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 .
This communication is in response to Amendment filed on June 9, 2026. Claims 1-20 are pending. Claims 1, 3, 7-9, 1115, 16, and 19 are amended.
Response to Arguments
Referring to the objection to the title, Applicant’s amendment is acknowledged. As such, the objection to the title is withdrawn.
Referring to the 35 USC 101 rejection of claims 1-20, as amended, Applicant’s arguments have been considered but are not found persuasive.
Applicant argues that the aggregating, generating embedding vectors, encoding the vectors into a single latent embedding space, utilizing a similarity search model to recommend design templates to the user cannot be performed in the human mind. However, Examiner respectfully disagrees. Examiner submits that the aggregating step is a mental step that can be achieved by mentally making a grouping of at least one user event based on a user determined selection from a database of design templates. Furthermore, the specification describes the “user signals” as “individual user events” [para 26]. Furthermore, generating the embedding vectors from the combined user signal are mental steps that can be achieved by mentally generating embedding vectors based on select user events and metadata criteria of the digital design templates. The utilizing of a similarity search model to recommend design templates to the user is also a mental step. The step of identifying digital design templates that satisfy a user determined similarity threshold to recommend can be performed by the human mind through the use of a generic similarity search model. Thus, the claimed limitations can be performed by the human mind.
The encoding of the vectors into a single latent embedding space is an additional element and is insignificant extra-solution activity as selecting and outputting information for display and analysis, as identified in MPEP 2106.05(g) in Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354-55, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016, and does not provide integration into a practical application.
Applicant argues that the claims, as amended, recite improvements to the technology by delimiting the type of data to be displayed and how to display it, and to provide the digital design templates as recommendations across the multiple disparate software applications. However, Examiner respectfully disagrees. The mere display of recommendations of design templates based on a similarity measure determined by a generic similarity search model computing software does not show a technical improvement.
Applicant furthermore argues that the claims recite improvements to a computing technology by using a more efficient computation model- a single transformer neural network comprising self-attention layers- to improve efficiency, accuracy and operational flexibility of identifying digital design templates across computer networks. However, the recitation of utilizing a single transformer neural network comprising self-attention layers is merely applying the mental step of generating embedding vectors through the use of a neural network with the self attention layers and does not integrate the judicial exception into a practical application or teach significantly more.
For at least these reasons, the 35 USC 101 rejections of the pending claims are maintained and further in view of the new grounds of rejection as addressed below.
Applicant’s arguments with respect to claims 1-20 have been considered but are moot in view of the new grounds of rejection.
Information Disclosure Statement
The listing of references in the specification is not a proper information disclosure statement. 37 CFR 1.98(b) requires a list of all patents, publications, or other information submitted for consideration by the Office, and MPEP § 609.04(a) states, "the list may not be incorporated into the specification but must be submitted in a separate paper." Therefore, unless the references have been cited by the examiner on form PTO-892, they have not been considered.
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 an abstract idea without significantly more.
Claim 1 recites:
extracting metadata from digital design templates;
aggregating one or more user events into a combined user signal based on detecting one or more user events of a user with respect to a database of digital design templates;
generating, utilizing a single transformer neural network comprising self-attention layers, both a user embedding vector from the combined user signal and a plurality of embedding vectors for the digital design templates based on the metadata from the digital design templates;
encoding the user embedding vector from the combined user signal and the plurality of embedding vectors from the digital design templates into a single latent embedding space;
identifying, utilizing a similarity search model, a subset of digital design templates from the digital design templates to recommend to the user by comparing the user embedding vector with the plurality of embedding vectors within the single latent embedding space to identify one or more embedding vectors that satisfy a similarity threshold to the user embedding vector.
Step 1: The claim as a whole falls within one or more statutory categories.
Step 2A prong 1: At least claim 1 recites limitations that are abstract ideas.
The limitation “aggregating one or more user events into a combined user signal based on detecting one or more user events of a user with respect to a database of digital design templates” is a mental step that can be achieved by mentally making a grouping of at least one user event based on a user determined selection from a database of design templates. Furthermore, the specification describes the “user signals” as “individual user events” [para 26]. Thus, the claimed limitation can be performed by the human mind.
The limitation “generating both a user embedding vector from the combined user signal and a plurality of embedding vectors for the digital design templates based on the metadata from the digital design templates” are mental steps that can be achieved by mentally generating embedding vectors based on select user events and metadata criteria of the digital design templates. Thus, the claimed limitations can be performed by the human mind.
Furthermore, the limitation “identifying, utilizing a similarity search model, a subset of digital design templates from the digital design templates to recommend to the user by comparing the user embedding vector with the plurality of embedding vectors within the single latent embedding space to identify one or more embedding vectors that satisfy a similarity threshold to the user embedding vector” is also a mental step. The step of identifying digital design templates that satisfy a user determined similarity threshold to recommend can be performed by the human mind.
Step 2A prong 2: Claim 1 recites the limitation ‘extracting metadata from digital design templates from a database of digital design templates’. This limitation is an additional element and is insignificant extra-solution activity as retrieval of select metadata from digital design template data (i.e. mere data gathering) such as 'obtaining information' as identified in MPEP 2106.05(g) and does not provide integration into a practical application.
The limitation “encoding the user embedding vector from the combined user signal and the plurality of embedding vectors from the digital design templates into a single latent embedding space” is also an additional element and is insignificant extra-solution activity as selecting and outputting information for display and analysis, as identified in MPEP 2106.05(g) in Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354-55, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016, and does not provide integration into a practical application.
Furthermore, Claim 1 recites the following additional elements “a single transformer neural network comprising self-attention layers” and “similarity search model”, note that these recited additional elements are a high-level recitation of generic computer software to perform the mental process and applied on a computer as in MPEP 2106.05(f), which does not provide integration into a practical application.
Step 2B: the conclusions for the additional elements representing mere implementation using a computer are carried over and do not provide significantly more.
With respect to the "extracting” and “encoding” limitations identified as insignificant extra-solution activity above, when re-evaluated these elements are well-understood, routine, and conventional as evidenced by the court cases in MPEP 2106.05(d)(II), "i. Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); … OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network);" and thus remains insignificant extra-solution activity that does not provide significantly more.
Therefore, claim 1 as a whole do not change this conclusion and the claim is ineligible.
Claim 16 recites:
extracting metadata from digital design templates from a database of digital design templates;
generating, utilizing a transformer neural network comprising self-attention layers, a plurality of embedding vectors for the digital design templates from the extracted metadata;
accessing a plurality of user events of a user corresponding to a user account, wherein the plurality of user events is from multiple disparate software applications;
generating, utilizing the transformer neural network, a user embedding vector from the plurality of user events of a user from the multiple disparate software applications; and
identifying, utilizing a similarity search model, a subset of digital design templates from the database of digital design templates to recommend to the user account of the user by identifying one or more embedding vectors of the plurality of embedding vectors that satisfy a similarity threshold to the user embedding vector; and
providing the subset of digital design templates as recommendations across the multiple disparate software applications.
Step 1: The claim as a whole falls within one or more statutory categories.
Step 2A prong 1: At least claim 16 recites limitations that are abstract ideas.
The limitation “generating a plurality of embedding vectors for the digital design templates from the extracted metadata” and “generating a user embedding vector from the plurality of user events of a user from the multiple disparate software applications” are mental steps that can be achieved by mentally generating embedding vectors based on select user events and metadata criteria of the digital design templates. Thus, the claimed limitations can be performed by the human mind.
The limitation “accessing a plurality of user events of a user corresponding to a user account, wherein the plurality of user events is from multiple disparate software applications” is a mental step. A user can visually select user events corresponding to a user account from a displayed listing. Thus the claimed limitation can be performed by the human mind.
Furthermore, the limitation “identifying a subset of digital design templates from the database of digital design templates to recommend to the user account of the user by identifying one or more embedding vectors of the plurality of embedding vectors that satisfy a similarity threshold to the user embedding vector” is also a mental step. The step of identifying digital design templates that satisfy a user determined similarity threshold to recommend can be performed by the human mind.
Step 2A prong 2: Claim 16 recites the limitation ‘extracting metadata from digital design templates from a database of digital design templates’. This limitation is an additional element and is insignificant extra-solution activity as retrieval of select metadata from digital design template data (i.e. mere data gathering) such as 'obtaining information' as identified in MPEP 2106.05(g) and does not provide integration into a practical application.
The limitation “providing the subset of digital design templates as recommendations across the multiple disparate software applications” is also an additional element and is insignificant extra-solution activity as selecting and outputting information for display, as identified in MPEP 2106.05(g) in Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354-55, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016, and does not provide integration into a practical application.
Furthermore, Claim 16 recites the following additional elements “at least one processing device”, “a transformer neural network comprising self-attention layers” and “similarity search model”, note that these recited additional elements are a high-level recitation of generic computer components and software to perform the mental process and applied on a computer as in MPEP 2106.05(f), which does not provide integration into a practical application.
Step 2B: the conclusions for the additional elements representing mere implementation using a computer are carried over and do not provide significantly more.
With respect to the "extracting” and ”providing” limitations identified as insignificant extra-solution activity above, when re-evaluated these elements are well-understood, routine, and conventional as evidenced by the court cases in MPEP 2106.05(d)(II), "i. Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); … OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network);" and thus remains insignificant extra-solution activity that does not provide significantly more.
Therefore, claim 16 as a whole do not change this conclusion and the claim is ineligible.
Claim 9 recites:
extracting metadata from digital design templates;
detecting, utilizing a real-time user event aggregation model, one or more user events of a user with respect to the digital design templates;
generating a combined user signal based on the one or more user events and by assigning, utilizing an exponential decay function, weights to the one or more user events to generate one or more weighted user events;
generating, utilizing a single transformer neural network comprising self-attention layers, both a user embedding vector from the combined user signal and a plurality of embedding vectors for the digital design templates based on the metadata from the digital design templates;
encoding the user embedding vector from the combined user signal and the plurality of embedding vectors from the digital design templates into a single latent embedding space; and
identifying, utilizing a similarity search model, a subset of digital design templates from the digital design templates to recommend to the user by comparing the user embedding vector with the plurality of embedding vectors within the single latent embedding space to identify one or more embedding vectors that satisfy a similarity threshold to the user embedding vector.
Step 1: The claim as a whole falls within one or more statutory categories.
Step 2A prong 1: At least claim 9 recites limitations that are abstract ideas.
The limitation “detecting, one or more user events of a user with respect to the digital design templates” is a mental step that can be achieved by mentally detecting the occurrence of one or more user events with respect to the templates. Thus, the claimed limitation can be performed by the human mind.
The limitation “generating a combined user signal based on the one or more user events and by assigning, utilizing an exponential decay function, weights to the one or more user events to generate one or more weighted user events” are mental steps that can be achieved by mentally making a grouping of at least one user event based on a user determined selection from a database of design templates and by mentally assigning a weight to the user events. Furthermore, the specification describes the “user signals” as “individual user events” [para 26]. Thus, the claimed limitations can be performed by the human mind.
The limitations “generating a user embedding vector from the combined user signal and a plurality of embedding vectors for the digital design templates based on the metadata from the digital design templates” are mental steps that can be achieved by mentally generating embedding vectors based on select user events and metadata criteria of the digital design templates. Thus, the claimed limitations can be performed by the human mind.
Furthermore, the limitation “identifying a subset of digital design templates from the digital design templates to recommend to the user by comparing the user embedding vector with the plurality of embedding vectors within the single latent embedding space to identify one or more embedding vectors that satisfy a similarity threshold to the user embedding vector” is also a mental step. The step of identifying digital design templates that satisfy a user determined similarity threshold to recommend from displayed embedding vectors within a space can be performed by the human mind.
Step 2A prong 2: Claim 9 recites the limitation ‘extracting metadata from digital design templates’. This limitation is an additional element and is insignificant extra-solution activity as retrieval of select metadata from digital design template data (i.e. mere data gathering) such as 'obtaining information' as identified in MPEP 2106.05(g) and does not provide integration into a practical application.
The limitation “encoding the user embedding vector from the combined user signal and the plurality of embedding vectors from the digital design templates into a single latent embedding space” is also an additional element and is insignificant extra-solution activity as selecting and outputting information for display and analysis, as identified in MPEP 2106.05(g) in Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354-55, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016, and does not provide integration into a practical application.
Furthermore, Claim 9 recites the following additional elements “a system”, “one or more memory components”, “one or more processing devices”, “single transformer neural network comprising self-attention layers”, “real-time user event aggregation model”, and “similarity search model”, note that these recited additional elements are a high-level recitation of generic computer components and software to perform the mental process and applied on a computer as in MPEP 2106.05(f), which does not provide integration into a practical application.
Step 2B: the conclusions for the additional elements representing mere implementation using a computer are carried over and do not provide significantly more.
With respect to the "extracting” and “encoding” limitations identified as insignificant extra-solution activity above, when re-evaluated these elements are well-understood, routine, and conventional as evidenced by the court cases in MPEP 2106.05(d)(II), "i. Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); … OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network);" and thus remains insignificant extra-solution activity that does not provide significantly more.
Therefore, the claim 9 as a whole does not change this conclusion and the claim is ineligible.
Claims 2, 10 and 17 depend from claims 1, 9 and 16 and thus include all the limitations of claims 1, 9 and 16, therefore claims 2, 10 and 17 recite the same abstract ideas of "mental processes".
Claims 2, 10 and 17 furthermore recite: wherein extracting the metadata further comprises: extracting from the digital design templates at least one of title, text, description, categories, topics, or tasks; and determining stylistic elements of the digital design templates based on at least one of the title, the text, the description, the categories, the topics, or the tasks.
Step 1: Claims 2, 10 and 17 as a whole fall within one or more statutory categories.
Step 2A prong 1: Claims 2, 10 and 17 recite limitations that are abstract ideas.
The limitation “determining stylistic elements of the digital design templates based on at least one of the title, the text, the description, the categories, the topics, or the tasks” is a mental step. One can mentally make a determination of stylistic elements of the template based on a criteria. Thus, the claimed limitation can be performed by the human mind.
Step 2A prong 2: Claims 2, 10 and 17 recite the limitation “wherein extracting the metadata further comprises: extracting from the digital design templates at least one of title, text, description, categories, topics, or tasks”. This extracting step further defines the extracting step in claims 1, 9 and 16 from which the claims depend and was considered as insignificant extra-solution activity as retrieval/receiving of data (i.e. mere data gathering) such as 'obtaining information' as identified in MPEP 2106.05(g) and does not provide integration into a practical application.
Step 2B:
With respect to the "extracting” limitation identified as insignificant extra-solution activity above, when re-evaluated this element is well-understood, routine, and conventional as evidenced by the court cases in MPEP 2106.05(d)(II), "i. Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); … OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network);" and thus remains insignificant extra-solution activity that does not provide significantly more.
Therefore, claims 2, 10 and 17 as a whole do not change this conclusion and the claims are ineligible.
Claim 3 depends from claim 1 and thus include all the limitations of claim 1, therefore claim 3 recites the same abstract ideas of "mental processes".
Claim 3 recites: wherein generating the plurality of embedding vectors further comprises: utilizing the single transformer neural network comprising a sentence transformer to generate the plurality of embedding vectors that represent content within the digital design templates; and storing the plurality of embedding vectors in a database corresponding with the digital design templates.
Step 1: Claim 3 as a whole falls within one or more statutory categories.
Step 2A prong 1: Claim 3 recites limitations that are abstract ideas.
The limitation “generate the plurality of embedding vectors that represent content within the digital design templates” is a mental step that can be achieved by mentally determining and generating embedding vectors based on content criteria. Thus, the claimed limitation can be performed by the human mind.
Step 2A prong 2: Claim 3 recites the limitation “storing the plurality of embedding vectors in a database corresponding with the digital design templates”. This storing step is considered post-solution extra-solution activity and is using of a computer or other machinery in its ordinary capacity for tasks such as storing data and does not integrate a judicial exception into a practical application or provide significantly more.
Furthermore, Claim 3 recites the following additional elements “single transformer neural network comprising a sentence transformer”, note that these recited additional elements are a high-level recitation of generic computer software to perform the mental process and applied on a computer as in MPEP 2106.05(f), which does not provide integration into a practical application.
Step 2B:
the conclusions for the additional elements representing mere implementation using a computer are carried over and do not provide significantly more.
With respect to the “storing” limitation identified as insignificant extra-solution activity above when re-evaluated this element is well-understood, routine, and conventional as evidenced by the court cases in MPEP 2106.05(d)(II), “iv. Storing and retrieving information in memory, Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015); OIP Techs., 788 F.3d at 1363, 115 USPQ2d at 1092-93” and thus remains insignificant extra-solution activity that does not provide significantly more. Therefore, claim 3 as a whole does not change this conclusion and the claim is ineligible.
Claims 4, 12 and 18 depend from claims 1, 9 and 16 and thus include all the limitations of claims 1, 9 and 16, therefore claims 4, 12 and 18 recite the same abstract ideas of "mental processes".
Claims 4, 12 and 18 recite: detecting one or more events of the user in real-time by utilizing a user event stream aggregation model, wherein the one or more events include digital design template remixes, digital design template exports, and search queries performed by the user; and generating the user embedding vector from the detected one or more events.
Step 1: Claims 4, 12 and 18 as a whole fall within one or more statutory categories.
Step 2A prong 1: Claims 4, 12 and 18 recite limitations that are abstract ideas.
The limitation “detecting one or more events of the user in real-time, wherein the one or more events include digital design template remixes, digital design template exports, and search queries performed by the user; and generating the user embedding vector from the detected one or more events” are mental steps that can be achieved by mentally detecting user events and generating embedding vectors based on the user event criteria. Thus, the claimed limitations can be performed by the human mind.
Step 2A prong 2: Claims 4, 12 and 18 recites the following additional elements “a user event stream aggregation model”, note that these recited additional elements are a high-level recitation of generic computer software to perform the mental process and applied on a computer as in MPEP 2106.05(f), which does not provide integration into a practical application.
Step 2B:
The conclusions for the additional elements representing mere implementation using a computer are carried over and do not provide significantly more.
Therefore, claims 4, 12 and 18 as a whole do not change this conclusion and the claims are ineligible.
Claim 5 depends from claim 1 and thus include all the limitations of claim 1, therefore claim 5 recites the same abstract ideas of "mental processes".
Claim 5 recites: generating a weight for each of the one or more user events of the user by utilizing an exponential decay function.
Step 1: Claim 5 as a whole falls within one or more statutory categories.
Step 2A prong 1: Claim 5 recites limitations that are abstract ideas.
The limitation “generating a weight for each of the one or more user events of the user by utilizing an exponential decay function” is a mental step that can be achieved by a user mentally determining a weight for each user event by using the exponential decay function. Thus, the claimed limitation can be performed by the human mind.
Step 2A prong 2: Claim 5 does not recite any additional elements that would integrate the judicial exception into a practical application.
Step 2B:
Claim 5 does not include any additional elements that would provide significantly more.
Therefore, claim 5 as a whole does not change this conclusion and the claim is ineligible.
Claim 6 depends from claim 5 and thus include all the limitations of claim 5, therefore claim 6 recites the same abstract ideas of "mental processes".
Claim 6 recites: generating the weight for each of the one or more user events based on at least one of recency of the one or more user events or individualized geo-seasonal intent, wherein generating the weight based on individualized geo-seasonal intent comprises determining an individualized annual pattern of behavior; and identifying a number of digital design templates of the subset of digital design templates based on the weight for each of the one or more user events.
Step 1: Claim 6 as a whole falls within one or more statutory categories.
Step 2A prong 1: Claim 6 recites limitations that are abstract ideas.
The limitations “generating the weight for each of the one or more user events based on at least one of recency of the one or more user events or individualized geo-seasonal intent, wherein generating the weight based on individualized geo-seasonal intent comprises determining an individualized annual pattern of behavior; and identifying a number of digital design templates of the subset of digital design templates based on the weight for each of the one or more user events” are mental steps that can be achieved by a user mentally determining a weight for each user event based on certain selection criteria and identifying a number of templates based on the determined weight. Thus, the claimed limitations can be performed by the human mind.
Step 2A prong 2: Claim 6 does not recite any additional elements that would integrate the judicial exception into a practical application.
Step 2B:
Claim 6 does not include any additional elements that would provide significantly more.
Therefore, claim 6 as a whole does not change this conclusion and the claim is ineligible.
Claim 7 depends from claim 1 and thus include all the limitations of claim 1, therefore claim 7 recites the same abstract ideas of "mental processes".
Claim 7 recites: detecting that the one or more user events of the user includes a digital design template export and a search query performed by the user; and
assigning, utilizing an exponential decay function, a first weight to the digital design template export and a second weight to the search query performed by the user, wherein the first weight is greater than the second weight.
Step 1: Claim 7 as a whole falls within one or more statutory categories.
Step 2A prong 1: Claim 7 recites limitations that are abstract ideas.
The limitation “detecting that the one or more user events of the user includes a digital design template export and a search query performed by the user” is a mental step. A user can visually determine that a user event relates to a digital design template export or a search query. Thus, the claimed limitation can be performed by the human mind.
The limitation “assigning, utilizing an exponential decay function, a first weight to the digital design template export and a second weight to the search query performed by the user, wherein the first weight is greater than the second weight” is a mental step. A user can mentally assign weights to different items of data based on certain criteria – in this case, an exponential decay function. Thus, the claimed limitation can be performed by the human mind.
Step 2A prong 2: Claim 7 does not recite any additional elements that would integrate the judicial exception into a practical application.
Step 2B:
Claim 7 does not include any additional elements that would provide significantly more.
Therefore, claim 7 as a whole does not change this conclusion and the claim is ineligible.
Claim 8 depends from claim 7 and thus include all the limitations of claim 7, therefore claim 8 recites the same abstract ideas of "mental processes".
Claim 8 recites: wherein comparing the user embedding vector with one or more embedding vectors of the plurality of embedding vectors for the digital design templates further comprises performing real-time pairwise distance computations across the single latent embedding space to identify, in real-time, the subset of digital design templates.
Step 1: Claim 8 as a whole falls within one or more statutory categories.
Step 2A prong 1: Claim 8 recite limitations that are abstract ideas.
The limitation “performing real-time pairwise distance computations across the single latent embedding space to identify, in real-time, the subset of digital design templates” are mathematical calculations and fall under the ‘mathematical concepts’ grouping of abstract ideas.
Step 2A prong 2: Claim 8 does not recite any additional elements that would integrate the judicial exception into a practical application.
Step 2B:
Claim 8 does not include any additional elements that would provide significantly more.
Therefore, claim 8 as a whole does not change this conclusion and the claim is ineligible.
Claim 15 depends from claim 9 and thus include all the limitations of claim 9, therefore claim 15 recites the same abstract ideas of "mental processes".
Claim 15 recites: wherein identifying the subset of digital design templates further comprises: comparing, utilizing the similarity search model, the user embedding vector with one or more embedding vectors of the plurality of embedding vectors for the digital design templates within the single latent embedding space to determine pairwise distance computations; and determining which embedding vectors of the one or more embedding vectors satisfy the similarity threshold to the user embedding vector based on the pairwise distance computations.
Step 1: Claim 15 as a whole falls within one or more statutory categories.
Step 2A prong 1: Claim 15 recite limitations that are abstract ideas.
The limitation “comparing, utilizing the similarity search model, the user embedding vector with one or more embedding vectors of the plurality of embedding vectors for the digital design templates within the single latent embedding space to determine pairwise distance computations; and determining which embedding vectors of the one or more embedding vectors satisfy the similarity threshold to the user embedding vector based on the pairwise distance computations” are mental steps that can be achieved by comparing user embedding vectors based on a selected similarity threshold based on distance computations. Thus, the claimed limitations can be performed by the human mind.
Step 2A prong 2: Claim 15 recites the following additional elements “the similarity search model”, note that these recited additional elements are a high-level recitation of generic computer software to perform the mental process and applied on a computer as in MPEP 2106.05(f), which does not provide integration into a practical application.
Step 2B:
The conclusions for the additional elements representing mere implementation using a computer are carried over and do not provide significantly more.
Therefore, claim 15 as a whole does not change this conclusion and the claims are ineligible.
Claim 11 depends from claim 9 and thus include all the limitations of claim 9, therefore claim 11 recites the same abstract ideas of "mental processes".
Claim 11 recites:
accessing a plurality of user events corresponding to a user account of the user, wherein the plurality of user events is from multiple disparate software applications; and
providing the subset of digital design templates from the digital design templates to recommend to the user across the multiple disparate software applications.
Step 1: Claim 11 as a whole falls within one or more statutory categories.
Step 2A prong 1: Claim 11 recite limitations that are abstract ideas.
The limitation “accessing a plurality of user events corresponding to a user account of the user, wherein the plurality of user events is from multiple disparate software applications” is a mental step. A user can visually select user events corresponding to a user account from a displayed listing. Thus the claimed limitation can be performed by the human mind.
Step 2A prong 2:
The limitation “providing the subset of digital design templates from the digital design templates to recommend to the user across the multiple disparate software applications” is also an additional element and is insignificant extra-solution activity as selecting and outputting information for display, as identified in MPEP 2106.05(g) in Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354-55, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016, and does not provide integration into a practical application.
Step 2B:
With respect to the ”providing” limitations identified as insignificant extra-solution activity above, when re-evaluated these elements are well-understood, routine, and conventional as evidenced by the court cases in MPEP 2106.05(d)(II), "i. Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); … OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network);" and thus remains insignificant extra-solution activity that does not provide significantly more.
Therefore, the claim 11 as a whole does not change this conclusion and the claim is ineligible.
Claims 13, 14, and 19 depend from claims 9 and 16 and thus include all the limitations of claims 9 and 16, therefore claims 13, 14, and 19 recite the same abstract ideas of "mental processes".
Claims 13, 14, and 19 recite:
(claim 13) wherein assigning the weights further comprises assigning, utilizing the exponential decay function, a first weight to the digital design template export and a second weight to the search query performed by the user, wherein the first weight is greater than the second weight;
(claim 14) wherein assigning the weights to the one or more user events further comprises identifying a number of digital design templates of the subset of digital design templates based on the weights for each of the one or more user events; and
(claim 19) wherein the operations further comprise: generating a weight for each of the plurality of user events of the user by utilizing an exponential decay function; and identifying a number of digital design templates of the subset of digital design templates based on the weight for each of the plurality of user events.
Step 1: Claims 13, 14, and 19 as a whole fall within one or more statutory categories.
Step 2A prong 1: Claims 13, 14, and 19 recites limitations that are abstract ideas.
The limitations “assigning/generating the weights further comprises assigning, utilizing the exponential decay function, a first weight to the digital design template export and a second weight to the search query performed by the user, wherein the first weight is greater than the second weight” and ” assigning the weights to the one or more user events further comprises identifying a number of digital design templates of the subset of digital design templates based on the weights for each of the plurality of user events” are mental steps that can be achieved by a user mentally determining/generating weights to assign to templates and search queries and mentally identifying templates based on the weights for user events. Thus, the claimed limitations can be performed by the human mind.
Step 2A prong 2: Claims 13, 14, and 19 do not recite any additional elements that would integrate the judicial exception into a practical application.
Step 2B:
Claims 13, 14, and 19 do not include any additional elements that would provide significantly more.
Therefore, claims 13, 14, and 19 as a whole do not change this conclusion and the claims are ineligible.
Claim 20 depends from claim 16 and thus include all the limitations of claim 16, therefore claim 20 recites the same abstract ideas of "mental processes".
Claim 20 recites: wherein identifying the subset of digital design templates further comprises comparing, utilizing the similarity search model, the user embedding vector with the one or more embedding vectors for the digital design templates within a single latent space to determine in real-time which embedding vectors of the one or more embedding vectors satisfy the similarity threshold to the user embedding vector.
Step 1: Claim 20 as a whole falls within one or more statutory categories.
Step 2A prong 1: Claim 20 recites limitations that are abstract ideas.
The limitation “identifying the subset of digital design templates further comprises comparing the user embedding vector with the one or more embedding vectors for the digital design templates within a single latent space to determine in real-time which embedding vectors of the one or more embedding vectors satisfy the similarity threshold to the user embedding vector” are mental steps that can be achieved by comparing user embedding vectors based on a selected similarity threshold. Thus, the claimed limitations can be performed by the human mind.
Step 2A prong 2: Claim 20 recites the following additional elements “the similarity search model”, note that these recited additional elements are a high-level recitation of generic computer software to perform the mental process and applied on a computer as in MPEP 2106.05(f), which does not provide integration into a practical application.
Step 2B:
The conclusions for the additional elements representing mere implementation using a computer are carried over and do not provide significantly more.
Therefore, claim 20 as a whole does not change this conclusion and the claim is ineligible.
To expedite a complete examination of the instant application, the claims rejected under 35 U.S.C. 101 (nonstatutory} above are further rejected as set forth below in anticipation of applicant amending these claims to place them within the four statutory categories of the invention.
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 for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
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.
Claims 1, 2, 16, 17 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over US PgPub 2021/0174164 by Hsieh et al (hereafter Hsieh), in view of US PgPub 2023/0325391 by Li et al (hereafter Li), and further in view of US Patent 12,511,679 issued to Batina et al (hereafter Batina).
Referring to claim 1, Hsieh discloses a method comprising [Abstract]:
extracting metadata from content items [content repository 110 is populated from data from a data source looking to make use of the system, e.g. product related data, content related data or crawled data may be uploaded to the content repository, para 47, see Fig 1];
aggregating one or more user events into a combined user signal based on detecting one or more user events of a user with respect to a database of content items [user feature data includes a (i.e. combined) set of interaction data pertaining to user interactions with the content items, para 59-60];
generating, utilizing a single transformer neural network, both a user embedding vector from the combined user signal [wherein a user embedding 130 is created from user feature data input into a user neural network model (U-NN) 130 and processed through a matchmaking neural network model in a shared dimensional space to yield a user shared-item embedding S220, para 79-80, Fig 1; Fig 5, elements S210-S220; user shared-item embeddings represent user interaction event data, para 60, 76; wherein a single item-related neural network can be used in place of a C-NN120 and a U-NN 130, para 61] and a plurality of embedding vectors for the content items based on the metadata from the content items [wherein a content neural network (C-NN) 120 functions as a neural network model of items constituting incorporation and mapping of content data items in relation to each other, para 49; C-NN 120 can process actual semantic information embedded within a photo media content item to identify key objects within, or from text media to identify important concepts and keywords, para 51; a transformer may be used to enrich the content embeddings, para 51; content items are represented as embeddings or vectors, para 76; processors 1002A-N include Machine learning/Deep learning processing units such as a Tensor Processing unit, para 211, Fig 21];
encoding the user embedding vector from the combined user signal and the plurality of embedding vectors from the content items into a single latent embedding space [matchmaking neural network (MmNN) 140 maps the user embeddings and content embeddings of the U-NN 130 and C-NN 120 models into a shared vector space, para 64-65];
identifying, utilizing a similarity search model, a subset of content items from the content items to recommend to the user by comparing the user embedding vector with the plurality of embedding vectors within the single latent embedding space to identify one or more embedding vectors that satisfy a similarity threshold to the user embedding vector [matchmaking neural network (MmNN) 140 maps the user embeddings and content embeddings of the U-NN 130 and C-NN 120 models into a shared vector space, para 64-65; the MmNN 140 model converts the content and user embeddings into content and user shared-item embeddings respectively, para 65; wherein an analysis of a shared-item embedding is applied in the MmNN 140 model with respect to a user shared-item embedding in selecting at least one content item, para 79, Fig 4, element S140, Fig 5, element S230; analysis of shared-item embeddings include identifying one or more shared-item embeddings satisfying a proximity condition such as a count of content-shared item embeddings within a specified distance threshold from a user shared-item embedding to determine content or users with some degree of similarity, para 128; Euclidean distance thresholds, para 181-183].
Referring to claim 1, while Hsieh discloses all of the above claimed subject matter, and also discloses that the content items include photo/graphical items [para 50] and a single item-related neural network can be used in place of a C-NN120 and a U-NN 130 [para 61], it remains silent as to the graphical content items specifically being digital design templates; and that the single transformer neural network comprises self-attention layers.
Li discloses that visual content library 150 includes a vast library of visual assets comprises a template library 158 that includes templates containing image, icon and illustration content types [see Fig 1B, element 158, para 43-46].
Hsieh and Li are analogous art that are directed to the same field of endeavor- the analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the photo/graphical content items of Hsieh with the templates of Li because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the image, icon and illustration template assets of Li further refine the type of photo/graphical content items taught by Hsieh.
Still referring to claim 1, while Hsieh/Li discloses all of the above claimed subject and also discloses that a single item-related neural network can be used in place of a C-NN120 and a U-NN 130 [Hsieh, para 61], it remains silent as to the single transformer neural network comprising self-attention layers.
Batina discloses a transformer 50 that includes encoder 52 and decoder 54 which each include a plurality of neural network layers at least one of which includes a self-attention layer [col. 10, lines 20-30, Fig 6].
Hsieh, Li and Batina are analogous art that are directed to the same field of endeavor- the analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the single item-related neural network of Hsieh to include the self-attention layers of the encoder/decoder of Batina because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the self-attention neural network layers of Batina further refines the transformer and single item-related neural network Hsieh.
Referring to claim 16, Hsieh discloses a non-transitory computer readable medium storing executable instructions which, when executed by at least one processing device [processor readable storage medium 1005, processors 1002A-N, para 210, Fig 21], cause the at least one processing device to perform operations comprising:
extracting metadata from content items from a database of content items [content repository 110 is populated from data from a data source looking to make use of the system, e.g. product related data, content related data or crawled data may be uploaded to the content repository, para 47, see Fig 1];
generating, utilizing a transformer neural network, a plurality of embedding vectors for the content items from the extracted metadata [wherein a content neural network (C-NN) 120 functions as a neural network model of items constituting incorporation and mapping of content data items in relation to each other, para 49; C-NN 120 can process actual semantic information embedded within a photo media content item to identify key objects within, or from text media to identify important concepts and keywords, para 51; a transformer may be used to enrich the content embeddings, para 51; content items are represented as embeddings or vectors, para 76; processors 1002A-N include Machine learning/Deep learning processing units such as a Tensor Processing unit, para 211, Fig 21];
accessing a plurality of user events of a user corresponding to a user account [user profile feature data includes a set of interaction data pertaining to user interactions with the content items, para 59-60], wherein the plurality of user events is from multiple disparate software applications [instances of the method can operate for different accounts, websites, applications and resources can be shared across different accounts in recommendations operations for multiple accounts, para 86-87];
generating, utilizing the transformer neural network, a user embedding vector from the plurality of user events of a user from the multiple disparate software applications [wherein a user embedding 130 is created from user feature data input into a user neural network model (U-NN) 130 and processed through a matchmaking neural network model in a shared dimensional space to yield a user shared-item embedding S220, para 79-80, Fig 1; Fig 5, elements S210-S220; user shared-item embeddings represent user interaction event data, para 60, 76; wherein a single item-related neural network can be used in place of a C-NN120 and a U-NN 130, para 61; instances of the method can operate for different accounts, websites, applications and resources can be shared across different accounts in recommendations operations for multiple accounts, para 86-87]; and
identifying, utilizing a similarity search model, a subset of digital design templates from the database of digital design templates to recommend to the user account of the user by identifying one or more embedding vectors of the plurality of embedding vectors that satisfy a similarity threshold to the user embedding vector [matchmaking neural network (MmNN) 140 maps the user embeddings and content embeddings of the U-NN 130 and C-NN 120 models into a shared vector space, para 64-65; the MmNN 140 model converts the content and user embeddings into content and user shared-item embeddings respectively, para 65; wherein an analysis of a shared-item embedding is applied in the MmNN 140 model with respect to a user shared-item embedding in selecting at least one content item, para 79, Fig 4, element S140, Fig 5, element S230; analysis of shared-item embeddings include identifying one or more shared-item embeddings satisfying a proximity condition such as a count of content-shared item embeddings within a specified distance threshold from a user shared-item embedding to determine content or users with some degree of similarity, para 128; Euclidean distance thresholds, para 181-183; user profile, para 58-59]; and
providing the subset of digital design templates as recommendations across the multiple disparate software applications [content-centric personalized recommendations, para 9, 45; resources can be shared across different accounts in recommendations operations for multiple accounts, para 86-87].
Referring to claim 16, while Hsieh discloses all of the above claimed subject matter, and also discloses that the content items include photo/graphical items [para 50], it remains silent as to the graphical content items specifically being digital design templates; that the single transformer neural network comprises self-attention layers.
Li discloses that visual content library 150 includes a vast library of visual assets comprises a template library 158 that includes templates containing image, icon and illustration content types [see Fig 1B, element 158, para 43-46].
Hsieh and Li are analogous art that are directed to the same field of endeavor- the analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the photo/graphical content items of Hsieh with the templates of Li because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the image, icon and illustration template assets of Li further refine the type of photo/graphical content items taught by Hsieh.
Still referring to claim 16, while Hsieh/Li discloses all of the above claimed subject and also discloses that a single item-related neural network can be used in place of a C-NN120 and a U-NN 130 [Hsieh, para 61], it remains silent as to the single transformer neural network comprising self-attention layers.
Batina discloses a transformer 50 that includes encoder 52 and decoder 54 which each include a plurality of neural network layers at least one of which includes a self-attention layer [col. 10, lines 20-30, Fig 6].
Hsieh, Li and Batina are analogous art that are directed to the same field of endeavor- the analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the single item-related neural network of Hsieh to include the self-attention layers of the encoder/decoder of Batina because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the self-attention neural network layers of Batina further refines the transformer and single item-related neural network Hsieh.
Referring to claim 2, Hsieh/Li/Batina discloses that extracting the metadata from the digital design templates comprises extracting at least one of title of digital design templates, text, description, categories, topics, or tasks [Hsieh, content title/name, description, categories, tags, authors, para 100].
Referring to claim 17, Hsieh/Li/Batina discloses that extracting the metadata further comprises: extracting from the digital design templates at least one of title, text, description, categories, topics, or tasks [Hsieh, content title/name, description, categories, tags, authors, para 100]; and determining stylistic elements of the digital design templates based on at least one of the title, the text, the description, the categories, the topics, or the tasks [Hsieh, colors, para 100; content properties collected for a particular piece of content can change depending on the content type and objectives for content suggestions, para 100].
Referring to claim 20, Hsieh/Li/Batina discloses that identifying the subset of digital design templates further comprises comparing, utilizing the similarity search model, the user embedding vector with the one or more embedding vectors for the digital design templates to determine in real-time which embedding vectors of the one or more embedding vectors satisfy the similarity threshold to the user embedding vector [Hsieh, matchmaking neural network (MmNN) 140 maps the user embeddings and content embeddings of the U-NN 130 and C-NN 120 models into a shared vector space, para 64-65; the MmNN 140 model converts the content and user embeddings into content and user shared-item embeddings respectively, para 65; wherein an analysis of a shared-item embedding is applied in the MmNN 140 model with respect to a user shared-item embedding in selecting at least one content item, para 79, Fig 4, element S140, Fig 5, element S230; analysis of shared-item embeddings include identifying one or more shared-item embeddings satisfying a proximity condition such as a count of content-shared item embeddings within a specified distance threshold from a user shared-item embedding to determine content or users with some degree of similarity, para 128; Euclidean distance thresholds, para 181-183; Hsieh, distance computation between user/subject and contents/object pairs, para 46; Li, real-time personalization, para 164].
Claims 9-11, 14 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Hsieh, in view of Li, in view of Batina, and further in view of US PGPub 2019/0114417 by Subbarayan et al (hereafter Subbarayan).
Referring to claim 9, Hsieh discloses a system [computing system, para 207, 209, Fig 21] comprising: one or more memory components [processor readable storage medium 1005, para 210, Fig 21]; and one or more processing devices coupled to the one or more memory components, the one or more processing devices to perform operations comprising [processors 1002A-N, para 210-212, Fig 21]:
extracting metadata from content items [content repository 110 is populated from data from a data source looking to make use of the system, e.g. product related data, content related data or crawled data may be uploaded to the content repository, para 47, see Fig 1];
detecting one or more user events of a user with respect to the content items [a user device is configured to transmit data captured locally during use of relevant application(s) (i.e. in real-time) to a local or remote ML algorithm and provide supplemental training data that can serve to fine-tune or increase the effectiveness of the ML algorithm, para 33];
generating a combined user signal based on the one or more user events [user feature data includes a (i.e. combined) set of interaction data pertaining to user interactions with the content items, para 59-60] and by assigning weights to the one or more user events to generate one or more weighted user events [wherein the item matchmaking model (CML) applies a rank-based weighting scheme such as weighted approximate rank pairwise loss to the items in the shared embedding space to characterize a user’s relative preference for different items, wherein preference indicates an item liked by the user, para 109];
generating, utilizing a single transformer neural network, both a user embedding vector from the combined user signal [wherein a user embedding 130 is created from user feature data input into a user neural network model (U-NN) 130 and processed through a matchmaking neural network model in a shared dimensional space to yield a user shared-item embedding S220, para 79-80, Fig 1; Fig 5, elements S210-S220; user shared-item embeddings represent user interaction event data, para 60, 76; wherein a single item-related neural network can be used in place of a C-NN120 and a U-NN 130, para 61] and a plurality of embedding vectors for the content items based on the metadata from the content items [wherein a content neural network (C-NN) 120 functions as a neural network model of items constituting incorporation and mapping of content data items in relation to each other, para 49; C-NN 120 can process actual semantic information embedded within a photo media content item to identify key objects within, or from text media to identify important concepts and keywords, para 51; a transformer may be used to enrich the content embeddings, para 51; content items are represented as embeddings or vectors, para 76; processors 1002A-N include Machine learning/Deep learning processing units such as a Tensor Processing unit, para 211, Fig 21];
encoding the user embedding vector from the combined user signal and the plurality of embedding vectors from the content items into a single latent embedding space [matchmaking neural network (MmNN) 140 maps the user embeddings and content embeddings of the U-NN 130 and C-NN 120 models into a shared vector space, para 64-65]; and
identifying, utilizing a similarity search model, a subset of content items from the content items to recommend to the user by comparing the user embedding vector with the plurality of embedding vectors within the single latent embedding space to identify one or more embedding vectors that satisfy a similarity threshold to the user embedding vector [matchmaking neural network (MmNN) 140 maps the user embeddings and content embeddings of the U-NN 130 and C-NN 120 models into a shared vector space, para 64-65; the MmNN 140 model converts the content and user embeddings into content and user shared-item embeddings respectively, para 65; wherein an analysis of a shared-item embedding is applied in the MmNN 140 model with respect to a user shared-item embedding in selecting at least one content item, para 79, Fig 4, element S140, Fig 5, element S230; analysis of shared-item embeddings include identifying one or more shared-item embeddings satisfying a proximity condition such as a count of content-shared item embeddings within a specified distance threshold from a user shared-item embedding to determine content or users with some degree of similarity, para 128; Euclidean distance thresholds, para 181-183; user profile, para 58-59].
Referring to claim 9, while Hsieh discloses all of the above claimed subject matter, and also discloses that the content items include photo/graphical items [para 50] and a single item-related neural network can be used in place of a C-NN120 and a U-NN 130 [para 61], it remains silent as to the graphical content items specifically being digital design templates; that the user event aggregation model is a real-time model, and the weighting of the items utilizing an exponential decay function and that the single transformer neural network comprises self-attention layers.
Li discloses that visual content library 150 includes a vast library of visual assets comprises a template library 158 that includes templates containing image, icon and illustration content types [see Fig 1B, element 158, para 43-46] and that a user device is configured to transmit data captured locally during use of relevant application(s) (i.e. in real-time) to a local or remote ML algorithm and provide supplemental training data that can serve to fine-tune or increase the effectiveness of the ML algorithm [para 33].
Hsieh and Li are analogous art that are directed to the same field of endeavor- the analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the photo/graphical content items of Hsieh with the templates of Li and to modify the input user feature data to be done through a real-time model because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the image, icon and illustration template assets of Li further refine the type of photo/graphical content items taught by Hsieh. Furthermore, the capturing of data through the real-time aggregation of Li further defines the method through which user feature data is input in Hsieh.
Still referring to claim 9, while Hsieh/Li discloses all of the above claimed subject and also discloses that a single item-related neural network can be used in place of a C-NN120 and a U-NN 130 [Hsieh, para 61], it remains silent as to the weighting of the items utilizing an exponential decay function; and the single transformer neural network comprising self-attention layers.
Batina discloses a transformer 50 that includes encoder 52 and decoder 54 which each include a plurality of neural network layers at least one of which includes a self-attention layer [col. 10, lines 20-30, Fig 6].
Hsieh, Li and Batina are analogous art that are directed to the same field of endeavor- the analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the single item-related neural network of Hsieh to include the self-attention layers of the encoder/decoder of Batina because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the self-attention neural network layers of Batina further refines the transformer and single item-related neural network Hsieh.
Referring to claim 9, while Hsieh/Li/Batina discloses all of the above claimed subject matter, and also discloses assigning weights to the one or more user events to generate one or more weighted user events [Hsieh, wherein the item matchmaking model (CML) applies a rank-based weighting scheme such as weighted approximate rank pairwise loss to the items in the shared embedding space to characterize a user’s relative preference for different items, wherein preference indicates an item liked by the user, para 109], it remains silent as to the weighting of the items utilizing an exponential decay function.
Subbarayan discloses utilizing an exponential decay function to derive weighting coefficients that increase in relative distance between two symbols in a sequence of API call items [para 52 and 54].
Hsieh, Li, Batina and Subbarayan are analogous art that are directed to the same field of endeavor- the analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the rank-based weighting scheme of Hsieh with the utilization of exponential decay function to derive weighting coefficients in Subbarayan because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the exponential decay weighting scheme of Subbarayan is a variation of a type of weighting scheme of the content items taught by Hsieh.
Referring to claim 10, Hsieh/Li/Batina/Subbarayan discloses that extracting the metadata from the digital design templates further comprises extracting stylistic elements of the digital design templates based on at least one of title of digital design templates, text, description, categories, topics, or tasks [Hsieh, content properties including colors, content title/name, description, categories, tags, authors, para 100; Hsieh, content properties collected for a particular piece of content can change depending on the content type and objectives for content suggestions, para 100].
Referring to claim 11, Hsieh/Li/Batina/Subbarayan discloses that the operations further comprise: accessing a plurality of user events of a user corresponding to a user account [Hsieh, user profile feature data includes a set of interaction data pertaining to user interactions with the content items, para 59-60], wherein the plurality of user events is from multiple disparate software applications [Hsieh, instances of the method can operate for different accounts, websites, applications and resources can be shared across different accounts in recommendations operations for multiple accounts, para 86-87]; and providing the subset of digital design templates as recommendations across the multiple disparate software applications [Hsieh, content-centric personalized recommendations, para 9, 45; resources can be shared across different accounts in recommendations operations for multiple accounts, para 86-87; Li, templates containing image, icon and illustration content types, Fig 1B, element 158, para 43-46].
Referring to claim 14, Hsieh/Li/Batina/Subbarayan discloses that assigning the weights to the one or more user events further comprises identifying a number of digital design templates of the subset of digital design templates based on the weights for each of the one or more user events [Hsieh, weighted ranking is used to train the matchmaking model in analysis that returns a prioritized set of candidate content items based in part on the calculated personalization scores, para 127].
Referring to claim 15, Hsieh/Li/Batina/Subbarayan discloses that identifying the subset of digital design templates further comprises: comparing, utilizing the similarity search model, the user embedding vector with one or more embedding vectors of the plurality of embedding vectors for the digital design templates within the single latent embedding space to determine pairwise distance computations; and determining which embedding vectors of the one or more embedding vectors satisfy the similarity threshold to the user embedding vector based on the pairwise distance computations [Hsieh, matchmaking neural network (MmNN) 140 maps the user embeddings and content embeddings of the U-NN 130 and C-NN 120 models into a shared vector space, para 64-65; the MmNN 140 model converts the content and user embeddings into content and user shared-item embeddings respectively, para 65; wherein an analysis of a shared-item embedding is applied in the MmNN 140 model with respect to a user shared-item embedding in selecting at least one content item, para 79, Fig 4, element S140, Fig 5, element S230; analysis of shared-item embeddings include identifying one or more shared-item embeddings satisfying a proximity condition such as a count of content-shared item embeddings within a specified distance threshold from a user shared-item embedding to determine content or users with some degree of similarity, para 128; Euclidean distance thresholds, para 181-183; Li, templates containing image, icon and illustration content types, Fig 1B, element 158, para 43-46].
Claims 3, 4 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Hsieh, in view of Li, in view of Batina, as applied to claims 1 and 16, and further in view of US PGPub 2022/0012296 by Marey.
Referring to claim 3, Hsieh/Li/Batina discloses all of the above claimed subject matter, and also discloses generating the plurality of embedding vectors that represent content within the digital design templates; and storing the plurality of embedding vectors in a database corresponding with the digital design templates [Hsieh, generated shared-item embeddings are stored as part of a matchmaking data model, para 65], it remains silent as to utilizing a sentence transformer to generate the embedding vectors.
Marey discloses that a sentence embedding machine learning model is employed to recommend a semantically similar post and word/sentence embeddings are created from the text in social media post items [para 40-41].
Hsieh, Li, Batina and Marey are analogous art that are directed to the same field of endeavor- analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the generation of the shared item embeddings of Hsieh to include sentence embeddings created in Marey because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the sentence embeddings created from social media post content items in Marey further refines the type of content shared-item embeddings in Hsieh.
Referring to claim 4, Hsieh/Li/Batina discloses all of the above claimed subject matter, and also discloses detecting one or more events of the user in real-time by utilizing a user event stream aggregation model [Hsieh, user device configured to transmit data captured locally during use of relevant application(s) (i.e. in real-time) to a local or remote ML algorithm and provide supplemental training data that can serve to fine-tune or increase the effectiveness of the ML algorithm, para 33], wherein the one or more events include search queries performed by the user [Hsieh, product search results, para 135]; and generating the user embedding vector from the detected one or more events [Hsieh, user embedding created from user feature data input into a user neural network model (U-NN) 130 to yield a user shared-item embedding S220, para 79-80, Fig 1; Fig 5, elements S210-S220; user shared-item embeddings represent user interaction event data, para 60, 76].
However, it remains silent as to the one or more events of the user including digital design template remixes and digital design template exports. Marey discloses that social media post templates may be inputted or extracted from past social media post records including historical social media conversations and modified prior to providing the post as a recommended post [para 34-35].
Hsieh, Li, Batina and Marey are analogous art that are directed to the same field of endeavor- analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the user events of Hsieh to include the identification and modification of templates similar to a user’s social media posts in past historical social media conversations of Marey because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the modified templates in Marey further refine the type of user interaction event data in Hsieh.
Referring to claim 18, Hsieh/Li/Batina discloses all of the above claimed subject matter, and also discloses detecting one or more events of the user in real-time by utilizing a user event stream aggregation model [Hsieh, user device configured to transmit data captured locally during use of relevant application(s) (i.e. in real-time) to a local or remote ML algorithm and provide supplemental training data that can serve to fine-tune or increase the effectiveness of the ML algorithm, para 33], wherein the one or more events include search queries performed by the user [Hsieh, product search results, para 135].
However, it remains silent as to the one or more events of the user including digital design template remixes and digital design template exports. Marey discloses that social media post templates may be inputted or extracted from past social media post records including historical social media conversations and modified prior to providing the post as a recommended post [para 34-35].
Hsieh, Li, Batina and Marey are analogous art that are directed to the same field of endeavor- analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the user events of Hsieh to include the identification and modification of templates similar to a user’s social media posts in past historical social media conversations of Marey because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the modified templates in Marey further refine the type of user interaction event data in Hsieh.
Claims 5, 6 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Hsieh, in view of Li, in view of Batina, as applied to claims 1 and 16 above, and further in view of Subbarayan.
Referring to claim 5, Hsieh/Li/Batina discloses all of the above claimed subject matter, and also discloses: generating a weight for each of the one or more user events of the user [Hsieh, wherein the item matchmaking model (CML) applies a rank-based weighting scheme such as weighted approximate rank pairwise loss to the items in the shared embedding space to characterize a user’s relative preference for different items, wherein preference indicates an item liked by the user, para 109], however it remains silent as to the weighting of the items utilizing an exponential decay function.
Subbarayan discloses utilizing an exponential decay function to derive weighting coefficients that increase in relative distance between two symbols in a sequence of API call items [para 52 and 54].
Hsieh, Li, Batina and Subbarayan are analogous art that are directed to the same field of endeavor- the analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the rank-based weighting scheme of Hsieh with the utilization of exponential decay function to derive weighting coefficients in Subbarayan because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the exponential decay weighting scheme of Subbarayan is a variation of a type of weighting scheme of the content items taught by Hsieh.
Referring to claim 6, Hsieh/Li/Batina/Subbarayan discloses generating the weight for each of the one or more user events based on at least one of recency of the one or more user events or individualized geo-seasonal intent, wherein generating the weight based on individualized geo-seasonal intent comprises determining an individualized annual pattern of behavior [Hsieh, personalization score calculated by taking into account additional factors such as content trending properties (e.g. recent change in popularity) and incorporated into the reranking process used to filter or set candidate items, para 134; personalization score is based on comparisons of content items most recently visited, purchased, favorited or otherwise interacted with by the user, para 138; Subbarayan, weighting using exponential decay function, para 52,54]; and identifying a number of digital design templates of the subset of digital design templates based on the weight for each of the one or more user events [Hsieh, weighted ranking is used to train the matchmaking model in analysis that returns a prioritized set of candidate content items based in part on the calculated personalization scores, para 127].
Referring to claim 19, Hsieh/Li/Batina discloses all of the above claimed subject matter, and also discloses: generating a weight for each of the plurality of user events of the user [Hsieh, wherein the item matchmaking model (CML) applies a rank-based weighting scheme such as weighted approximate rank pairwise loss to the items in the shared embedding space to characterize a user’s relative preference for different items, wherein preference indicates an item liked by the user, para 109] and identifying a number of digital design templates of the subset of digital design templates based on the weight for each of the plurality of user events [Hsieh, weighted ranking is used to train the matchmaking model in analysis that returns a prioritized set of candidate content items based in part on the calculated personalization scores, para 127].
However it remains silent as to the weighting of the items utilizing an exponential decay function.
Subbarayan discloses utilizing an exponential decay function to derive weighting coefficients that increase in relative distance between two symbols in a sequence of API call items [para 52 and 54].
Hsieh, Li, Batina and Subbarayan are analogous art that are directed to the same field of endeavor- the analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the rank-based weighting scheme of Hsieh with the utilization of exponential decay function to derive weighting coefficients in Subbarayan because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the exponential decay weighting scheme of Subbarayan is a variation of a type of weighting scheme of the content items taught by Hsieh.
Claims 7 and 8 are rejected under 35 U.S.C. 103 as being unpatentable over Hsieh, in view of Li, in view of Batina, as applied to claim 1, in view of Subbarayan, and further in view of Marey.
Referring to claim 7, Hsieh/Li/Batina discloses all of the above claimed subject matter, and also discloses detecting that the one or more user events include a search query performed by the user [Hsieh, product search results, para 135]; and
assigning the weights further comprises assigning a first weight to the digital design template export and a second weight to the search query performed by the user, wherein the first weight is greater than the second weight [Hsieh, personalization score calculated by taking into account additional factors such as content trending properties (e.g. recent change in popularity) and incorporated into the reranking process used to filter or set candidate items, para 134; personalization score is based on comparisons of content items most recently visited, purchased, favorited or otherwise interacted with by the user, para 138; Hsieh, product search results, para 135].
However, it remains silent as to the one or more events of the user including a digital design template export; and the weighting of the items utilizing an exponential decay function.
Subbarayan discloses utilizing an exponential decay function to derive weighting coefficients that increase in relative distance between two symbols in a sequence of API call items [para 52 and 54; Subbarayan, weighting using exponential decay function involves contexts including positions further away from the symbol being analyzed associated with lesser weights than symbols positioned closer to the symbol being analyzed, para 54].
Hsieh, Li and Subbarayan are analogous art that are directed to the same field of endeavor- the analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the rank-based weighting scheme of Hsieh with the utilization of exponential decay function to derive weighting coefficients in Subbarayan because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the exponential decay weighting scheme of Subbarayan is a variation of a type of weighting scheme of the content items taught by Hsieh.
Still referring to claim 7, while Hsieh/Li/Batina/Subbarayan discloses all of the above claimed subject matter, it remains silent as to the one or more events of the user including a digital design template export.
Marey discloses that social media post templates may be inputted or extracted from past social media post records including historical social media conversations and modified prior to providing the post as a recommended post [para 34-35].
Hsieh, Li, Batina, Subbarayan and Marey are analogous art that are directed to the same field of endeavor- analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the user events of Hsieh to include the modification of templates similar to a user’s social media posts in past historical social media conversations of Marey because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the modified templates in Marey further refine the type of user interaction event data in Hsieh.
Referring to claim 8, Hsieh/Li/Batina/Subbarayan/Marey discloses that comparing the user embedding vector with one or more embedding vectors of the plurality of embedding vectors for the digital design templates further comprises performing real-time pairwise distance computations across the single latent embedding space to identify, in real-time, the subset of digital design templates [Hsieh, distance computation between user/subject and contents/object pairs, para 46; Li, real-time personalization, para 164].
Claims 12 and 13 are rejected under 35 U.S.C. 103 as being unpatentable over Hsieh, in view of Li, in view of Batina, in view of Subbarayan, as applied to claim 9, and further in view of Marey.
Referring to claim 12, Hsieh/Li/Batina/Subbarayan discloses all of the above claimed subject matter, and also discloses detecting that the one or more user events include a search query performed by the user [Hsieh, product search results, para 135].
However, it remains silent as to the one or more events of the user including a digital design template export. Marey discloses that social media post templates may be inputted or extracted from past social media post records including historical social media conversations and modified prior to providing the post as a recommended post [para 34-35].
Hsieh, Li, Batina, Subbarayan and Marey are analogous art that are directed to the same field of endeavor- analysis of content items. It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to modify the user events of Hsieh to include the modification of templates similar to a user’s social media posts in past historical social media conversations of Marey because it would achieve predictable results.
The ordinary skilled artisan would have been motivated to make this modification because the modified templates in Marey further refine the type of user interaction event data in Hsieh.
Referring to claim 13, Hsieh/Li/Batina/Subbarayan/Marey discloses that assigning the weights further comprises assigning, utilizing the exponential decay function, a first weight to the digital design template export and a second weight to the search query performed by the user, wherein the first weight is greater than the second weight [Hsieh, personalization score calculated by taking into account additional factors such as content trending properties (e.g. recent change in popularity) and incorporated into the reranking process used to filter or set candidate items, para 134; personalization score is based on comparisons of content items most recently visited, purchased, favorited or otherwise interacted with by the user, para 138; Subbarayan, weighting using exponential decay function involves contexts including positions further away from the symbol being analyzed associated with lesser weights than symbols positioned closer to the symbol being analyzed, para 54; Hsieh, product search results, para 135; Marey, modified social media post templates, para 34-35].
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 nonprovisional extension fee (37 CFR 1.17(a)) 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 mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to CHERYL M SHECHTMAN whose telephone number is (571)272-4018. The examiner can normally be reached on M-F: 10am-6:30pm.
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, Amy Ng can be reached on 571-270-1698. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
CHERYL M SHECHTMANPatent Examiner
Art Unit 2164
/C.M.S/
/AMY NG/Supervisory Patent Examiner, Art Unit 2164
Prosecution Insights