DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Continued Examination Under 37 CFR 1.114
2. A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 3/30/2026 has been entered.
Accordingly, claims 1-3, 5-13, and 15-22 are pending in this application. Claims 1, 7-11, and 17-22 are currently amended.
Response to Arguments
3. Applicant’s arguments with respect to amended pending claims filed on 3/30/2026 have been fully considered. In view of the claim amendment filed, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection is made.
Further, regarding the new limitations recited in claims 1, 7-11, and 17-22, it is submitted that they are properly addressed by the new ground of rejection.
Furthermore, it is also submitted that all limitations in pending claims, including those not specifically argued, are properly addressed. The reason is set forth in the rejections. See claim analysis below for detail.
Claim Rejections - 35 USC § 103
4. 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 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.
5. 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.
6. Claims 1-3, 5-13, and 15-22 are rejected under 35 U.S.C. 103 as being unpatentable over Volkovs et al. (US 20230267367 A1) in view of Grappin et al. (US 20250045281 A1).
Regarding Claim 1, Volkovs discloses a computer-implemented method to recommend content items ([Abstract]: A recommendation system generates item recommendations for a user), the method comprising:
identifying content items for recommendation to a user, wherein identifying the(Fig. 1; [0019]: The recommendation system 130 identifies relevant items that users are likely to be interested in or will interact with and suggests the identified items to users)
obtaining a prior content item embedding for a prior content item associated with the user (Fig. 1, Figs. 5A-5C; [0019]-[0020]: For example, a recommendation system 130 included in a video streaming server may identify and suggest movies that a user may like based on movies that the user has previously viewed… the recommendation system 130 evaluates the likelihood of a user liking the item based on embeddings in a latent space); and
selecting, prior to assigning respective ranks, a first set of the content items, wherein the first set of content items are associated with first respective content item embeddings that are within a threshold distance of the prior content item embedding (Figs. 5A-5C; [0020]-[0028]: In general, the distance (i.e., a proximity measure) a between a user and an item in the latent space is used to determine the likelihood a user may be interested in the item… After scoring content items, a number of highest-scoring content items may be selected and recommended to the user… the recommendation module 200 may apply such a recommendation model with the embeddings to determine recommendation scores for a particular user).
However, Volkovs does not explicitly teach “assigning the respective ranks to individual content items in the
On the other hand, in the same field of endeavor, Grappin teaches
assigning the respective ranks to individual content items in the([0232]-[0235]: The server 102 is configured to input into the second model 406 during the given training iteration the vector 604 representative of a predicted digital item of the first type on which the given user is likely to perform a rare event, the user vector 606, and item vectors associated with the plurality of digital items to be ranked… perform ranking of digital items in the plurality of digital items in a listwise manner);
selecting, based on the respective ranks, one or more candidate content items from the([0243]: In response to the inputs, the second model 406 is configured to generate the ranked list 420… the server 102 may select one or more items from the ranked list 420 for inclusion into the recommended set of items; [0261]: The method 700 continues to step 712 with the server configured to select a set of content items from the ranked list of content items as the content for the user 20); and
providing the selected one or more candidate content items to a client device associated with the user for display in a user interface (Fig. 7; [0263]: The method 700 continues to step 714 with the server 102 configured to transmit the content for display to the user 20 in the application running on the electronic device 106).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teaching of Volkovs to incorporate the teachings of Grappin to include assigning the respective ranks to individual content items personalized to the user, selecting candidate content items, and providing the selected items to a client device for display in a user interface.
The motivation for doing so would be to increase the likelihood of a user interacting with a recommended item, as recognized by Grappin ([Abstract] of Grappin: A rank of an item is indicative of a likelihood of that the user interacts with the given item of the second type if the user performs an event of the pre-determined type on the given predicted item of the first type).
Regarding Claim 2, the combined teachings of Volkovs and Grappin disclose the computer-implemented method of claim 1.
Grappin further teaches wherein identifying the content items further includes:
obtaining user feature embeddings based on user features of the user ([0140]: Additionally, or alternatively, the server 102 may generate one or more user vectors by employ a variety of embedding algorithms; Fig. 4; [0206]: In some embodiments of the present technology, the first model 404 may be embodied as a matrix factorization model configured to receive embeddings for users and content items);
generating a user embedding based on the user feature embeddings using a first trained deep neural network (DNN) (Fig. 4; [0227]: In other embodiments, the second model 406 may be implemented as a Deep Neural Network (DNN)… For example, a two-tower NN can be executed by the server 102 in which the model learns a non-linear function that maps features to an embedding); and
selecting a second set of content items that are associated with second respective content item embeddings that are within a threshold distance of the user embedding, wherein assigning the respective ranks includes assigning respective ranks to individual content items in the second set of content items ([0232]-[0235]: The server 102 is configured to input into the second model 406 during the given training iteration the vector 604 representative of a predicted digital item of the first type on which the given user is likely to perform a rare event, the user vector 606, and item vectors associated with the plurality of digital items to be ranked… perform ranking of digital items in the plurality of digital items in a listwise manner) and
selecting the one or more candidate items includes selecting from the second set of content items ([0243]: In response to the inputs, the second model 406 is configured to generate the ranked list 420… the server 102 may select one or more items from the ranked list 420 for inclusion into the recommended set of items; [0261]: The method 700 continues to step 712 with the server configured to select a set of content items from the ranked list of content items as the content for the user 20).
Regarding Claim 3, the combined teachings of Volkovs and Grappin disclose the computer-implemented method of claim 2.
Grappin further teaches wherein the first trained DNN is from a first tower of a two tower model that includes a second trained DNN from a second tower (Fig. 1; [0227]: For example, a two-tower NN can be executed by the server 102 in which the model learns a non-linear function that maps features to an embedding), and
wherein the first trained DNN and the second trained DNN are trained to output user embeddings and content item embeddings (Fig. 1; [0227]: For example, a DNN model may use a softmax function that treats the problem as a multi-class prediction where… the model learns a non-linear function that maps features to an embedding) that are close in vector space for user-content item pairs that have a groundtruth association and that are separated in vector space for user-content item pairs that do not have the groundtruth association ([Figs. 1, 4; [0214]: It should be noted that the server 102 is configured to use the penalty score 510 in order to adjust the first model 404 such that the first model 404 in a sense “learns” from the given training set and generates predicted outputs that are closer to the ground-truth).
Regarding Claim 5, the combined teachings of Volkovs and Grappin disclose the computer-implemented method of claim 1.
Grappin further teaches wherein identifying the content items includes:
obtaining user feature embeddings based on user features of the user ([0140]: Additionally, or alternatively, the server 102 may generate one or more user vectors by employ a variety of embedding algorithms; Fig. 4; [0206]: In some embodiments of the present technology, the first model 404 may be embodied as a matrix factorization model configured to receive embeddings for users and content items);
generating a user embedding based on the user feature embeddings using a first trained deep neural network (DNN) (Fig. 4; [0227]: In other embodiments, the second model 406 may be implemented as a Deep Neural Network (DNN)… For example, a two-tower NN can be executed by the server 102 in which the model learns a non-linear function that maps features to an embedding);
selecting a second set of content items that includes content items that are associated with respective content item embeddings that are within a threshold distance of the user embedding (Figs. 1, 6; [0228]-[0230]: The server 102 is configured to access item data associated with a plurality of digital items. For example, the server 102 may be configured to retrieve and/or generate inter alia item vectors 608, 610 and 612 for respective digital items… For example, the first model 404 may generate a given output vector and identify a given item vector associated with a digital item of the first type that is closest to the given output vector; See also paras [0207]-[0214]);
obtaining a prior content item embedding for a prior content item associated with the user ([0228]: In some embodiments, it is contemplated that the training set 602 may include item vectors for digital items with which the given user previous interacted on the recommendation platform 112);
selecting a second set of content items that includes content items that are associated with respective content item embeddings that are within a threshold distance of the prior content item embedding ([0228]: In other embodiments, it is contemplated that the training set 602 may include item vectors for digital items of the second type); and
merging the first set of content items and the second set of content items, wherein the merging is performed prior to selecting the one or more candidate items ([0229]: The user vector 606 may be generated based on user data associated with the given user and previously stored in the database 110).
Regarding Claim 6, the combined teachings of Volkovs and Grappin disclose the computer-implemented method of claim 1.
Grappin further teaches wherein assigning the respective ranks is based on one or more of user interests of the user, content item inventory associated with the candidate content items, play history of the user, purchase history of the user, or recommendation context ([0097]: The ranks of respective digital items from this list may be determined based on inter alia user interests of the user 20).
Regarding Claim 7, the combined teachings of Volkovs and Grappin disclose the computer-implemented method of claim 2.
Grappin further teaches wherein the candidate content items are virtual experiences that include one or more developer items ([0158]: As an example, a digital item may comprise textual and/or visual information to be displayed on the recommendation application 107, such as a post, a comment, a picture, a gallery of pictures, a video, a product and/or a service, or any other form of digital item suitable for being recommended by the recommendation platform 112), and wherein a respective content item embedding for a virtual experience is a learned embedding based on respective developer item embeddings of the one or more developer items ([0199]-[0200]: During this training phase, the given NN adapts to the situation being learned and changes its structure such that the given NN will be able to provide reasonable predicted outputs for given inputs in a new situation (based on what was learned)… including filtering, clustering, vector embedding, and the like).
Regarding Claim 8, the combined teachings of Volkovs and Grappin disclose the computer-implemented method of claim .
Grappin further teaches wherein the candidate content items are virtual experiences that include a plurality of assets that include one or more of audio assets, visual assets, or text assets (Fig. 1; [0158]: As an example, a digital item may comprise textual and/or visual information to be displayed on the recommendation application 107, such as a post, a comment, a picture, a gallery of pictures, a video, a product and/or a service, or any other form of digital item suitable for being recommended by the recommendation platform 112), and wherein a respective content item embedding for a virtual experience is an asset embedding based on the plurality of assets associated with the virtual experience ([0156]: The one or more SP vectors may be generated by the server 102 employing one or more machine learning algorithms and/or one or more embedding techniques as described above; [Nonfunctional descriptive material describing the content items]).
Regarding Claim 9, the combined teachings of Volkovs and Grappin disclose the computer-implemented method of claim 2.
Grappin further teaches wherein the candidate content items are virtual experiences that include a plurality of assets and one or more developer items (Fig. 1; [0158]: As an example, a digital item may comprise textual and/or visual information to be displayed on the recommendation application 107, such as a post, a comment, a picture, a gallery of pictures, a video, a product and/or a service, or any other form of digital item suitable for being recommended by the recommendation platform 112), and wherein a respective content item embedding for a virtual experience is a concatenation of an asset embedding based on the plurality of assets associated with the virtual experience and a learned embedding based on respective developer item embeddings of the one or more developer items ([0156]: The one or more SP vectors may be generated by the server 102 employing one or more machine learning algorithms and/or one or more embedding techniques as described above; [Nonfunctional descriptive material describing the content items]).
Regarding Claim 10, the combined teachings of Volkovs and Grappin disclose the computer-implemented method of claim 2.
Grappin further teaches wherein the candidate content items are one or more content items for purchase ([0019]: The frequent events may also represent “purchases” of the first digital item by a large number of users), and wherein the content item embedding for a content item for purchase is a respective item feature embedding of the one or more content items ([0156]: The one or more SP vectors may be generated by the server 102 employing one or more machine learning algorithms and/or one or more embedding techniques as described above; [Nonfunctional descriptive material describing the content items]) .
Regarding Claim 11, Volkovs discloses a non-transitory computer-readable medium with instructions stored thereon that, when executed by one or more hardware processors, cause the one or more hardware processors to perform operations ([0075]-[0076]: In one embodiment, a software module is implemented with a computer program product comprising a computer-readable medium containing computer program code, which can be executed by a computer processor … Such a computer program may be stored in a non-transitory, tangible computer readable storage medium), the operations comprising:
identifying content items for recommendation to a user, wherein identifying the(Fig. 1; [0019]: The recommendation system 130 identifies relevant items that users are likely to be interested in or will interact with and suggests the identified items to users)
obtaining a prior content item embedding for a prior content item associated with the user (Fig. 1, Figs. 5A-5C; [0019]-[0020]: For example, a recommendation system 130 included in a video streaming server may identify and suggest movies that a user may like based on movies that the user has previously viewed… the recommendation system 130 evaluates the likelihood of a user liking the item based on embeddings in a latent space); and
selecting, prior to assigning respective ranks, a first set of the content items, wherein the first set of content items are associated with first respective content item embeddings that are within a threshold distance of the prior content item embedding (Figs. 5A-5C; [0020]-[0028]: In general, the distance (i.e., a proximity measure) a between a user and an item in the latent space is used to determine the likelihood a user may be interested in the item… After scoring content items, a number of highest-scoring content items may be selected and recommended to the user… the recommendation module 200 may apply such a recommendation model with the embeddings to determine recommendation scores for a particular user).
However, Volkovs does not explicitly teach “assigning the respective ranks to individual content items in the
On the other hand, in the same field of endeavor, Grappin teaches
assigning the respective ranks to individual content items in the([0232]-[0235]: The server 102 is configured to input into the second model 406 during the given training iteration the vector 604 representative of a predicted digital item of the first type on which the given user is likely to perform a rare event, the user vector 606, and item vectors associated with the plurality of digital items to be ranked… perform ranking of digital items in the plurality of digital items in a listwise manner);
selecting, based on the respective ranks, one or more candidate content items from the([0243]: In response to the inputs, the second model 406 is configured to generate the ranked list 420… the server 102 may select one or more items from the ranked list 420 for inclusion into the recommended set of items; [0261]: The method 700 continues to step 712 with the server configured to select a set of content items from the ranked list of content items as the content for the user 20); and
providing the selected one or more candidate content items to a client device associated with the user for display in a user interface (Fig. 7; [0263]: The method 700 continues to step 714 with the server 102 configured to transmit the content for display to the user 20 in the application running on the electronic device 106).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teaching of Volkovs to incorporate the teachings of Grappin to include assigning the respective ranks to individual content items personalized to the user, selecting candidate content items, and providing the selected items to a client device for display in a user interface.
The motivation for doing so would be to increase the likelihood of a user interacting with a recommended item, as recognized by Grappin ([Abstract] of Grappin: A rank of an item is indicative of a likelihood of that the user interacts with the given item of the second type if the user performs an event of the pre-determined type on the given predicted item of the first type).
Regarding Claim 12, the combined teachings of Volkovs and Grappin disclose the non-transitory computer-readable medium of claim 11.
Grappin further teaches wherein identifying the content items further comprises:
obtaining user feature embeddings based on user features of the user ([0140]: Additionally, or alternatively, the server 102 may generate one or more user vectors by employ a variety of embedding algorithms; Fig. 4; [0206]: In some embodiments of the present technology, the first model 404 may be embodied as a matrix factorization model configured to receive embeddings for users and content items);
generating a user embedding based on the user feature embeddings using a first trained deep neural network (DNN).
selecting a second set of content items that are associated with second respective content item embeddings that are within a threshold distance of the user embedding ([0243]: In response to the inputs, the second model 406 is configured to generate the ranked list 420… the server 102 may select one or more items from the ranked list 420 for inclusion into the recommended set of items; [0261]: The method 700 continues to step 712 with the server configured to select a set of content items from the ranked list of content items as the content for the user 20),
wherein assigning the respective ranks includes assigning respective ranks to individual content items in the second set of content items and selecting the one or more candidate items includes selecting from the second set of content items ([0232]-[0235]: The server 102 is configured to input into the second model 406 during the given training iteration the vector 604 representative of a predicted digital item of the first type on which the given user is likely to perform a rare event, the user vector 606, and item vectors associated with the plurality of digital items to be ranked… perform ranking of digital items in the plurality of digital items in a listwise manner).
Regarding Claim 13, the combined teachings of Volkovs and Grappin disclose the non-transitory computer-readable medium of claim 12.
Grappin further teaches wherein the first trained DNN is from a first tower of a two tower model that includes a second trained DNN from a second tower (Fig. 1; [0227]: For example, a two-tower NN can be executed by the server 102 in which the model learns a non-linear function that maps features to an embedding), and
wherein the first trained DNN and the second trained DNN are trained to output user embeddings and the second respective content item embeddings (Fig. 1; [0227]: For example, a DNN model may use a softmax function that treats the problem as a multi-class prediction where… the model learns a non-linear function that maps features to an embedding) that are close in vector space for user-content item pairs that have a groundtruth association and that are separated in vector space for user-content item pairs that do not have the groundtruth association ([Figs. 1, 4; [0214]: It should be noted that the server 102 is configured to use the penalty score 510 in order to adjust the first model 404 such that the first model 404 in a sense “learns” from the given training set and generates predicted outputs that are closer to the ground-truth).
Regarding Claim 15, the combined teachings of Volkovs and Grappin disclose the non-transitory computer-readable medium of claim 11.
Grappin further teaches wherein identifying the content items further comprises:
obtaining user feature embeddings based on user features of the user([0140]: Additionally, or alternatively, the server 102 may generate one or more user vectors by employ a variety of embedding algorithms; Fig. 4; [0206]: In some embodiments of the present technology, the first model 404 may be embodied as a matrix factorization model configured to receive embeddings for users and content items);
generating a user embedding based on the user feature embeddings using a first trained deep neural network (DNN) (Fig. 4; [0227]: In other embodiments, the second model 406 may be implemented as a Deep Neural Network (DNN)… For example, a two-tower NN can be executed by the server 102 in which the model learns a non-linear function that maps features to an embedding);
selecting a second set of content items that includes content items that are associated with respective content item embeddings that are within a threshold distance of the user embedding (Figs. 1, 6; [0228]-[0230]: The server 102 is configured to access item data associated with a plurality of digital items. For example, the server 102 may be configured to retrieve and/or generate inter alia item vectors 608, 610 and 612 for respective digital items… For example, the first model 404 may generate a given output vector and identify a given item vector associated with a digital item of the first type that is closest to the given output vector; See also paras [0207]-[0214]); and
merging the first set of content items and the second set of content items, wherein the merging is performed prior to selecting the one or more candidate items ([0229]: The user vector 606 may be generated based on user data associated with the given user and previously stored in the database 110).
Regarding Claim 16, the combined teachings of Volkovs and Grappin disclose the non-transitory computer-readable medium of claim 11.
Grappin further teaches wherein assigning the respective ranks is based on one or more of user interests of the user, content item inventory associated with the candidate content items, play history of the user, purchase history of the user, or recommendation context ([0097]: The ranks of respective digital items from this list may be determined based on inter alia user interests of the user 20).
Regarding Claim 17, the combined teachings of Volkovs and Grappin disclose the non-transitory computer-readable medium of claim 11.
Grappin further teaches wherein the candidate content items are virtual experiences that include at least one of include a plurality of assets or one or more developer items ([0158]: As an example, a digital item may comprise textual and/or visual information to be displayed on the recommendation application 107, such as a post, a comment, a picture, a gallery of pictures, a video, a product and/or a service, or any other form of digital item suitable for being recommended by the recommendation platform 112), and
wherein, when the candidate content items are virtual experiences that include the one or more developer items, the respective content item embedding for a virtual experience is a learned embedding based on respective developer item embeddings of the one or more developer items ([0199]-[0200]: During this training phase, the given NN adapts to the situation being learned and changes its structure such that the given NN will be able to provide reasonable predicted outputs for given inputs in a new situation (based on what was learned)… including filtering, clustering, vector embedding, and the like), or wherein, when the candidate content items are virtual experiences that include the plurality of assets and the one or more developer items, the content item embedding for the virtual experience is a concatenation of an asset embedding based on the plurality of assets associated with the virtual experience and a learned embedding based on respective developer item embeddings of the one or more developer items ([0140]: Additionally, or alternatively, the server 102 may generate one or more user vectors by employ a variety of embedding algorithms).
Regarding Claim 18, the combined teachings of Volkovs and Grappin disclose the non-transitory computer-readable medium of claim 11.
Grappin further teaches wherein the candidate content items are virtual experiences that include a plurality of assets that include one or more of audio assets, visual assets, or text assets (Fig. 1; [0158]: As an example, a digital item may comprise textual and/or visual information to be displayed on the recommendation application 107, such as a post, a comment, a picture, a gallery of pictures, a video, a product and/or a service, or any other form of digital item suitable for being recommended by the recommendation platform 112), and
wherein a respective content item embedding for a virtual experience is an asset embedding based on the plurality of assets associated with the virtual experience ([0156]: The one or more SP vectors may be generated by the server 102 employing one or more machine learning algorithms and/or one or more embedding techniques as described above; [Nonfunctional descriptive material describing the content items]).
Regarding Claim 19, the combined teachings of Volkovs and Grappin disclose the non-transitory computer-readable medium of claim 11.
Grappin further teaches wherein the candidate content items are one or more content items for purchase ([0019]: The frequent events may also represent “purchases” of the first digital item by a large number of users), and wherein a respective content item embedding for a content item for purchase is a respective item feature embedding of the one or more content items ([0156]: The one or more SP vectors may be generated by the server 102 employing one or more machine learning algorithms and/or one or more embedding techniques as described above; [Nonfunctional descriptive material describing the content items]).
Regarding Claim 20, Volkovs discloses a computing device, comprising:
one or more hardware processors; and a non-transitory computer readable medium coupled to the one or more hardware processors, with instructions thereon, that when executed by the one or more hardware processors to perform operations ([0075]-[0076]: In one embodiment, a software module is implemented with a computer program product comprising a computer-readable medium containing computer program code, which can be executed by a computer processor … Such a computer program may be stored in a non-transitory, tangible computer readable storage medium), the operations comprising:
identifying content items for recommendation to a user, wherein identifying the(Fig. 1; [0019]: The recommendation system 130 identifies relevant items that users are likely to be interested in or will interact with and suggests the identified items to users)
obtaining a prior content item embedding for a prior content item associated with the user (Fig. 1, Figs. 5A-5C; [0019]-[0020]: For example, a recommendation system 130 included in a video streaming server may identify and suggest movies that a user may like based on movies that the user has previously viewed… the recommendation system 130 evaluates the likelihood of a user liking the item based on embeddings in a latent space); and
selecting, prior to assigning respective ranks, a first set of the content items, wherein the first set of content items are associated with first respective content item embeddings that are within a threshold distance of the prior content item embedding (Figs. 5A-5C; [0020]-[0028]: In general, the distance (i.e., a proximity measure) a between a user and an item in the latent space is used to determine the likelihood a user may be interested in the item… After scoring content items, a number of highest-scoring content items may be selected and recommended to the user… the recommendation module 200 may apply such a recommendation model with the embeddings to determine recommendation scores for a particular user).
However, Volkovs does not explicitly teach “assigning the respective ranks to individual content items in theset of content items; and providing the selected one or more candidate content items to a client device associated with the user for display in a user interface.”
On the other hand, in the same field of endeavor, Grappin teaches
assigning the respective ranks to individual content items in the([0232]-[0235]: The server 102 is configured to input into the second model 406 during the given training iteration the vector 604 representative of a predicted digital item of the first type on which the given user is likely to perform a rare event, the user vector 606, and item vectors associated with the plurality of digital items to be ranked… perform ranking of digital items in the plurality of digital items in a listwise manner);
selecting, based on the respective ranks, one or more candidate content items from the([0243]: In response to the inputs, the second model 406 is configured to generate the ranked list 420… the server 102 may select one or more items from the ranked list 420 for inclusion into the recommended set of items; [0261]: The method 700 continues to step 712 with the server configured to select a set of content items from the ranked list of content items as the content for the user 20); and
providing the selected one or more candidate content items to a client device associated with the user for display in a user interface (Fig. 7; [0263]: The method 700 continues to step 714 with the server 102 configured to transmit the content for display to the user 20 in the application running on the electronic device 106).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teaching of Volkovs to incorporate the teachings of Grappin to include assigning the respective ranks to individual content items personalized to the user, selecting candidate content items, and providing the selected items to a client device for display in a user interface.
The motivation for doing so would be to increase the likelihood of a user interacting with a recommended item, as recognized by Grappin ([Abstract] of Grappin: A rank of an item is indicative of a likelihood of that the user interacts with the given item of the second type if the user performs an event of the pre-determined type on the given predicted item of the first type).
Regarding Claim 21, the combined teachings of Volkovs and Grappin disclose the computer-implemented method of claim 1.
Volkovs further teaches wherein selecting, prior to assigning the respective ranks, the first set of content items comprises:
for a content item embedding of a plurality of content item embeddings corresponding to the content items, calculating a vector distance between the prior content item embedding and a respective content item embedding ([0072]: In this example, the difference may be due to the Euclidean distance separation between positive and negative pairs);
comparing the vector distance between the prior content item embedding and the respective content item embedding to the threshold distance ([Abstract]: In training, the weight of individual user-item pairs in affecting the user and item embeddings may be determined based on the distance of the particular user-item pair between user embedding and item embedding, as well as the comparative distance for other items of the same type for that user and for the distance of user-item pairs for other users); and
in response to the vector distance being within the threshold distance, selecting the respective content item embedding for inclusion in the first set of content items ([0007]: In some embodiments, the training set of user-item pairs is selected actively by identifying “hard” items for a particular user, such as positive items within a threshold distance of the closest negative item, or the negative items within a threshold distance of the farthest positive item).
Regarding Claim 22, the combined teachings of Volkovs and Grappin disclose the non-transitory computer-readable medium of claim 11.
Volkovs further teaches wherein selecting, prior to assigning the respective ranks, the first set of content items comprises:
for a content item embedding of a plurality of content item embeddings corresponding to the content items, calculating a vector distance between the prior content item embedding and a respective content item embedding ([Abstract]: A recommendation system generates item recommendations for a user based on a distance between a user embedding and item embeddings);
comparing the vector distance between the prior content item embedding and the respective content item embedding to the threshold distance ([0007]-[009]: In some embodiments, the training set of user-item pairs is selected actively by identifying “hard” items for a particular user, such as positive items within a threshold distance of the closest negative item, or the negative items within a threshold distance of the farthest positive item); and
in response to the vector distance being within the threshold distance, selecting the respective content item embedding for inclusion in the first set of content items ([0007]-[009]: Similarly, difficult positive items for a particular user (i.e., user-item pairs) may be selected that are within a threshold of the closest negative item).
Conclusion
27. Any inquiry concerning this communication or earlier communications from the examiner should be directed to SHIRLEY D. HICKS whose telephone number is (571)272-3304. The examiner can normally be reached Mon - Fri 7:30 - 4:00.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Charles Rones can be reached on (571) 272-4085. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/S.D.H./Examiner, Art Unit 2168
/CHARLES RONES/Supervisory Patent Examiner, Art Unit 2168
Prosecution Insights