item and the additional item in the one or more prior orders; generating an item embedding and a user embedding by applying a trained tensor decomposition model to the tensor, wherein the tensor decomposition model is trained to generate item embeddings and user embeddings that occupy a common latent space based on tensors input to the tensor decomposition model.
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 (and parallel claim 11) in combination that overcome the prior art are:
generating, from the prior orders, a tensor having more than two dimensions, each element of the tensor corresponding to a user, an item, and an additional item and having a value indicating a connection between the item and the additional item in the one or more prior orders; generating an item embedding and a user embedding by applying a trained tensor decomposition model to the tensor, wherein the tensor decomposition model is trained to generate item embeddings and user embeddings that occupy a common latent space based on tensors input to the tensor decomposition model.
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 high level disclosure of the mathematical concepts 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 the claims 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.
With respect to claims 3-7 and 13-17, the closest prior art was found to be the following:
Regarding claims 3 and 13, Erez US 20190266482 which teaches multiplying of a vector by a row of a matrix for embedding vectors classifying items which creates a scalar score. This scalar score indicating the similarity between an embedded vector and an item stored in a matrix. (see [0070 and 0079]). However it would not have been obvious to one of ordinary skill in the art alone or in combination with Biswas in view of Carbune, to teach wherein generating the score for each candidate item of the set, the score for the candidate item of the set based on the order embedding for the order and the item embedding of the candidate item comprises: determining a product of the order embedding and the user embedding for the user; determining the score for the candidate item of the set as a measure of similarity between the determined product and the item embedding for the candidate item of the set.
Regarding claims 4 and 14, Biswas discloses the application of dot product for matrices in [0027, 36, 54 and 64] in order to make a recommendation of a product. However the reference alone nor in combination with Carbune or Erez does not teach wherein the measure of similarity between the determined product and the item embedding for the candidate item of the set comprises a dot product between the determined product and the item embedding for the candidate item of the set, nor would it have been obvious to combine to one of ordinary skill in the art before the effective filing date of the claimed invention.
Regarding claims 5 and 15, Palnitkar teaches determining a score for content using the average between two embeddings, however does not teach alone or in combination with Biswas, Carbune, and/or Erez wherein generating the order embedding for the order from item embeddings for the one or more items included in the order comprises: generating the order embedding as a product of an average of the item embeddings for the one or more items included in the order and the user embedding. Further, it would not have been obvious to combine to one of ordinary skill in the art before the effective filing date of the claimed invention.
Claims 6, 7, 16, and 17 depend on claims 5 and 15 and the same reasoning applies as is stated for claim 5.
As independent claims 1 and 11 are free of prior art, dependent claims 2-10 and 12-19, 21 are also free of prior art by at least virtue of dependency.
For these reasons claims 1-19 and 21 are determined to be free of prior art, however remain rejected under 35 USC 101.
Consideration of Declaration under 37 CFR 1.132
The declaration under 37 CFR 1.132 filed 2/26/2026 is insufficient to overcome the rejection of claims 1-19 and 21 based upon 35 USC 101 as set forth in the last Office action because: the declaration does not provide support for the level of detail required to show technical improvement in concert with the disclosure and claimed invention needed to overcome 35 USC 101. This reasoning is further detailed in the response to arguments which includes excerpts from the declaration.
Response to Arguments
Applicant's arguments filed 2/23/2026 have been fully considered but they are not persuasive. It is also noted that the “Declaration of Shishir Kumar Prasad under 37 CFR 1.132” filed on 2/23/2026 has also been considered, however is not found sufficient to overcome the rejection. The reasons are set forth below.
With respect to the remarks on page 9 regarding the “data size and data sparsity”, while the generating the order embedding from the item embedding for item in the order and using that order embedding to score candidate item may avoid repeated use of the resource intensive machine learning models, the improvement here lies in the abstract idea. The choice of the mathematical approach to the scoring of the candidate items improves the process and thereby consequentially, less resources are needed. The technology itself is not being improved, at most the dataset and selection of the mathematical computation is being improved. This is further shown as the application of the tensor decomposition model to generate the item embeddings and user embeddings can reduce the dimensionality of the large tensors and then provide meaningful (improved) representations of the user and items. The improvement lies here in the mathematical approach and the improved output of recommendations, which is at most an improvement to the abstract idea. Consequentially, the data and usability of the data, as it has been improved, is better data for the computer system to use. The machine learning and computer system here are additional elements, however merely used (“apply it”) to carry out the abstract idea.
With respect to the remarks on pages 10-11 in view of the Desjardins decision, the examiner asserts that using the two-part analysis, the claims remain rejected under 35 USC 101. Examiner notes that the fact patterns of the instant case are different from those set forth in Ex Parte Desjardins, and different fact patterns may have different eligibility outcomes. In Ex Parte Desjardins, the claimed invention was a method of training a machine learning model on a series of tasks, and technical improvements as a result of the model training were identified as reduced storage, reduced system complexity, and preservation of performance attributes associated with earlier tasks during subsequent computational tasks. While the ARP in Ex Parte Desjardins determined that these improvements were sufficient to reverse the 101 rejection of the claims at hand in Ex Parte Desjardins, analogous improvements are not apparent in the instant claims. Furthermore, as discussed below, neither Applicant’s specification nor the instant claims set forth analogous improvements. Accordingly, under the analysis set forth according to the MPEP, discussed below, the amended claims stand as ineligible.
In concert with the response above, the alleged technical improvement actually is an improvement to the abstract idea and reduction in computation resources is merely consequential to the improvement abstract idea. That is, the limitations of "generating an order embedding for the order from item embeddings for the one or more items included in the order"; "identifying a set of candidate items identified by the order"; "generating a score for each candidate item of the set, the score for a candidate item of the set based on the order embedding for the order and an item embedding of the candidate item"; "selecting one or more candidate items based on the generated scores" are part of the abstract idea, shown in the rejection above. The details provided in the disclosure conclude that the implementation of the mathematical operation improve the data itself and thereby reduce the computational resources needed and improve the recommendation data itself. The examiner further notes, the details of the machine learning in the specification (starting at [0033]) provide a high level and general machine learning technique and that the machine learning model can be any number of machine learning models carrying out the computations discussed above and recited in the claimed invention. Further in [0037] the tensor decomposition model can be any of a number of decomposition techniques. The further steps of "generating, from the prior orders, a tensor having more than two dimensions, each element of the tensor corresponding to a user, an item, and an additional item and having a value indicating a connection between the item and the additional item in the one or more prior orders"; "generating an item embedding and a user embedding by applying a tensor decomposition model to the tensor, wherein the tensor decomposition model is trained to generate item embeddings and user embeddings that occupy a common latent space based on tensors input to the tensor decomposition model" that are said to reduce the dimensionality of the large tensors to generate meaningful embedding representations of the users and items based on sparse data, are also part of the abstract idea. As discussed above, the data being improved and thereby the usability of the data stored/used by the computer is merely consequential from the improvement to the abstract idea. While the specification and the claims recite sufficient disclosure to conclude the abstract idea itself may be improved as the background and summary of the instant specification sets forth the instant application as improving the recommendation process over conventional recommendation systems, there are no details showing the technology itself is improved.
The claims and the disclosure recite an application of a mathematical model which improves the recommendations that are output and then used by the system.
Therefore, the remarks considered in concert with the “Declaration of Shishir Kumar Prasad under 37 CFR 1.132” are not found to be sufficient to overcome the rejection under 35 USC 101 remains.
Relevant Art Not Cited
Liang 20180060302 [0245] In some examples, at block 1320, the corrected text segment 608 can be selected from among the plurality of stored candidate text segments based at least in part on the representation(s), e.g., the respective distances. For example, the candidate text segment corresponding to the smallest of the distances can be selected as the corrected text segment 608. In some examples, the corrected text segment 608 can be selected only if the respective distance is less than a selected threshold, or in response to the respective distance being less than a selected threshold. If all distances exceed the threshold, in some examples, the second text segment 608 can be selected to be the corrected text segment 608. In some examples, block 1320 can include selecting the corrected text segment 1608, e.g., using locality-sensitive hashing as described above.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to VICTORIA E. FRUNZI whose telephone number is (571)270-1031. The examiner can normally be reached Monday- Friday 7-4 (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, Marissa 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.
VICTORIA E. FRUNZI
Primary Examiner
Art Unit TC 3689
/VICTORIA E. FRUNZI/Primary Examiner, Art Unit 3689 4/8/2026
Prosecution Insights