DETAILED ACTION
Status of Claims
Claims 1-20 submitted on 07/06/2026 are pending and have been examined. Claims 1, 3, 5-13, and 16 have been amended.
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 .
Priority
Acknowledgment is made of applicant's provisional application. The certified copy has been filed in parent Application No. 63/627,304, filed on 01/31/2024.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception without significantly more. The claims recite an abstract idea. This judicial exception is not integrated into a practical application. The claim(s) do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Step 1
Claims 1-9 are directed to a machine, claims 10-15 are directed to an article of manufacture, and claims 16-20 are directed to a process (see MPEP 2106.03).
Step 2A, Prong 1
Claim 1, taken as representative, recites at least the following limitations that recite an abstract idea:
a system, comprising: obtain a request to select a set of personalized information items from a collection of information items to a user;
classify the user in a first user class of a plurality of user classes, wherein the first user class has a predefined relationship with a first object class of a plurality of object classes;
identify, among a plurality of users, a subset of users that are classified in the first user class;
identify, in the collection of information items, a subset of information items that are classified in the first object class;
identify the one or more personalized information items within a time limit utilizing a collaborative filtering model, by:
obtaining an input including engagement information of the subset of users and the subset of information items;
based on the input, determining the collaborative filtering model by:
determining a user embedding of each of the subset of users;
determining a similarity level between the user and each of the subset of users based on the user embeddings of the subset of users; and
applying the similarity levels of the subset of users with the user as weights to determine, for each of the subset of information items, an accumulated object numbers based on a weighted sum of respective object numbers of the set of related users; and
selecting the one or more personalized information items of the user based on the accumulated object number of each of the subset of information items; and
in response to identifying the one or more personalized information items, generate instructions configured to display an interface including the one or more personalized information items associated with the user.
The above limitation, under its broadest reasonable interpretation, falls within the “Certain Methods of Organizing Human Activity” grouping of abstract ideas, enumerated in MPEP 2106.04(a)(2)(II), in that it recites a commercial interaction. See ¶0003 of the instant specification where it is discussed that the invention improves product recommendation. Under the broadest reasonable interpretation, the above claims also fall within the “Mathematical concepts” grouping of abstract ideas, enumerated in MPEP 2106.04(a)(2)(I), in that they recite mathematical relationships, mathematical formulas or equations, mathematical calculations. Claims 10 and 16 recite similar limitations as claim 1.
Thus, under Prong 1 of Step 2A, claims 1, 10, and 16 recite an abstract idea.
Step 2A, Prong 2
Claim 1 includes the following additional elements that are bolded:
a system, comprising: at least one processor; and a non-transitory memory having instructions stored thereon that, when executed by the at least one processor, cause the at least one processor to: obtain a request to select a set of personalized information items from a collection of information items to a user;
classify the user in a first user class of a plurality of user classes, wherein the first user class has a predefined relationship with a first object class of a plurality of object classes;
identify, among a plurality of users, a subset of users that are classified in the first user class;
identify, in the collection of information items, a subset of information items that are classified in the first object class;
identify the one or more personalized information items within a time limit utilizing a collaborative filtering model, by:
obtaining an input including engagement information of the subset of users and the subset of information items;
based on the input, determining the collaborative filtering model by:
determining a user embedding of each of the subset of users;
determining a similarity level between the user and each of the subset of users based on the user embeddings of the subset of users; and
applying the similarity levels of the subset of users with the user as weights to determine, for each of the subset of information items, an accumulated object numbers based on a weighted sum of respective object numbers of the set of related users; and
selecting the one or more personalized information items of the user based on the accumulated object number of each of the subset of information items; and
in response to identifying the one or more personalized information items, generate instructions configured to display an interface including the one or more personalized information items on a client device associated with the user.
Claims 10 and 16 include the same additional elements as claim 1. In addition, claim 10 includes additional elements such as a non-transitory computer-readable storage medium, having instructions stored thereon, which when executed by one or more processors cause the one or more processors to. In addition, claim 16 includes additional elements such as a system including a non-transitory memory having instructions stored thereon and at least one processor operatively coupled to the non-transitory memory and configured to execute the instructions.
The additional elements recited in claims 1, 10, and 16 merely invoke such elements as a tool to perform the abstract idea and generally link the use of the abstract idea to a particular technological environment of client devices (see MPEP 2106.05(f) and MPEP 2106.05(h). These additional elements are described at a high level in Applicant’s specification without any meaningful detail about their structure or configuration (see ¶0026, ¶0034, and Fig. 1).
As such, under Prong 2 of Step 2A, when considered both individually and as a whole, the additional elements do not integrate the judicial exception into a practical application and, thus, claims 1, 10, and 16 are directed to an abstract idea.
Step 2B
As noted above, while the recitation of the additional elements in independent claims 1, 10, and 16 are acknowledged, claims 1, 10, and 16 merely invoke such additional elements as a tool to perform the abstract idea and generally link the use of the abstract idea to a particular technological environment (see MPEP 2106.05(f) and MPEP 2106.05(h)).
Even when considered as an ordered combination, the additional elements of claims 1, 10, and 16 do not add anything that is not already present when they are considered individually. Therefore, under Step 2B, there are no meaningful limitations in claims 1, 10, and 16 that transform the judicial exception into a patent eligible application such that the claims amount to significantly more than the judicial exception itself (see MPEP 2106.05).
As such, independent claims 1, 10, and 16 are ineligible.
Dependent claims 4, 6-9, 11-14, 17, and 18 when analyzed as a whole, are held to be patent ineligible under 35 U.S.C. 101 because they do not add “significantly more” to the abstract idea. More specifically, dependent claims 4, 6-9, 11-14, 17, and 18 merely further define the abstract limitations of claims 1, 10, and 16 or provide further embellishments of the limitations recited in independent claims 1, 10, and 16. Claims 4, 6-9, 11-14, 17, and 18 do not introduce any further additional elements. Thus, dependent claims 4, 6-9, 11-14, 17, and 18 are ineligible.
Furthermore, it is noted that certain dependent claims recite additional elements supplemental to those recited in independent claims 1, 10, and 16: an object database (claims 2, 3, 19, and 20), user database (claim 5), an online shopping application via one of an internet browser and a dedicated application (claim 15), and a user interface of the online shopping application on a screen (claim 15). However, these elements do not integrate the abstract idea into a practical application because they merely amount to using a computer to apply the abstract idea to a particular technological environment or field of use and thus do not act to integrate the abstract idea into a practical application of the abstract idea. Additionally, the additional elements do not amount to significantly more because they merely amount to using a computer to apply the abstract idea and amount to no more than a general link of the use of the abstract idea to a particular technological environment.
Thus, dependent claims 2, 3, 5, 15, 19, and 20 are ineligible.
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.
Claim(s) 1, 3-6, 10-13, 15, 16, 18, and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vasishta et al. (US 10,157,411 B1 [previously cited]) in view of Ruan et al. (US 2025/0005644 A1).
Regarding Claim 1, Vasishta et al., hereinafter, Vasishta, discloses a system, comprising: at least one processor; and a non-transitory memory having instructions stored thereon that, when executed by the at least one processor, cause the at least one processor to: obtain a request to select a set of personalized information items from a collection of information items to a user (Figs. 1 and 4; Col. 4, lines 41-47[when a user is viewing a “source” item, the recommendation service 30 can select a “target” item based on the source item, the RFM-to-item association mappings 28, and the user's RFM score.] in view of Col. 9, lines 9-12[users' item viewing histories, item download histories, item ratings, shopping cart adds, and/or other types of item selection events] and Abstract [The recommendation system can select items to recommend to particular users based at least partly on the detected behavior-based correlations.] in view of Col. 9, lines 29-40[All of the methods and tasks described above may be performed and fully automated by a computer system… computing device typically includes a hardware processor (or multiple hardware processors) that includes digital logic circuitry for executing program instructions or modules stored in a memory or other computer-readable storage medium.]; Examiner notes that “RFM” is an acronym for “Recency, Frequency, and Monetary amount,” see Abstract of Vasishta);
classify the user in a first user class of a plurality of user classes, wherein the first user class has a predefined relationship with a first object class of a plurality of object classes (Figs. 2 and 3; Col. 5, lines 10-15[segment the users into one of 125 segments. The 125 segments correspond to combinations of RFM scores according to the scoring scheme described above. For example, 111 represents one segment while 411 represents another segment] in view of Col. 6, lines 49-53[The purchase history of all the users in the selected segment may share some common characteristics because the users in the selected segment also share a common characteristic (e.g. same RFM score).] in view of Abstract[The scores can be RFM (Recency, Frequency, and Monetary amount) scores based on at least purchase or selection recency, purchase or selection frequency, and in some instances, the monetary amount spent]);
identify, among a plurality of users, a subset of users that are classified in the first user class (Fig. 3-4; Col. 5, lines 34-37[the correlation detector 22 can cycle through all the possible RFM scores, identify all the users with that RFM score]);
identify, in the collection of information items, a subset of information items that are classified in the first object class (Fig. 4; Col. 6, line 57 to Col. 7, line 3[The correlation detector engine 22 can identify if a particular segment purchases name brand (higher) quality items or generic items more frequently than the general user population. The correlation detector engine 22 can also identify the mean and variance of the price distribution of items purchased for a particular segment. This can indicate price sensitivity and the likelihood of the particular segment purchasing an item worth beyond a certain amount. Using these correlations, the system may attempt to move a particular user up in the segment chain by recommending items that may be at the higher end of price range for that segment, but still within the tolerance. If done gradually, the user may be able to move up in the segmentation hierarchy]);
identify the one or more personalized information items utilizing a collaborative filtering, by: obtaining an input including engagement information of the subset of users and the subset of information items (Fig. 4; Col. 6, line 57 to Col. 7, line 3[The correlation detector engine 22 can identify if a particular segment purchases name brand (higher) quality items or generic items more frequently than the general user population. The correlation detector engine 22 can also identify the mean and variance of the price distribution of items purchased for a particular segment. This can indicate price sensitivity and the likelihood of the particular segment purchasing an item worth beyond a certain amount. Using these correlations, the system may attempt to move a particular user up in the segment chain by recommending items that may be at the higher end of price range for that segment, but still within the tolerance. If done gradually, the user may be able to move up in the segmentation hierarchy.]);
based on the input, determining the collaborative filtering by: determining a user of each of the subset of users (Fig. 3; Col. 6, lines 22-26[analyzing the purchase behavior for each user in a segment at block 306 may be used to create behavioral associations between users with particular scores and the items they are likely to be interested in purchasing from the electronic catalog]);
determining a similarity level between the user and each of the subset of users based on the user of the subset of users (Fig. 3; Col. 6, lines 15-25[In block 304, the users can be divided into multiple segments based on their calculated scores… the users can be divided into one of 125 segments. The users in each segment may share a particular characteristic] in view of Col. 5, lines 52-57[users with an RFM score of 551 very rarely purchase jewelry items in comparison to users in other segments; (2) users with RFM scores of 44× (where “x” represents “don't care”) purchase audio tracks and audio electronics items significantly more frequently than users in other segments]); and
applying the similarity levels of the subset of users with the user as weights to determine, for each of the subset of information items, object numbers based on a weighted of respective object numbers of the set of related users (Fig. 3; Col. 6, lines 6-40[In block 302, the RFM score generator 24 generates scores for all or a subset of the users of an electronic catalog (e.g. shopping site)… In block 304, the users can be divided into multiple segments based on their calculated scores… For example, the correlation detector engine 22 can select a particular segment (e.g. all the users with a score of 515)]); and
selecting the one or more personalized information items of the user based on object number of each of the subset of information items (Figs. 3 and 4; Col. 8, line 66 to Col. 9, line 3[If item D does not fall in the user's RFM segment, item D may be filtered out so that it is not recommended; or its ranking in the recommendation set may be decreased to reduce its likelihood of being recommended] in view of Col. 4, lines 63-67[The recommendation engine 30 can select a list of items to recommend based on the RFM-to-item association 28 and the narrow down the recommendations based on items-to-items association or other filters.]); and
in response to identifying the one or more personalized information items, generate instructions configured to display an interface including the one or more personalized information items on a client device associated with the user (Fig. 4[element 410]; Col. 7, lines 50-58[In block 408, the recommendation service 30 can subsequently filter or re-rank the set of candidate recommendations based on the correlations corresponding to the user's RFM score. For example, using negative correlations, the recommendation service 30 can remove from the set of candidate recommendations that the user is not likely to purchase. The recommendation engine can transmit the recommendations to a user device at block 410 without further processing.] in view of Claim 1[ matching the current user's generated score with one of the plurality of segments; selecting, by a recommendation service, catalog items to recommend to the current user based at least partly on the detected segment characterizing correlations between the matched segment and particular catalog items and/or catalog item attributes; and causing the selected catalog items to be presented to current user via a user interface]).
Although Vasishta discloses identifying items by obtaining an input of engagement information, Vasishta does not explicitly disclose identifying items within a time limit utilizing a filtering model.
Although Vasishta discloses filtering by determining a user of a subset of users, Vasishta does not explicitly disclose a filtering model by determining a user embedding and user embeddings.
Although Vasishta discloses determining an accumulated object number based on a weighted score, Vasishta does not explicitly disclose an accumulated number based on a sum of numbers of users and selecting items based on the accumulated number.
However, Ruan et al., hereinafter, Ruan, discloses identifying items within a time period utilizing a model, user embeddings, and a weighted a sum of a number of users (¶0064[To do so, the machine learning training module 230 may first compute a user embedding by combining the short-term and long-term embedding for each user. For example, a user embedding may be computed as a weighted sum of a long-term embedding for a user and a short-term embedding for the user] in view of ¶0036[Session data for a session of a customer may identify the session … a time frame associated with the session. Session data also may describe a customer's interactions with the online system 140, such as information describing items with which the customer interacted, item data for the items, information describing types of the interactions (e.g., browsing, selecting, or searching for items, adding items to a shopping list, etc.), times of the interactions, dwell time, etc.]).
The system of Ruan is applicable to the system of Vasishta as they share characteristics and capabilities, namely, they are both targeted to recommending products to users. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the filtering of items to recommend to users as disclosed by Vasishta to include user embedding and identifying items within a time period as taught by Ruan. One of ordinary skill in the art would have been motivated to expand the system of Vasishta in order to identify candidate items for presentation to users (Abstract).
Regarding Claim 3, Vasishta in view of Ruan teaches the system of claim 1, Vasishta discloses wherein the non-transitory memory further includes instructions to: extract, from an object database, each of the subset of information items, the object database storing the collection of information items (Fig. 1; Col. 3, lines 40-53[As illustrated in FIG. 1, the system includes one or more physical servers 10 that provide user access to an electronic catalog of items represented in a database or collection of databases 12. The catalog items may include or consist of items that can be purchased via the system (e.g., consumer electronics products, household appliances, apparel items, sporting goods, software products, digital content items, etc.). The database 12 may also store data regarding how the items are arranged within categories or browse nodes of a hierarchical browse structure, as is known in the art. Data regarding the catalog items and the browse structure is accessible via a catalog service 14 that provides access to descriptions of the various items available for purchase.]); and
apply an object classifier to classify each of the subset of information items to the first object class among the plurality of object classes (Fig. 3[element 310]; Col. 7, lines 19-31[The correlation detector engine 22 can repeat the above process to identify the correlations for each of the segments. In block 310, the correlation detector engine 22 can store the correlations for each segment in the RFM-to-item mapping 28. In an embodiment, the RFM-to-item mapping data repository 28 can include a list of scores and all the items associated with the respective scores. The RFM-to-item mapping data repository 28 can be updated once a month or at any other appropriate time interval. In other embodiments, the RFM-to-item mapping data repository 28 is updated in real time when there is a need to recommend items. The RFM-to-item mapping data repository 28 can also selectively be updated.]).
Regarding Claim 4, Vasishta in view of Ruan teaches the system of claim 3, Vasishta discloses wherein each of the subset of information items includes one or more of: title, short description, long description, product type taxonomy, shelf taxonomy, and item attributes of a respective object (Fig. 3; Col. 7, lines 10-18[In some embodiments, the correlations may additionally or alternatively be detected based on user behaviors other than purchases. For example, the correlations between RFM scores and items (or item attributes) could additionally or alternatively be based on shopping cart add events, item rating events, wish list add events, item review submissions, and other types of user-generated events that reflect user preferences for particular items or item types.]).
Regarding Claim 5, Vasishta in view of Ruan teaches the system of claim 1, Vasishta discloses wherein the non-transitory memory further includes instructions to: extract, from a user database, user information of each of the subset of users, the user database storing user information of the plurality of users (Fig. 1; Col. 3, line 59 to Col. 4, line 11[As users browse and make purchases of items represented in the electronic catalog, the system records one or more types of item selection events in a data repository 20, which may include multiple distinct log files and/or databases. The events may include, for example, item purchase events, item viewing events (which may be based on visits to item detail pages), shopping cart add events, wish list add events, and other types of user-directed events that evidence users' interests in particular catalog items… One or more types of recorded events or event histories are analyzed by a correlation detection service 22 to detect behavioral relationships]); and
apply a user classifier to classify each of the subset of users to the first user class based on the user information of the respective user (Fig. 3; Col. 5, lines 5-15[FIG. 2 illustrates example purchase histories for three different users, their corresponding RFM scores and how the correlation engine 22 can generate an RFM-to-item mapping 28 based on the RFM scores. In the illustrated example, the RFM score generator 24 used the RFM subdivision scheme described above to segment the users into one of 125 segments. The 125 segments correspond to combinations of RFM scores according to the scoring scheme described above. For example, 111 represents one segment while 411 represents another segment.]).
Regarding Claim 6, Vasishta in view of Ruan teaches the system of claim 5, Vasishta discloses wherein the user classifier includes a user engagement extraction and a classification, and the instructions to apply the user classifier further comprise instructions to (Fig. 2; Col. 5, lines 10-15), for each of the subset of users:
obtain the user information of the respective user including one or more of: a purchase history, a search history, and a browsing history of the respective user (Fig. 2; Col. 5, lines 30-35[The correlation detection engine 22 can use the RFM scores and the purchase histories to create an RFM-to-item mapping 28.]);
apply the user engagement extraction to generate a user engagement feature based on the user information of the respective user (Fig. 3; Col. 5, lines 10-15[segment the users into one of 125 segments. The 125 segments correspond to combinations of RFM scores according to the scoring scheme described above. For example, 111 represents one segment while 411 represents another segment] in view of Col. 6, lines 49-53[The purchase history of all the users in the selected segment may share some common characteristics because the users in the selected segment also share a common characteristic (e.g. same RFM score).] in view of Abstract[The scores can be RFM (Recency, Frequency, and Monetary amount) scores based on at least purchase or selection recency, purchase or selection frequency, and in some instances, the monetary amount spent]); and
apply the classification to determine the first user class based on the user engagement feature (Figs. 2-4; Col. 5, lines 10-15[segment the users into one of 125 segments. The 125 segments correspond to combinations of RFM scores according to the scoring scheme described above. For example, 111 represents one segment while 411 represents another segment] in view of Col. 6, lines 49-53[The purchase history of all the users in the selected segment may share some common characteristics because the users in the selected segment also share a common characteristic (e.g. same RFM score).] in view of Abstract[The scores can be RFM (Recency, Frequency, and Monetary amount) scores based on at least purchase or selection recency, purchase or selection frequency, and in some instances, the monetary amount spent]).
Although Vasishta discloses user engagement extraction and classifications, Vasishta does not explicitly teach an extraction model and a classification model and applying the extraction model and the classification model to determine features.
However, Ruan discloses identifying items within a time period utilizing a model, user embeddings, and a weighted a sum of a number of users (¶0064[To do so, the machine learning training module 230 may first compute a user embedding by combining the short-term and long-term embedding for each user. For example, a user embedding may be computed as a weighted sum of a long-term embedding for a user and a short-term embedding for the user] in view of ¶0036[Session data for a session of a customer may identify the session … a time frame associated with the session. Session data also may describe a customer's interactions with the online system 140, such as information describing items with which the customer interacted, item data for the items, information describing types of the interactions (e.g., browsing, selecting, or searching for items, adding items to a shopping list, etc.), times of the interactions, dwell time, etc.]).
The system of Ruan is applicable to the system of Vasishta as they share characteristics and capabilities, namely, they are both targeted to recommending products to users. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the filtering of items to recommend to users as disclosed by Vasishta to include user embedding and identifying items within a time period as taught by Ruan. One of ordinary skill in the art would have been motivated to expand the system of Vasishta in order to identify candidate items for presentation to users (Abstract).
Regarding Claim 10, Vasishta discloses a non-transitory computer-readable storage medium, having instructions stored thereon, which when executed by one or more processors cause the one or more processors to (Fig. 1; Col. 9, lines 29-40[All of the methods and tasks described above may be performed and fully automated by a computer system. The computer system may, in some cases, include multiple distinct computers or computing devices (e.g., physical servers, workstations, storage arrays, etc.) that communicate and interoperate over a network to perform the described functions. Each such computing device typically includes a hardware processor (or multiple hardware processors) that includes digital logic circuitry for executing program instructions or modules stored in a memory or other computer-readable storage medium.]):
obtain a request to select a set of personalized information items from a collection of information items to a user (Fig. 4; Col. 4, lines 41-47[when a user is viewing a “source” item, the recommendation service 30 can select a “target” item based on the source item, the RFM-to-item association mappings 28, and the user's RFM score.] in view of Col. 9, lines 9-12[users' item viewing histories, item download histories, item ratings, shopping cart adds, and/or other types of item selection events] and Abstract [The recommendation system can select items to recommend to particular users based at least partly on the detected behavior-based correlations.]);
classify the user in a first user class of a plurality of user classes, wherein the first user class has a predefined relationship with a first object class of a plurality of object classes (Fig.3; Col. 5, lines 10-15[segment the users into one of 125 segments. The 125 segments correspond to combinations of RFM scores according to the scoring scheme described above. For example, 111 represents one segment while 411 represents another segment] in view of Col. 6, lines 49-53[The purchase history of all the users in the selected segment may share some common characteristics because the users in the selected segment also share a common characteristic (e.g. same RFM score).] in view of Abstract[The scores can be RFM (Recency, Frequency, and Monetary amount) scores based on at least purchase or selection recency, purchase or selection frequency, and in some instances, the monetary amount spent]);
identify, among a plurality of users, a subset of users that are classified in the first user class (Fig. 3-4; Col. 5, lines 34-37[The correlation detector 22 can identify the purchasing behavior of each segment. In an embodiment, the correlation detector 22 can cycle through all the possible RFM scores, identify all the users with that RFM score]);
identify, in the collection of information items, a subset of information items that are classified in the first object class (Fig. 4; Col. 6, line 57 to Col. 7, line 3[The correlation detector engine 22 can identify if a particular segment purchases name brand (higher) quality items or generic items more frequently than the general user population. The correlation detector engine 22 can also identify the mean and variance of the price distribution of items purchased for a particular segment. This can indicate price sensitivity and the likelihood of the particular segment purchasing an item worth beyond a certain amount. Using these correlations, the system may attempt to move a particular user up in the segment chain by recommending items that may be at the higher end of price range for that segment, but still within the tolerance. If done gradually, the user may be able to move up in the segmentation hierarchy]);
identify the one or more personalized information items utilizing a collaborative filtering, by: obtaining an input including engagement information of the subset of users and the subset of information items (Fig. 4; Col. 6, line 57 to Col. 7, line 3[The correlation detector engine 22 can identify if a particular segment purchases name brand (higher) quality items or generic items more frequently than the general user population. The correlation detector engine 22 can also identify the mean and variance of the price distribution of items purchased for a particular segment. This can indicate price sensitivity and the likelihood of the particular segment purchasing an item worth beyond a certain amount. Using these correlations, the system may attempt to move a particular user up in the segment chain by recommending items that may be at the higher end of price range for that segment, but still within the tolerance. If done gradually, the user may be able to move up in the segmentation hierarchy.]);
based on the input, determining the collaborative filtering by: determining a user of each of the subset of users (Fig. 3; Col. 6, lines 22-26[analyzing the purchase behavior for each user in a segment at block 306 may be used to create behavioral associations between users with particular scores and the items they are likely to be interested in purchasing from the electronic catalog]);
determining a similarity level between the user and each of the subset of users based on the user of the subset of users (Fig. 3; Col. 6, lines 15-25[In block 304, the users can be divided into multiple segments based on their calculated scores… the users can be divided into one of 125 segments. The users in each segment may share a particular characteristic] in view of Col. 5, lines 52-57[users with an RFM score of 551 very rarely purchase jewelry items in comparison to users in other segments; (2) users with RFM scores of 44× (where “x” represents “don't care”) purchase audio tracks and audio electronics items significantly more frequently than users in other segments]); and
applying the similarity levels of the subset of users with the user as weights to determine, for each of the subset of information items, object numbers based on a weighted of respective object numbers of the set of related users (Fig. 3; Col. 6, lines 6-40[In block 302, the RFM score generator 24 generates scores for all or a subset of the users of an electronic catalog (e.g. shopping site)… In block 304, the users can be divided into multiple segments based on their calculated scores… For example, the correlation detector engine 22 can select a particular segment (e.g. all the users with a score of 515)]); and
selecting the one or more personalized information items of the user based on object number of each of the subset of information items (Figs. 3 and 4; Col. 8, line 66 to Col. 9, line 3[If item D does not fall in the user's RFM segment, item D may be filtered out so that it is not recommended; or its ranking in the recommendation set may be decreased to reduce its likelihood of being recommended] in view of Col. 4, lines 63-67[The recommendation engine 30 can select a list of items to recommend based on the RFM-to-item association 28 and the narrow down the recommendations based on items-to-items association or other filters.]); and
in response to identifying the one or more personalized information items, generate instructions configured to display an interface including the one or more personalized information items on a client device associated with the user (Fig. 4[element 410]; Col. 7, lines 50-58[In block 408, the recommendation service 30 can subsequently filter or re-rank the set of candidate recommendations based on the correlations corresponding to the user's RFM score. For example, using negative correlations, the recommendation service 30 can remove from the set of candidate recommendations that the user is not likely to purchase. The recommendation engine can transmit the recommendations to a user device at block 410 without further processing.] in view of Claim 1[ matching the current user's generated score with one of the plurality of segments; selecting, by a recommendation service, catalog items to recommend to the current user based at least partly on the detected segment characterizing correlations between the matched segment and particular catalog items and/or catalog item attributes; and causing the selected catalog items to be presented to current user via a user interface]).
Although Vasishta discloses identifying items by obtaining an input of engagement information, Vasishta does not explicitly disclose identifying items within a time limit utilizing a filtering model.
Although Vasishta discloses filtering by determining a user of a subset of users, Vasishta does not explicitly disclose a filtering model by determining a user embedding and user embeddings.
Although Vasishta discloses determining an accumulated object number based on a weighted score, Vasishta does not explicitly disclose an accumulated number based on a sum of numbers of users and selecting items based on the accumulated number.
However, Ruan discloses identifying items within a time period utilizing a model, user embeddings, and a weighted a sum of a number of users (¶0064[To do so, the machine learning training module 230 may first compute a user embedding by combining the short-term and long-term embedding for each user. For example, a user embedding may be computed as a weighted sum of a long-term embedding for a user and a short-term embedding for the user] in view of ¶0036[Session data for a session of a customer may identify the session … a time frame associated with the session. Session data also may describe a customer's interactions with the online system 140, such as information describing items with which the customer interacted, item data for the items, information describing types of the interactions (e.g., browsing, selecting, or searching for items, adding items to a shopping list, etc.), times of the interactions, dwell time, etc.]).
The system of Ruan is applicable to the system of Vasishta as they share characteristics and capabilities, namely, they are both targeted to recommending products to users. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the filtering of items to recommend to users as disclosed by Vasishta to include user embedding and identifying items within a time period as taught by Ruan. One of ordinary skill in the art would have been motivated to expand the system of Vasishta in order to identify candidate items for presentation to users (Abstract).
Regarding Claim 11, Vasishta in view of Ruan teaches the non-transitory computer-readable storage medium of claim 10, Vasishta discloses wherein the engagement information of the subset of users includes one or more of: a purchase history, a search history, and a browsing history of each of the subset of users (Fig. 2; Col. 5, lines 30-35[The correlation detection engine 22 can use the RFM scores and the purchase histories to create an RFM-to-item mapping 28.]).
Regarding Claim 12, Vasishta in view of Ruan teaches the non-transitory computer-readable storage medium of claim 11, Vasishta discloses applying the collaborative filtering further comprising: ranking the subset of information items for the user based on the similarity level between the user and each of the subset of users (Fig. 4; Col. 7, lines 45-60[In block 406, the recommendation service 30 can obtain a ranked set of candidate recommendations. The ranked set of candidate recommendations can be generated using any recommendation algorithm (e.g. item-to-item associations). In block 408, the recommendation service 30 can subsequently filter or re-rank the set of candidate recommendations based on the correlations corresponding to the user's RFM score. For example, using negative correlations, the recommendation service 30 can remove from the set of candidate recommendations that the user is not likely to purchase. The recommendation engine can transmit the recommendations to a user device at block 410 without further processing.]).
Although Vasishta discloses applying filtering of content, Vasishta does not explicitly disclose a filtering model.
However, Ruan discloses identifying items within a time period utilizing a model, user embeddings, and a weighted a sum of a number of users (¶0064[To do so, the machine learning training module 230 may first compute a user embedding by combining the short-term and long-term embedding for each user. For example, a user embedding may be computed as a weighted sum of a long-term embedding for a user and a short-term embedding for the user] in view of ¶0036[Session data for a session of a customer may identify the session … a time frame associated with the session. Session data also may describe a customer's interactions with the online system 140, such as information describing items with which the customer interacted, item data for the items, information describing types of the interactions (e.g., browsing, selecting, or searching for items, adding items to a shopping list, etc.), times of the interactions, dwell time, etc.]).
The system of Ruan is applicable to the system of Vasishta as they share characteristics and capabilities, namely, they are both targeted to recommending products to users. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the filtering of items to recommend to users as disclosed by Vasishta to include user embedding and identifying items within a time period as taught by Ruan. One of ordinary skill in the art would have been motivated to expand the system of Vasishta in order to identify candidate items for presentation to users (Abstract).
Regarding Claim 13, Vasishta in view of Ruan teaches the non-transitory computer-readable storage medium of claim 12, Vasishta discloses ranking the subset of information items for the user further comprising:
for each of the subset of information items, wherein the respective information item corresponds to a respective object (Fig. 1; Col. 3, lines 40-50[The catalog items may include or consist of items that can be purchased via the system (e.g., consumer electronics products, household appliances, apparel items, sporting goods, software products, digital content items, etc.).]):
determining that a set of related users has engaged the respective object based on the engagement information (Fig. 4; Col. 4, lines 9-25[The catalog items may include or consist of items that can be purchased via the system (e.g., consumer electronics products, household appliances, apparel items, sporting goods, software products, digital content items, etc.) ... The RFM-to-item association mappings 28 are used by a recommendation service or engine 50 to provide item recommendations to users. Items may be recommended via many different channels. For example, RFM-to-item association mappings 28 may be used to provide personalized recommendations that are based on a particular user's RFM score. The detected correlations can be used to filter, re-rank, or supplement the recommendations generated by any type of recommendation system or algorithm]);
generating the weighted object number for each of the set of related users as a product of a total number of object times engaged by the related user and the similarity level between the user and the related user (Fig. 2; Col. 6, lines 15-25[In block 304, the users can be divided into multiple segments based on their calculated scores. When using RFM scores according to the scoring scales described above (i.e., a scale of 1 to 5 for each of recency, frequency, and monetary), the users can be divided into one of 125 segments.] in view of Col. 3, lines 5-35[In the illustrated embodiments, each RFM score is thus a set of three individual scores (R, F and M) on a scale of 1 to 5. For example, an RFM score of 1-2-3 (which may alternatively be written as “123”) represents the following score set: R=1, F=2, M=3. In other embodiments, different scales or scoring methods may be used. The RFM score can also be combined into a single number. For example, a single number can be obtained by multiplying the recency score, frequency score, and the monetary score by 100, 10, and 1, respectively and adding them all together (e.g. 5*100+1*10+1=511). The RFM scores can reveal important behavior patterns for the users. For instance, the RFM score of 555 may correspond to highly valuable users who shopped recently, have bought regularly, and spend lots of money. The RFM score of 51×, 41× may indicate new users. The RFM score of 135 may indicate users who used to visit the shopping site often and spent big, but haven't shopped for a while. The RFM score of 111, 112 may refer to users who purchased an item a while ago and have not spent much since then. Finally, the RFM score of 155, 255, 145, 245, 154, and 254 may indicate high valued users who have stopped coming back. Furthermore, associating the RFM scores with items, as described more in detail below, can improve the recommendation service.]); and
generating the accumulated object numbers as a sum of the weighted object number of the set of related users (Fig. 2; Col. 6, lines 15-25[In block 304, the users can be divided into multiple segments based on their calculated scores. When using RFM scores according to the scoring scales described above (i.e., a scale of 1 to 5 for each of recency, frequency, and monetary), the users can be divided into one of 125 segments.] in view of Col. 3, lines 5-35[In the illustrated embodiments, each RFM score is thus a set of three individual scores (R, F and M) on a scale of 1 to 5. For example, an RFM score of 1-2-3 (which may alternatively be written as “123”) represents the following score set: R=1, F=2, M=3. In other embodiments, different scales or scoring methods may be used. The RFM score can also be combined into a single number. For example, a single number can be obtained by multiplying the recency score, frequency score, and the monetary score by 100, 10, and 1, respectively and adding them all together (e.g. 5*100+1*10+1=511). The RFM scores can reveal important behavior patterns for the users. For instance, the RFM score of 555 may correspond to highly valuable users who shopped recently, have bought regularly, and spend lots of money. The RFM score of 51×, 41× may indicate new users. The RFM score of 135 may indicate users who used to visit the shopping site often and spent big, but haven't shopped for a while. The RFM score of 111, 112 may refer to users who purchased an item a while ago and have not spent much since then. Finally, the RFM score of 155, 255, 145, 245, 154, and 254 may indicate high valued users who have stopped coming back. Furthermore, associating the RFM scores with items, as described more in detail below, can improve the recommendation service.]); and
determining that the accumulated object numbers of the one or more personalized information items are higher than those of a remainder of the subset of information items for the user (Col. 6, line 65 to Col. 7, line 3[Using these correlations, the system may attempt to move a particular user up in the segment chain by recommending items that may be at the higher end of price range for that segment, but still within the tolerance. If done gradually, the user may be able to move up in the segmentation hierarchy.]).
Regarding Claim 15, Vasishta in view of Ruan teaches the non-transitory computer-readable storage medium of claim 10, Vasishta discloses further comprising instructions to: enable execution of an online shopping application via one of an Internet browser and a dedicated application on the client device associated with the user (Fig. 1; Col. 3, lines 35-50[The system may, for example be accessible via a web site, a specialized mobile application, an interactive television system, and/or other types of systems that provide interactive browse functionality. As illustrated in FIG. 1, the system includes one or more physical servers 10 that provide user access to an electronic catalog of items represented in a database or collection of databases 12. The catalog items may include or consist of items that can be purchased via the system (e.g., consumer electronics products, household appliances, apparel items, sporting goods, software products, digital content items, etc.).]), including:
displaying a user interface of the online shopping application on a screen (Fig. 1; Claim 1[causing the selected catalog items to be presented to current user via a user interface]);
receiving a user action on an actionable information item displayed on the user interface (Fig. 2l Col. 4, lines 15-25[the RFM score for a user can be calculated in real time when the user visits the shopping site or performs a particular type of action. The real time RFM score can be updated in the repository 26. The correlation detection service 22 can use the user specific RFM scores and the purchase histories of users to detect behavioral relationships between the RFM scores and items.]); and
in response to the user action, generating the request to recommend one or more personalized information items to the user (Fig. 1; Col. 4, lines 30-40[The RFM-to-item association mappings 28 are used by a recommendation service or engine 50 to provide item recommendations to users. Items may be recommended via many different channels. For example, RFM-to-item association mappings 28 may be used to provide personalized recommendations that are based on a particular user's RFM score. The detected correlations can be used to filter, re-rank, or supplement the recommendations generated by any type of recommendation system or algorithm.]).
Regarding Claim 16, Vasishta discloses a method, comprising: at a system including a non-transitory memory having instructions stored thereon and at least one processor operatively coupled to the non-transitory memory and configured to execute the instructions: obtaining a request to select a set of personalized information items from a collection of information items to a user (Figs. 1 and 4; Col. 4, lines 41-47[when a user is viewing a “source” item, the recommendation service 30 can select a “target” item based on the source item, the RFM-to-item association mappings 28, and the user's RFM score.] in view of Col. 9, lines 9-12[users' item viewing histories, item download histories, item ratings, shopping cart adds, and/or other types of item selection events] and Abstract [The recommendation system can select items to recommend to particular users based at least partly on the detected behavior-based correlations.] in view of Col. 9, lines 29-40[All of the methods and tasks described above may be performed and fully automated by a computer system… computing device typically includes a hardware processor (or multiple hardware processors) that includes digital logic circuitry for executing program instructions or modules stored in a memory or other computer-readable storage medium.]; Examiner notes that “RFM” is an acronym for “Recency, Frequency, and Monetary amount,” see Abstract of Vasishta);
classifying the user in a first user class of a plurality of user classes, wherein the first user class has a predefined relationship with a first object class of a plurality of object classes (Figs. 2 and 3; Col. 5, lines 10-15[segment the users into one of 125 segments. The 125 segments correspond to combinations of RFM scores according to the scoring scheme described above. For example, 111 represents one segment while 411 represents another segment] in view of Col. 6, lines 49-53[The purchase history of all the users in the selected segment may share some common characteristics because the users in the selected segment also share a common characteristic (e.g. same RFM score).] in view of Abstract[The scores can be RFM (Recency, Frequency, and Monetary amount) scores based on at least purchase or selection recency, purchase or selection frequency, and in some instances, the monetary amount spent]);
identifying, among a plurality of users, a subset of users that are classified in the first user class (Fig. 3-4; Col. 5, lines 34-37[the correlation detector 22 can cycle through all the possible RFM scores, identify all the users with that RFM score]);
identifying, in the collection of information items, a subset of information items that are classified in the first object class (Fig. 4; Col. 6, line 57 to Col. 7, line 3[The correlation detector engine 22 can identify if a particular segment purchases name brand (higher) quality items or generic items more frequently than the general user population. The correlation detector engine 22 can also identify the mean and variance of the price distribution of items purchased for a particular segment. This can indicate price sensitivity and the likelihood of the particular segment purchasing an item worth beyond a certain amount. Using these correlations, the system may attempt to move a particular user up in the segment chain by recommending items that may be at the higher end of price range for that segment, but still within the tolerance. If done gradually, the user may be able to move up in the segmentation hierarchy]);
identifying the one or more personalized information items utilizing a collaborative filtering, by: obtaining an input including engagement information of the subset of users and the subset of information items (Fig. 4; Col. 6, line 57 to Col. 7, line 3[The correlation detector engine 22 can identify if a particular segment purchases name brand (higher) quality items or generic items more frequently than the general user population. The correlation detector engine 22 can also identify the mean and variance of the price distribution of items purchased for a particular segment. This can indicate price sensitivity and the likelihood of the particular segment purchasing an item worth beyond a certain amount. Using these correlations, the system may attempt to move a particular user up in the segment chain by recommending items that may be at the higher end of price range for that segment, but still within the tolerance. If done gradually, the user may be able to move up in the segmentation hierarchy.]);
based on the input, determining the collaborative filtering by: determining a user of each of the subset of users (Fig. 3; Col. 6, lines 22-26[analyzing the purchase behavior for each user in a segment at block 306 may be used to create behavioral associations between users with particular scores and the items they are likely to be interested in purchasing from the electronic catalog]);
determining a similarity level between the user and each of the subset of users based on the user of the subset of users (Fig. 3; Col. 6, lines 15-25[In block 304, the users can be divided into multiple segments based on their calculated scores… the users can be divided into one of 125 segments. The users in each segment may share a particular characteristic] in view of Col. 5, lines 52-57[users with an RFM score of 551 very rarely purchase jewelry items in comparison to users in other segments; (2) users with RFM scores of 44× (where “x” represents “don't care”) purchase audio tracks and audio electronics items significantly more frequently than users in other segments]); and
applying the similarity levels of the subset of users with the user as weights to determine, for each of the subset of information items, object numbers based on a weighted of respective object numbers of the set of related users (Fig. 3; Col. 6, lines 6-40[In block 302, the RFM score generator 24 generates scores for all or a subset of the users of an electronic catalog (e.g. shopping site)… In block 304, the users can be divided into multiple segments based on their calculated scores… For example, the correlation detector engine 22 can select a particular segment (e.g. all the users with a score of 515)]); and
selecting the one or more personalized information items of the user based on object number of each of the subset of information items (Figs. 3 and 4; Col. 8, line 66 to Col. 9, line 3[If item D does not fall in the user's RFM segment, item D may be filtered out so that it is not recommended; or its ranking in the recommendation set may be decreased to reduce its likelihood of being recommended] in view of Col. 4, lines 63-67[The recommendation engine 30 can select a list of items to recommend based on the RFM-to-item association 28 and the narrow down the recommendations based on items-to-items association or other filters.]); and
in response to identifying the one or more personalized information items, generating instructions configured to display an interface including the one or more personalized information items on a client device associated with the user (Fig. 4[element 410]; Col. 7, lines 50-58[In block 408, the recommendation service 30 can subsequently filter or re-rank the set of candidate recommendations based on the correlations corresponding to the user's RFM score. For example, using negative correlations, the recommendation service 30 can remove from the set of candidate recommendations that the user is not likely to purchase. The recommendation engine can transmit the recommendations to a user device at block 410 without further processing.] in view of Claim 1[ matching the current user's generated score with one of the plurality of segments; selecting, by a recommendation service, catalog items to recommend to the current user based at least partly on the detected segment characterizing correlations between the matched segment and particular catalog items and/or catalog item attributes; and causing the selected catalog items to be presented to current user via a user interface]).
Although Vasishta discloses identifying items by obtaining an input of engagement information, Vasishta does not explicitly disclose identifying items within a time limit utilizing a filtering model.
Although Vasishta discloses filtering by determining a user of a subset of users, Vasishta does not explicitly disclose a filtering model by determining a user embedding and user embeddings.
Although Vasishta discloses determining an accumulated object number based on a weighted score, Vasishta does not explicitly disclose an accumulated number based on a sum of numbers of users and selecting items based on the accumulated number.
However, Ruan discloses identifying items within a time period utilizing a model, user embeddings, and a weighted a sum of a number of users (¶0064[To do so, the machine learning training module 230 may first compute a user embedding by combining the short-term and long-term embedding for each user. For example, a user embedding may be computed as a weighted sum of a long-term embedding for a user and a short-term embedding for the user] in view of ¶0036[Session data for a session of a customer may identify the session … a time frame associated with the session. Session data also may describe a customer's interactions with the online system 140, such as information describing items with which the customer interacted, item data for the items, information describing types of the interactions (e.g., browsing, selecting, or searching for items, adding items to a shopping list, etc.), times of the interactions, dwell time, etc.]).
The method of Ruan is applicable to the method of Vasishta as they share characteristics and capabilities, namely, they are both targeted to recommending products to users. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the filtering of items to recommend to users as disclosed by Vasishta to include user embedding and identifying items within a time period as taught by Ruan. One of ordinary skill in the art would have been motivated to expand the method of Vasishta in order to identify candidate items for presentation to users (Abstract).
Regarding Claim 18, Vasishta in view of Ruan teaches the method of claim 16, Vasishta discloses wherein each of the collection of information items is classified to one or more object classes (Fig. 4; Col. 3, lines 40-55[The catalog items may include or consist of items that can be purchased via the system (e.g., consumer electronics products, household appliances, apparel items, sporting goods, software products, digital content items, etc.). The database 12 may also store data regarding how the items are arranged within categories or browse nodes of a hierarchical browse structure, as is known in the art. Data regarding the catalog items and the browse structure is accessible via a catalog service 14 that provides access to descriptions of the various items available for purchase] in view of Col. 6, line 57 to Col. 7, line 3[The correlation detector engine 22 can identify if a particular segment purchases name brand (higher) quality items or generic items more frequently than the general user population. The correlation detector engine 22 can also identify the mean and variance of the price distribution of items purchased for a particular segment. This can indicate price sensitivity and the likelihood of the particular segment purchasing an item worth beyond a certain amount. Using these correlations, the system may attempt to move a particular user up in the segment chain by recommending items that may be at the higher end of price range for that segment, but still within the tolerance. If done gradually, the user may be able to move up in the segmentation hierarchy]), and
each of the plurality of users is classified to one or more user classes (Fig. 3-4; Col. 5, lines 34-37[The correlation detector 22 can identify the purchasing behavior of each segment. In an embodiment, the correlation detector 22 can cycle through all the possible RFM scores, identify all the users with that RFM score]).
Regarding Claim 20, Vasishta in view of Ruan teaches the method of claim 16, Vasishta discloses further comprising: extracting, from an object database, each of the subset of information items, the object database storing the collection of information items (Fig. 1; Col. 3, lines 40-53[As illustrated in FIG. 1, the system includes one or more physical servers 10 that provide user access to an electronic catalog of items represented in a database or collection of databases 12. The catalog items may include or consist of items that can be purchased via the system (e.g., consumer electronics products, household appliances, apparel items, sporting goods, software products, digital content items, etc.). The database 12 may also store data regarding how the items are arranged within categories or browse nodes of a hierarchical browse structure, as is known in the art. Data regarding the catalog items and the browse structure is accessible via a catalog service 14 that provides access to descriptions of the various items available for purchase.]); and
applying an object classifier to classify each of the subset of information items to the first object class among the plurality of object classes (Fig. 3[element 310]; Col. 7, lines 19-31[The correlation detector engine 22 can repeat the above process to identify the correlations for each of the segments. In block 310, the correlation detector engine 22 can store the correlations for each segment in the RFM-to-item mapping 28. In an embodiment, the RFM-to-item mapping data repository 28 can include a list of scores and all the items associated with the respective scores. The RFM-to-item mapping data repository 28 can be updated once a month or at any other appropriate time interval. In other embodiments, the RFM-to-item mapping data repository 28 is updated in real time when there is a need to recommend items. The RFM-to-item mapping data repository 28 can also selectively be updated.]).
Claim(s) 2 and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vasishta in view of Ruan in view of Kimelfeld et al. (US 2014/0351020 A1 [previously cited]).
Regarding Claim 2, Vasishta in view of Ruan teaches the system of claim 1, Vasishta discloses further comprising: an object database storing the collection of information items (Col. 3, lines 40-55[The catalog items may include or consist of items that can be purchased via the system (e.g., consumer electronics products, household appliances, apparel items, sporting goods, software products, digital content items, etc.). The database 12 may also store data regarding how the items are arranged within categories or browse nodes of a hierarchical browse structure, as is known in the art. Data regarding the catalog items and the browse structure is accessible via a catalog service 14 that provides access to descriptions of the various items available for purchase.]),
wherein the collection of information items includes information items (Col. 3, lines 40-55[The catalog items may include or consist of items that can be purchased via the system (e.g., consumer electronics products, household appliances, apparel items, sporting goods, software products, digital content items, etc.). The database 12 may also store data regarding how the items are arranged within categories or browse nodes of a hierarchical browse structure, as is known in the art. Data regarding the catalog items and the browse structure is accessible via a catalog service 14 that provides access to descriptions of the various items available for purchase.]), and
wherein the subset of users has a first number (Ni) of users (Col. 2, lines 63-67[In some instances, a significant number of users of the site may make only a single purchase. In that case, it may be better to assign a score of 1 to all the users with one purchase and then divide the remaining frequencies evenly among the four segments.]), and
the subset of information items has a second number (N2) of information items (Col. 3, lines 40-45[As illustrated in FIG. 1, the system includes one or more physical servers 10 that provide user access to an electronic catalog of items represented in a database or collection of databases 12]),
where each of the first number (Ni) and the second number (N2) (Col. 2, lines 63-67[In some instances, a significant number of users of the site may make only a single purchase. In that case, it may be better to assign a score of 1 to all the users with one purchase and then divide the remaining frequencies evenly among the four segments.] in view of Col. 3, lines 40-45[As illustrated in FIG. 1, the system includes one or more physical servers 10 that provide user access to an electronic catalog of items represented in a database or collection of databases 12]).
Although Vasishta discloses storing items and users, Vasishta in view of Ruan does not explicitly teach more than 10 million items and wherein the number of items is greater than a respective threshold number NTH.
However, Kimelfeld et al., hereinafter, Kimelfeld, teaches more than 10 million items and a threshold number (¶0034[For this example, K=2, and N is the total number of items, such that N is equal to 100 million.] in view of ¶0060[determining when the random hash function number is greater than a threshold, the threshold is a function of a total number of the items; when the random hash function number is greater than the threshold, not adding the particular item value pair to the particular value range which results in the particular item value pair being randomly discarded; when the random hash function number is less than the threshold, adding the particular item value pair to the particular value range]).
The system of Kimelfeld is applicable to the system of Vasishta in view of Ruan as they share characteristics and capabilities, namely, they are all targeted to classifying commercial interactions online. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the user and item classifications as taught by Vasishta in view of Ruan to include more than 10 million items and a threshold as taught by Kimelfeld. One of ordinary skill in the art would have been motivated to expand the system of Vasishta in view of Ruan in order to extract information from a data set and transform it into an understandable structure for further use (¶0002).
Regarding Claim 19, Vasishta in view of Ruan teaches the method of claim 16, Vasishta discloses further comprising: storing in an object database the collection of information items (Col. 3, lines 40-55[The catalog items may include or consist of items that can be purchased via the system (e.g., consumer electronics products, household appliances, apparel items, sporting goods, software products, digital content items, etc.). The database 12 may also store data regarding how the items are arranged within categories or browse nodes of a hierarchical browse structure, as is known in the art. Data regarding the catalog items and the browse structure is accessible via a catalog service 14 that provides access to descriptions of the various items available for purchase.]),
wherein the collection of information items includes information items (Col. 3, lines 40-55[The catalog items may include or consist of items that can be purchased via the system (e.g., consumer electronics products, household appliances, apparel items, sporting goods, software products, digital content items, etc.). The database 12 may also store data regarding how the items are arranged within categories or browse nodes of a hierarchical browse structure, as is known in the art. Data regarding the catalog items and the browse structure is accessible via a catalog service 14 that provides access to descriptions of the various items available for purchase.]), and
wherein the subset of users has a first number (Ni) of users (Col. 2, lines 63-67[In some instances, a significant number of users of the site may make only a single purchase. In that case, it may be better to assign a score of 1 to all the users with one purchase and then divide the remaining frequencies evenly among the four segments.]), and
the subset of information items has a second number (N2) of information items (Col. 3, lines 40-45[As illustrated in FIG. 1, the system includes one or more physical servers 10 that provide user access to an electronic catalog of items represented in a database or collection of databases 12]),
where each of the first number (Ni) and the second number (N2) (Col. 2, lines 63-67[In some instances, a significant number of users of the site may make only a single purchase. In that case, it may be better to assign a score of 1 to all the users with one purchase and then divide the remaining frequencies evenly among the four segments.] in view of Col. 3, lines 40-45[As illustrated in FIG. 1, the system includes one or more physical servers 10 that provide user access to an electronic catalog of items represented in a database or collection of databases 12]).
Although Vasishta discloses storing items and users, Vasishta in view of Ruan does not explicitly teach more than 10 million items and wherein the number of items is greater than a respective threshold number NTH.
However, Kimelfeld teaches more than 10 million items and a threshold number (¶0034[For this example, K=2, and N is the total number of items, such that N is equal to 100 million.] in view of ¶0060[determining when the random hash function number is greater than a threshold, the threshold is a function of a total number of the items; when the random hash function number is greater than the threshold, not adding the particular item value pair to the particular value range which results in the particular item value pair being randomly discarded; when the random hash function number is less than the threshold, adding the particular item value pair to the particular value range]).
The method of Kimelfeld is applicable to the method of Vasishta in view of Ruan as they share characteristics and capabilities, namely, they are all targeted to classifying commercial interactions online. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the user and item classifications as taught by Vasishta in view of Ruan to include more than 10 million items and a threshold as taught by Kimelfeld. One of ordinary skill in the art would have been motivated to expand the method of Vasishta in view of Ruan in order to extract information from a data set and transform it into an understandable structure for further use (¶0002).
Claim(s) 7-9 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vasishta in view of Ruan in view of Garib Herrera et al. (US 11,758,210 B1 [previously cited]).
Regarding Claim 7, Vasishta in view of Ruan teaches the system of claim 5, Vasishta discloses wherein the instructions to apply the user classifier further comprise instructions to, for each of the subset of users:
generate for the first user class (Fig. 3; Col. 5, lines 10-15[segment the users into one of 125 segments. The 125 segments correspond to combinations of RFM scores according to the scoring scheme described above. For example, 111 represents one segment while 411 represents another segment] in view of Col. 6, lines 49-53[The purchase history of all the users in the selected segment may share some common characteristics because the users in the selected segment also share a common characteristic (e.g. same RFM score).] in view of Abstract[The scores can be RFM (Recency, Frequency, and Monetary amount) scores based on at least purchase or selection recency, purchase or selection frequency, and in some instances, the monetary amount spent]); and
in accordance with a determination that the first user class is user class, classify the respective user to the first user class (Fig. 3; Col. 5, lines 10-15[segment the users into one of 125 segments. The 125 segments correspond to combinations of RFM scores according to the scoring scheme described above. For example, 111 represents one segment while 411 represents another segment] in view of Col. 6, lines 49-53[The purchase history of all the users in the selected segment may share some common characteristics because the users in the selected segment also share a common characteristic (e.g. same RFM score).] in view of Abstract[The scores can be RFM (Recency, Frequency, and Monetary amount) scores based on at least purchase or selection recency, purchase or selection frequency, and in some instances, the monetary amount spent]).
Although Vasishta discloses generating a user class, Vasishta in view of Ruan does not explicitly teach generating a probability and that the probability for an entity is greater than a first threshold.
However, Garib Herrera et al., hereinafter, Herrera, teaches a probability being greater than a threshold (Fig. 3; Col. 13, lines 32-45[Still referring to FIG. 3, a set of rules 310 may govern the selection of one or more audience groups to represent the user 302 (e.g., and the audience in which user 302 is included) based on the likelihood that the user 302 belongs to one or more of the audience groups. When a probability that the user 302 is part of an audience group (e.g., as determined using the Equations above) exceeds a probability threshold, the rules 310 may require that a particular audience group always be selected, that a particular audience group be selected when its probability is above a threshold percentile of all audience group probabilities, that a particular audience group be selected based on the probability exceeding a threshold set as a result of another audience group being selected, and the like.]).
The system of Herrera is applicable to the system of Vasishta in view of Ruan as they share characteristics and capabilities, namely, they are all targeted to classifying commercial interactions online. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the user and item classifications as taught by Vasishta in view of Ruan to include a probability being greater than a threshold as taught by Herrera. One of ordinary skill in the art would have been motivated to expand the system of Vasishta in view of Ruan in order to customize content presentation (Col. 1, lines 11-12).
Regarding Claim 8, Vasishta in view of Ruan in view of Herrera teaches the system of claim 7, Vasishta discloses wherein the instructions to apply the user classifier further comprise instructions to, for each of the subset of users:
generate for a second user class (Figs. 2-4; Col. 5, lines 45-65[the correlation detector 22 searches for statistically significant correlations that differentiate a particular segment from most or all other segments; these “segment characterizing” correlations tend to be the most useful for generating item recommendations. The following are hypothetical examples: (1) users with an RFM score of 551 very rarely purchase jewelry items in comparison to users in other segments; (2) users with RFM scores of 44× (where “x” represents “don't care”) purchase audio tracks and audio electronics items significantly more frequently than users in other segments. The first of these correlations may be used to filter out, or to reduce the probability of, recommendations of jewelry items to users having RFM scores of 551. The second may be used to boost, or increase the likelihood of, recommendations of audio tracks and electronic items to users with RFM scores of 44×.] in view of Col. 5, lines 27-31[Similarly, the RFM score generator 24 can calculate RFM scores for the second user 204 and the third user 206 and store their scores in the user specific RFM data repository 26.]); and
in accordance with a determination that the second user class is user class, classify the respective user to the second user class, the respective user belonging to the first user class the second user class (Col. 5, lines 5-11[FIG. 2 illustrates example purchase histories for three different users, their corresponding RFM scores and how the correlation engine 22 can generate an RFM-to-item mapping 28 based on the RFM scores. In the illustrated example, the RFM score generator 24 used the RFM subdivision scheme described above to segment the users into one of 125 segments.]).
Although Vasishta discloses generating a user class, Vasishta in view of Ruan does not explicitly teach generating a probability and that the probability for an entity is greater than a second threshold and a user belonging to both a first and second user class.
However, Herrera teaches a probability being greater than a threshold and a user belonging to two classes (Fig. 3; Col. 7, lines 28-29[The audience segments may overlap, so one user may belong to many segments.] in view of Col. 13, lines 32-45[Still referring to FIG. 3, a set of rules 310 may govern the selection of one or more audience groups to represent the user 302 (e.g., and the audience in which user 302 is included) based on the likelihood that the user 302 belongs to one or more of the audience groups. When a probability that the user 302 is part of an audience group (e.g., as determined using the Equations above) exceeds a probability threshold, the rules 310 may require that a particular audience group always be selected, that a particular audience group be selected when its probability is above a threshold percentile of all audience group probabilities, that a particular audience group be selected based on the probability exceeding a threshold set as a result of another audience group being selected, and the like.]).
The system of Herrera is applicable to the system of Vasishta in view of Ruan as they share characteristics and capabilities, namely, they are all targeted to classifying commercial interactions online. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the user and item classifications as taught by Vasishta in view of Ruan to include a probability being greater than a threshold and a user belonging to two classes as taught by Herrera. One of ordinary skill in the art would have been motivated to expand the system of Vasishta in view of Ruan in order to customize content presentation (Col. 1, lines 11-12).
Regarding Claim 9, Vasishta in view of Ruan in view of Herrera teaches the system of claim 7, Vasishta discloses wherein the instructions to apply the user classifier further comprise instructions to, for each of the subset of users:
generate for a second user class (Figs. 2-4; Col. 5, lines 45-65[the correlation detector 22 searches for statistically significant correlations that differentiate a particular segment from most or all other segments; these “segment characterizing” correlations tend to be the most useful for generating item recommendations. The following are hypothetical examples: (1) users with an RFM score of 551 very rarely purchase jewelry items in comparison to users in other segments; (2) users with RFM scores of 44× (where “x” represents “don't care”) purchase audio tracks and audio electronics items significantly more frequently than users in other segments. The first of these correlations may be used to filter out, or to reduce the probability of, recommendations of jewelry items to users having RFM scores of 551. The second may be used to boost, or increase the likelihood of, recommendations of audio tracks and electronic items to users with RFM scores of 44×.] in view of Col. 5, lines 27-31[Similarly, the RFM score generator 24 can calculate RFM scores for the second user 204 and the third user 206 and store their scores in the user specific RFM data repository 26.]); and
in accordance with a determination that the second user class is user class, determine that the respective user does not belong to the second user class (Figs. 2 and 4; Col. 8, lines 55-65[Low-utility recommendations can be inhibited in various ways. One is to filter out, from a dataset of catalog items selected using a criteria (e.g. browsing history) all the items that do not match the user's RFM score. The filtered set of items may then be used to provide recommendations to users] in view of Col. 5, lines 45-65).
Although Vasishta discloses generating a user class, Vasishta in view of Ruan does not explicitly teach generating a probability and that the probability for an entity is not greater than a second threshold.
However, Herrera teaches a probability that is not greater than a threshold (Fig. 3; Col. 7, lines 55-60[The result may be a probability, and the system may evaluate whether the probability is high enough (e.g., exceeds a threshold) to classify users as being within an audience segment.]).
The system of Herrera is applicable to the system of Vasishta in view of Ruan as they share characteristics and capabilities, namely, they are all targeted to classifying commercial interactions online. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the user and item classifications as taught by Vasishta in view of Ruan to include a probability that is not greater than a threshold as taught by Herrera. One of ordinary skill in the art would have been motivated to expand the system of Vasishta in view of Ruan in order to customize content presentation (Col. 1, lines 11-12).
Claim(s) 14 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vasishta in view of Ruan in view of Lind et al. (US 2004/0267596 A1 [previously cited]).
Regarding Claim 14, Vasishta in view of Ruan teaches the non-transitory computer-readable storage medium of claim 10, Vasishta discloses wherein: in accordance with the predefined relationship, the first user class and the first object class are associated with a type of a plurality of types (Fig. 2; Col. 5, lines 5-11[FIG. 2 illustrates example purchase histories for three different users, their corresponding RFM scores and how the correlation engine 22 can generate an RFM-to-item mapping 28 based on the RFM scores. In the illustrated example, the RFM score generator 24 used the RFM subdivision scheme described above to segment the users into one of 125 segments.]);
for each of the subset of users classified to the first user class, type is a first user class (Fig. 3; Col. 6, lines 15-25[In block 304, the users can be divided into multiple segments based on their calculated scores. When using RFM scores according to the scoring scales described above (i.e., a scale of 1 to 5 for each of recency, frequency, and monetary), the users can be divided into one of 125 segments. The users in each segment may share a particular characteristic.]); and
for each of the subset of information items classified to the first object class, being bought for a of the type is a first object class (Fig. 3; Col. 6, lines 15-25[In block 304, the users can be divided into multiple segments based on their calculated scores. When using RFM scores according to the scoring scales described above (i.e., a scale of 1 to 5 for each of recency, frequency, and monetary), the users can be divided into one of 125 segments. The users in each segment may share a particular characteristic.]).
Although Vasishta discloses predefined relationships, Vasishta in view of Ruan does not explicitly teach a pet type of pet types, a respective probability of being an owner of a pet of the pet type is greater than a threshold and a respective probability of a pet of the pet type is greater than a threshold.
However, Lind et al., hereinafter, Lind, teaches a probability of being a pet owner and a threshold (¶0006[For example, consider an online buyer who previously bought a dog bowl and a dog bone. Probabilities can be determined that show that it is likely that this person owns a dog. The person might, therefore, be interested in dog related items such as dog collars, leashes, and brushes. Since these items are brought to the attention of the buyer, if it matches their needs, they are more likely to purchase those items than, for instance, an advertisement for catnip or a bird feeder.] in view of ¶0068[This value is a minimum acceptable count for incorporating effects from that particular association. Anything below this threshold is not utilized. If Weight of Evidence is desired 806, a minimum acceptable value or threshold is determined based on counts of parameters]).
The system of Lind is applicable to the system of Vasishta in view of Ruan as they share characteristics and capabilities, namely, they are all targeted to classifying commercial interactions online. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the user and item classifications as taught by Vasishta in view of Ruan to include a probability of being a pet owner and a threshold as taught by Lind. One of ordinary skill in the art would have been motivated to expand the system of Vasishta in view of Ruan in order to provide quickly creatable models that can efficiently predict probability that a user likes an item and/or similarities between items (Abstract).
Claim(s) 17 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vasishta in view of Ruan in view of Putcha et al. (US 2017/0220986 A1 [previously cited]).
Regarding Claim 17, Vasishta in view of Ruan teaches the method of claim 16, Vasishta discloses further comprising: in accordance with a determination that the user belongs to the first user class, enabling a display of a page to the user associated with the predefined relationship (Claim 1[matching the current user's generated score with one of the plurality of segments; selecting, by a recommendation service, catalog items to recommend to the current user based at least partly on the detected segment characterizing correlations between the matched segment and particular catalog items and/or catalog item attributes; and causing the selected catalog items to be presented to current user via a user interface, thereby enabling the current user to more efficiently locate and navigate to representations of catalog items of interest]),
wherein actionable information items representing the one or more personalized information items are displayed on the page or concurrently with the page (Claim 1[matching the current user's generated score with one of the plurality of segments; selecting, by a recommendation service, catalog items to recommend to the current user based at least partly on the detected segment characterizing correlations between the matched segment and particular catalog items and/or catalog item attributes; and causing the selected catalog items to be presented to current user via a user interface, thereby enabling the current user to more efficiently locate and navigate to representations of catalog items of interest]).
Although Vasishta discloses displaying a page to a user, Vasishta in view of Ruan does not explicitly teach a signup page to invite a user to create a profile and displaying the signup page with the signup page.
However, Putcha et al., hereinafter, Putcha, teaches a signup page to invite a user to create a profile (Fig. 1; ¶0043[In some embodiments, when a consumer 110 initially sets up an online account with the retail provider, the system 100 (i.e., order processing server 130 or another server on the system 100 dedicated to new consumer sign-up) is configured to permit the consumer 110 to generate a consumer profile including personal information of the consumer 110 (e.g., name, address, phone number, and the like), payment methods (e.g., credit card information), as well as substitute product settings of the consumer 110 (e.g., what products may or may not be offered to the consumer 110 if the product ordered by the consumer from the retail provider is out of stock at the retail sales facility 170). It will be appreciated that the consumer 110 may select new options and/or update previously selected options at any time after setting up the account with the retail provider. In some embodiments, the personal information of the consumer 110 and any options selected by the consumer 110 are stored in the consumer information database 140 for subsequent retrieval by the order processing server 130 (e.g., in response to a login request or product order placement by the consumer 110)]).
The method of Putcha is applicable to the method of Vasishta in view of Ruan as they share characteristics and capabilities, namely, they are all targeted to classifying commercial interactions online. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the user and item classifications as taught by Vasishta in view of Ruan to include sign up pages to create a user profile as taught by Putcha. One of ordinary skill in the art would have been motivated to expand the method of Vasishta in view of Ruan in order to purchase one or more substitute products when the products ordered by the consumers are out of stock (¶0002).
Response to Arguments
Applicant’s arguments on pages 10 and 11 of the remarks filed 07/06/2026, with respect to the previous 35 USC § 101 rejections have been fully considered but are not persuasive.
Applicant argues on pages 10 and 11 of the remarks that the amended claims integrate the abstract idea into a practical application by providing a technical improvement. Examiner respectfully disagrees. According to the MPEP 2106.04, the question of whether a claim is “directed to” a judicial exception in Step 2A is now evaluated using a two-prong inquiry. Prong One asks if the claim “recites” an abstract idea, law of nature, or natural phenomenon. Under that prong, the mere inclusion of a judicial exception such as a method of organizing human activity in a claim means that the claim “recites” a judicial exception (see MPEP 2106.04 [“The mere inclusion of a judicial exception such as a mathematical formula (which is one of the mathematical concepts identified as an abstract idea in MPEP § 2106.04(a)) in a claim means that the claim "recites" a judicial exception under Step 2A Prong One.”]). Additionally, MPEP 2106.04 instructs examiners to refer to the groupings of abstract ideas enumerated in MPEP 2106.04(a)(2) (i.e., mathematical concepts, certain methods of organizing human activities, and mental processes) in order to identify abstract ideas. As noted above and in the previous office action, the claims recite product recommendation based on engagement information. This is further illustrated in ¶0003 of the instant specification which discusses that the invention improves customer engagement with products. This is an abstract idea because it is a concept of business relations which makes it a method of organizing human activity (i.e., one of the groupings of abstract ideas enumerated in MPEP 2106.04(a)(2)).
With respect to the applicant’s argument that the amended claim integrates the abstract idea into a practical application, the MPEP sets forth, in Step 2A Prong Two, that a claim that recites a judicial exception is not directed to that judicial exception, if the claim as a whole "integrates the recited judicial exception into a practical application of that exception." The evaluation of Prong Two requires the use of the considerations (e.g. improving technology, effecting a particular treatment or prophylaxis, implementing with a particular machine, etc.) identified by the Supreme Court and the Federal Circuit, to ensure that the claim as a whole integrates [the] judicial exception into a practical application [that] will apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception, such that the claim is more than a drafting effort designed to monopolize the judicial exception.' In the instant case, the claims include additional elements such as at least one processor; and a non-transitory memory having instructions stored thereon that, when executed by the at least one processor, cause the at least one processor to: obtain, items on a client device associated with the user. While these elements are recited, they are merely peripherally incorporated in order to implement the abstract idea. Put another way, these additional elements are merely used to apply the abstract idea of providing product recommendation based on engagement information in a technological environment without effectuating any improvement or changes to the functioning of the additional elements or other technology. Applicant's disclosure does not articulate or suggest how these additional elements function, individually or in combination, in any manner other than using generic functionality nor does the disclosure articulate how the elements provide a technical improvement. Accordingly, the additional elements do not integrate the abstract idea into a practical application because they merely amount to using the computing components as a tool to perform the abstract idea.
Accordingly, Examiner maintains that the invention is directed to a judicial exception without
significantly more. The claims recite an abstract idea. This judicial exception is not integrated into a practical application. The claim(s) do not include additional elements that are sufficient to amount to significantly more than the judicial exception. Thus the 35 USC §101 rejections are maintained.
Applicant’s arguments on pages 11-14 of the remarks filed 07/06/2026, with respect to the previous 35 USC § 103 rejections have been fully considered but are mostly moot in view of the new 103 rejection of the amended claims.
Applicant argues that the references cited do not teach the newly amended limitation, “selecting the one or more personalized information items of the user based on the accumulated object number of each of the subset of information items.” Examiner respectfully disagrees. Vasishta discloses a recommendation engine that selects personalized items based on an RFM (Recency, Frequency, and Monetary amount) number corresponding to a user subset, see Vasishta Figs. 3 and 4 and Col. 8, line 66 to Col. 9, line 3. Examiner notes that selecting items based on an RFM number corresponding to a user subset is comparable to selecting one or more personalized information items of the user based on object number of each of the subset of information items. Vasishta does not explicitly disclose selecting based on the accumulated object number. However, the newly cited reference Ruan teaches selections based on an accumulated number, see ¶0064 of Ruan. Accordingly, Vasishta in view of Ruan teaches “selecting the one or more personalized information items of the user based on the accumulated object number of each of the subset of information items.”
Applicant argues that the references cited do not teach the newly amended limitation, “determining the collaborative filtering model by: determining a user embedding of each of the subset of users; determining a similarity level between the user and each of the subset of users based on the user embeddings of the subset of users; and applying the similarity levels of the subset of users with the user as weights to determine, for each of the subset of information items, an accumulated object numbers based on a weighted sum of respective object numbers of the set of related users.” Examiner respectfully disagrees. Vasishta discloses analyzing purchase behavior for each user in a segment or subset in order to create behavioral association between uses with particular scores and the items they are likely to be interested in purchasing, see Col. 6, lines 22-26 of Vasishta. Examiner notes that analyzing purchase behavior of users to create associations between users and items is comparable to determining a user of each subset of users.
Vasishta further discloses dividing users into segments or subsets based on calculated scores for the users, see Col. 6, lines 15-25 of Vasishta. Examiner notes that dividing users into subsets based on calculated scores for the users is comparable to determining a similarity level between the user and each of the subset of users based on the user of the subset of users.
Vasishta further discloses an “RFM score generator” that generates scores for users in order to divide the users into subsets based on their calculated scores and further selecting subsets based on their subset score, see Col. 6, lines 6-40. Examiner notes that dividing users into subsets based on similarity of their generated RFM score is comparable to applying the similarity levels of the subset of users with the user as weights to determine, for each of the subset of information items, object numbers based on a weighted of respective object numbers of the set of related users. Vasishta does not explicitly disclose a filtering model, determining a user embedding, determining a similarity based on user embeddings, or an accumulated object number based on a weighted sum. However, the newly cited references Ruan teaches identifying items utilizing a model, user embeddings, and a weighted sum of a number of users, see ¶0036 and ¶0064 of Ruan. Accordingly, Vasishta in view of Ruan teaches “determining the collaborative filtering model by: determining a user embedding of each of the subset of users; determining a similarity level between the user and each of the subset of users based on the user embeddings of the subset of users; and applying the similarity levels of the subset of users with the user as weights to determine, for each of the subset of information items, an accumulated object numbers based on a weighted sum of respective object numbers of the set of related users.”
Accordingly, references Vasishta, Kimelfeld, Herrera, Lind, and Putcha have been maintained, and reference Ruan has been added in view of the claim amendments.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any extension fee pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to AHOORA LADONI whose email is Ahoora.Ladoni@uspto.gov and telephone number is (703) 756-5617. The examiner can normally be reached M-F 0900–1700 ET.
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.
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/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.
/AHOORA LADONI/Examiner, Art Unit 3689
/MARISSA THEIN/Supervisory Patent Examiner, Art Unit 3689
Prosecution Insights