DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Objection
Claim 1, 4, 5 are objected to because of having following informalities:
Following abbreviated terms are recited “API”, “SVD”, “BERT”, “SHAP” without expanding/elaborating them.
Claim 2, 3, 6, 7 and 9-22 are objected for their dependency on objected base claims.
Appropriate corrections are required.
Response to Amendment
This Office action is in response to Applicant's amendment filed on 10/4/2024.
Claim 1-22 are pending. Claim 1, 4, 5, 8, 9, 10, 12, 13-16 are amended. Claim 18-22 are newly added. Claim 1-22 are rejected.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim 1 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Crabtree, Jason et al (PGPUB Document No. 20240348663), hereafter, referred to as “Crabtree”.
Regarding Claim 1, Bleicher teaches A system for enhancing digital interactions and marketing strategies, comprising: a processing system configured to(Bleicher, abstract discloses a system for marketing using shopping avatar “system, and method to facilitate personalized online shopping……. personalized shopping assistant application to match one or more products from the one or more product data sources to one or more personalized user shopping avatars”):
receive body scan data from a plurality of users captured via smartphone cameras(Bleicher, para 0063 discloses scanning of users by smartphone for receiving physical attributes of users “In some embodiments a dedicated or generic application on a smart phone, tablet or other computing device may be used to enable effective photographing or scanning of a user” ); extract physical measurements from the body scan data by processing captured images using machine learning algorithms to generate a comprehensive digital model including body circumferences, volumes, and surface area metrics(Bleicher, para 0086 discloses extracting anatomical attributes of users using machine learning “scanned or graphic data may be acquired for a user, for example, from standard photographs, 2D and/or 3D image scanning etc…….computer vision and/or machine learning algorithms may be applied, for example, to identify the Eyes, nose, nose bridge, temples, ears etc” );
synchronizing the extracted physical measurements with user profiles through secure API connections(Bleicher, para 0087 discloses generation of user profile using physical measurement data and additional data such as user’s history and preference data “At step 720 the processed user scanned data may be parameterized, for example, to extract the precise face length, width, height, proportions, nose width and volume….together with the user history data from 705 and user preferences data at 715, to generate a user's glasses shopping profile based on the user's face profile and other physical properties, as well as user behavior and user preference data”);
match users with relevant influencers and associated content based on the physical measurements matching the user with relevant content based on the profile(Bleicher, para 0088 discloses matching products/influencers based on user profile which has physical measurement data along with historical and preference data “At step 730 a matchmaking of the user glasses shopping profile to the glasses …..the advanced matching up of appropriate products, in accordance with the specific user's personal shopping profile and preferences”);
But Bleicher does not explicitly teach receive user-submitted preference data via an interactive onboarding questionnaire, wherein the preference data includes style categories, values, or content interests;
create user profiles incorporating the physical measurements, user-submitted preference data and additional user data by integrating social media data through OAuth protocols for secure access delegation;
implementing pixel tracking across social content and landing pages to capture conversion events;
wherein the matching comprises analyzing audience demographics, style preferences, physical congruence and conversion behavior using vector similarity scores weighted by contextual relevance factors:
wherein the system utilizes a hybrid engine combining collaborative filtering such as SVD matrix factorization and content-based filtering approaches deployed on enterprise- grade AI infrastructure;
wherein semantic similarity is computed using Sentence-BERT embeddings stored and queried through Milvus vector databases; wherein behavioral and interactional data are continuously learned and adapted through real-time feedback loops to dynamically improve match precision.
generate AI marketing personas analyzing social media metadata using deep learning to calculate SHAP value outputs; creating targeted demographic personas with specific income characteristics relative to product pricing; facilitate live shopping interactions between matched users through: implementing API connections with external commerce platforms; tracking clicks, cart additions, and purchase conversions across user interactions.
However, in the same field of endeavor of user profile creation Montgomery teaches receive user-submitted preference data via an interactive onboarding questionnaire, wherein the preference data includes style categories, values, or content interests(Montgomery, para 0040 disclose collecting data from questionnaire/surveys “the surveys include questions that gather demographic information about the surveyed individual as well as information about elements that are of interest to the individual. Demographic questions may include, for example, questions regarding age, gender, zip code, income level, education level, etc. Element questions may include questions about what types of movies, sports and other recreational activities are of interest to the individual, what types of products or services the individual purchases, etc. Market research data 703 may include data collected about individual consumers”);
create user profiles incorporating the physical measurements, user-submitted preference data and additional user data by(Montgomery, Fig. 4 and para 0033 disclose creation of user profile using user preference and social media data (any tracking data) “profile generation illustrated in the process flow 300, including examples of data types. Data inputs box 402 shows example viewer data that may be provided to generate viewer profiles. This data may include the data received from tracking viewers as described above. For example, viewer element data may indicate elements that a viewer "likes" or otherwise indicates an interest. Offline zip or zip code-based data may indicate general demographic data describing people who live in the same geographic area as a viewer (e.g., as derived from the viewers network address). Site data may describe web sites that the viewer visits after placement of the cookie, as described herein”; where prior art Andriukhin to be discussed next discloses accessing data via OAuth protocols);
implementing pixel tracking across social content and landing pages to capture conversion events(Montgomery, para 0030 discloses implementing pixel tracking (can be used for social media) and conversion “When a viewer (e.g., via a viewer device 110) downloads content including the tracking pixel, the tracking pixel causes the viewer device 110 to direct a cookie request to the targeting server 102 or another suitable server”; para 0033 further disclose conversion data “ Event data may describe things that the viewer does after placement of the cookie including, for example, being served an advertisement or other content, clicking through the advertisement or other content, converting the advertisement or other content, etc. Time of day and day of week data may indicate the times and dates at which the viewer performs the various activities described herein”);
facilitate live shopping interactions between matched users through: implementing API connections with external commerce platforms; tracking clicks, cart additions, and purchase conversions across user interactions(Montgomery, para 0030 facilitating shopping/commercial activity such as conversion (activities end up in purchase) “When a viewer (e.g., via a viewer device 110) downloads content including the tracking pixel, the tracking pixel causes the viewer device 110 to direct a cookie request to the targeting server 102 or another suitable server”; para 0033 further disclose conversion data “ Event data may describe things that the viewer does after placement of the cookie including, for example, being served an advertisement or other content, clicking through the advertisement or other content, converting the advertisement or other content, etc. Time of day and day of week data may indicate the times and dates at which the viewer performs the various activities described herein”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of tracking and extracting user activities data of Montgomery into feature of extracting user characteristics from image of Bleicher and Melcher to produce an expected result of profiling users based on user characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to track user activities and provide contents accordingly (Montgomery, para 0030).
But Bleicher and Montgomery don’t explicitly teach integrating social media data through OAuth protocols for secure access delegation; wherein the matching comprises analyzing audience demographics, style preferences, physical congruence and conversion behavior using vector similarity scores weighted by contextual relevance factors: wherein the system utilizes a hybrid engine combining collaborative filtering such as SVD matrix factorization and content-based filtering approaches deployed on enterprise- grade AI infrastructure;
wherein semantic similarity is computed using Sentence-BERT embeddings stored and queried through Milvus vector databases; wherein behavioral and interactional data are continuously learned and adapted through real-time feedback loops to dynamically improve match precision.
generate AI marketing personas analyzing social media metadata using deep learning to calculate SHAP value outputs; creating targeted demographic personas with specific income characteristics relative to product pricing;
However, in the same field of endeavor of securing access Andriukhin teaches integrating social media data through OAuth protocols for secure access delegation(Andriukhin, para 0045 discloses integrating or interfacing via secure APIs employing OAuth protocol “integration Layer 291 allows system 200 to interface with external systems via secure APIs, employing protocols like OAuth 2.0 for secure access”);
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of securing access of Andriukhin into feature of extracting user characteristics from image of Bleicher and Montgomery to produce an expected result of securing access. The modification would be obvious because one of ordinary skill in the art would be motivated to access data from third-party applications (social media) securely using OAuth protocol(Andriukhin, para 0045).
But Bleicher, Montgomery and Andriukhin don’t explicitly teach wherein the matching comprises analyzing audience demographics, style preferences, physical congruence and conversion behavior using vector similarity scores weighted by contextual relevance factors: wherein the system utilizes a hybrid engine combining collaborative filtering such as SVD matrix factorization and content-based filtering approaches deployed on enterprise- grade AI infrastructure; wherein semantic similarity is computed using Sentence-BERT embeddings stored and queried through Milvus vector databases; wherein behavioral and interactional data are continuously learned and adapted through real-time feedback loops to dynamically improve match precision.
generate AI marketing personas analyzing social media metadata using deep learning to calculate SHAP value outputs; creating targeted demographic personas with specific income characteristics relative to product pricing;
However, in the same field of endeavor of user profile creation Montgomery teaches wherein the matching comprises analyzing audience demographics, style preferences, physical congruence and conversion behavior using vector similarity scores weighted by contextual relevance factors: wherein the system utilizes a hybrid engine combining collaborative filtering such as SVD matrix factorization and content-based filtering approaches deployed on enterprise- grade AI infrastructure; wherein semantic similarity is computed using Sentence-BERT(Seyeditabari, col 12:8-16 discloses using BERT for contextual/semantic similarity measurement along with SVD (Singular Value Decomposition) principal “BERT having any number of layers, each of which may be configured to learn different contextual information from the respective representations. Alternatively, the model 485 may be any principal component analysis; singular value decomposition; deep learning system; nearest neighbor method or analysis; factorization method or analysis; generative model; gradient boosted decision tree; support vector machine; similarity measure”)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of measuring similarity of Seyeditabari into feature of extracting user characteristics from image of Bleicher, Montgomery and Andriukhin to produce an expected result of matching with related contents. The modification would be obvious because one of ordinary skill in the art would be motivated to accurately predict the match accurately using trained machine learning model(Seyeditabari, abstract).
But, Bleicher, Montogomery, Andriukhin and Seyeditabari don’t explicitly teach embeddings stored and queried through Milvus vector databases; wherein behavioral and interactional data are continuously learned and adapted through real-time feedback loops to dynamically improve match precision. generate AI marketing personas analyzing social media metadata using deep learning to calculate SHAP value outputs; creating targeted demographic personas with specific income characteristics relative to product pricing;
However, in the same field of endeavor of persona/avatar creation Crabtree teaches embeddings stored and queried through Milvus vector databases; wherein behavioral and interactional data are continuously learned and adapted through real-time feedback loops to dynamically improve match precision(Crabtree, para 0148 discloses using any vector database for storing embeddings “Each state can be represented as a high-dimensional vector embedding that captures the relevant features and metrics of the simulation at a given point in time. These state vectors can be stored in a vector database, enabling fast similarity search and retrieval of states based on their vector representations”). generate AI marketing personas analyzing social media metadata using deep learning to calculate SHAP value outputs(Crabtree, claim 3 further discloses generation of personas using generative AI “wherein the one or more hardware processors are further configured for generating and validating simulation scenarios using probabilistic programming languages, rules, agent policies and personas, population generation, generative models, such as Generative Adversarial Networks (GANs)”); creating targeted demographic personas with specific income characteristics relative to product pricing(Crabtree, para 0114 discloses using SHAP to AI models for output generation “Feature importance analysis can be applied to these AI models to understand which features contribute the most to their predictions or decisions. Techniques like permutation importance, SHAP (SHapley Additive explanations), or LIME (Local Interpretable Model-agnostic Explanations) can be used to quantify the importance of each feature”; where prior art Montogomery teaches user attributes such as demographical information);
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of using vector database of Crabtree into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin and Seyeditabari to produce an expected result of generating avatar/persona using user characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to make the similarity searches faster by storing vectors in vector database (Crabtree, para 0148).
Regarding Claim 18 (New), Bleicher, Montogomery, Andriukhin, Seyeditabari and Crabtree teach all the limitations of claim 1, and Montogomery further teaches wherein matching includes identifying influencers whose audiences demonstrate high conversion likelihood based on tracked engagement behavior and historical purchase activity(Montgomery, para 0030 facilitating shopping/commercial activity such as conversion (activities end up in purchase) “When a viewer (e.g., via a viewer device 110) downloads content including the tracking pixel, the tracking pixel causes the viewer device 110 to direct a cookie request to the targeting server 102 or another suitable server”; para 0033 further disclose conversion data “ Event data may describe things that the viewer does after placement of the cookie including, for example, being served an advertisement or other content, clicking through the advertisement or other content, converting the advertisement or other content, etc. Time of day and day of week data may indicate the times and dates at which the viewer performs the various activities described herein”),
Bleicher further teaches in combination with physical congruence and profile similarity(Bleicher, para 0088 discloses matching products/influencers based on user profile which has physical measurement data along with historical and preference data “At step 730 a matchmaking of the user glasses shopping profile to the glasses …..the advanced matching up of appropriate products, in accordance with the specific user's personal shopping profile and preferences”);
Regarding Claim 19 (New), Bleicher, Montogomery, Andriukhin, Seyeditabari and Crabtree teach all the limitations of claim 1, and Andriukhin further teaches wherein integrating social media data through OAuth protocols comprises: implementing OAuth 2.0 authorization framework for secure API integration with external social media platforms; utilizing OAuth 2.0 access tokens to authenticate and authorize data retrieval requests without exposing user credentials; maintaining secure API connections through OAuth 2.0 token refresh mechanisms to ensure continuous data synchronization (Andriukhin, para 0045 discloses integrating or interfacing via secure APIs employing OAuth protocol to external system (social media) “integration Layer 291 allows system 200 to interface with external systems via secure APIs, employing protocols like OAuth 2.0 for secure access”);
Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Crabtree, Jason et al (PGPUB Document No. 20240348663), hereafter, referred to as “Crabtree”, in further view of Rajaram; Santoash et al (PGPUB Document No. 20250124044), hereafter, referred to as “Rajaram”.
Regarding Claim 20 (New), Bleicher, Montogomery, Andriukhin, Seyeditabari and Crabtree teach all the limitations of claim 1 but don’t explicitly teach wherein the hybrid engine combining collaborative filtering and content-based filtering comprises: implementing TensorFlow machine learning models to process social media content and extract user behavioral patterns; utilizing Sentence-BERT transformer models to generate semantic embeddings for text-based content analysis; coordinating TensorFlow and Sentence-BERT processing pipelines to create unified user preference vectors that incorporate both behavioral and semantic similarity metrics;.
However, in the same field of endeavor of user activity analysis Rajaram teaches wherein the hybrid engine combining collaborative filtering and content-based filtering comprises: implementing TensorFlow machine learning models to process social media content and extract user behavioral patterns(Rajaram, para 0093 discloses using TensorFlow platform to analyze and identify and analyze user data (which can similarly be applied for social media data to identify certain type of data) “a context matching algorithm widget may use LLMs, Machine Learning libraries (e.g., scikit-learn, TensorFlow), and Vector databases (e.g., Faiss, Annoy, Pinecone) to identify information relevant to identify the document processed by an AI assistant 137……”); utilizing Sentence-BERT transformer models to generate semantic embeddings for text-based content analysis; coordinating TensorFlow and Sentence-BERT processing pipelines to create unified user preference vectors that incorporate both behavioral and semantic similarity metrics(Rajaram, para 0040 discloses similarity is being calculated for BERT embeddings/vector “Behavior Instruction is implemented by converting each defined situation in the behavior instructions into a vector embedding using a pre-trained language model (e.g., BERT, a Robustly Optimized BERT Pretraining Approach (ROBERTa), or a fine-tuned domain-specific model). These situation embeddings are stored along with their corresponding desired responses in the AssistantConfig. Then, when processing a user query, the user query is converted into a vector embedding using the same embedding model. The cosine similarity between the query embedding and all stored situation embeddings is computed. Then, if the highest similarity score exceeds a predetermined threshold (e.g., 0.85), the corresponding desired response is retrieved”; where para 0093 discloses a vector database for similarity calculation “Vector databases (e.g., Faiss, Annoy, Pinecone) to identify information relevant to identify the document processed by an AI assistant 137”),
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of using vector database of Rajaram into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin, Seyeditabari and Crabtree to produce an expected result of generating avatar/persona using user characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to make the similarity searches faster by storing vectors in vector database (Rajaram, para 0093).
Claim 2 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Crabtree, Jason et al (PGPUB Document No. 20240348663), hereafter, referred to as “Crabtree”, in further view of Courtier-Dutton, David et al (PGPUB Document No. 20140122504), hereafter, referred to as “Courtier-Dutton”, in further view of Khanwalkar, Manoj et al (US Patent No. 10091312), hereafter, referred to as “Khanwalkar”.
Regarding claim 2 (Original), ), Bleicher, Montogomery, Andriukhin, Seyeditabari and Crabtree teach all the limitations of claim 1 but they don’t explicitly teach wherein the processing system is further configured to: implement a double-blind review mechanism between businesses and influencers; track performance metrics of marketing campaigns; refine matching algorithms based on collected performance data.
However in the same field of endeavor of users’ feedback consideration Courtier-Dutton teaches wherein the processing system is further configured to: implement a double-blind review mechanism between businesses and influencers(Courtier-Dutton, para 0036 discloses utilizing double-blind method for review between two parties (business and influencers) “The review collection process is preferably double blind; the reviewers cannot choose which items they review and similarly the users requesting market research cannot choose which reviewers are presented with their media for consideration. The media is presented anonymously so as not to influence the opinion of the reviewer. ….)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of using double-blind reviewing method of Courtier-Dutton into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin, Seyeditabari and Crabtree to produce an expected result of generating avatar/persona using user characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to minimize opinion bias by implementing double-blind review mechanism (Courtier-Dutton, para 0036).
But, Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Courtier-Dutton don’t explicitly teach track performance metrics of marketing campaigns; refine matching algorithms based on collected performance data.
However in the same field of endeavor of tracking marketing data Khanwalkar teaches track performance metrics of marketing campaigns(Khanwalkar, col 5: 51-55 discloses tracking of marketing data “track the performance of marketing programs holistically across multiple digital touch points, which can help contribute to a more meaningful, accurate measurement of marketing spend (for example, by improving conversion tracking from mobile applications to web browsers; by increasing attribution windows; by enabling or improving view through attribution and/or cross-device attribution; and so on”); refine matching algorithms based on collected performance data (Khanwalkar, col 12:12-15 discloses refining/updating matching algorithm “The probabilistic matching algorithm may then be updated to test various combinations of probabilistic rules that may increase confidence levels in matching to unique devices”)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of tracking marketing performance of Khanwalkar into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Courtier-Dutton to produce an expected result of improving the marketing the campaign. The modification would be obvious because one of ordinary skill in the art would be motivated to improve the marketing conversion by implementing accurate user device identification (Khanwalkar, col 5: 51-55).
Claim 3 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Crabtree, Jason et al (PGPUB Document No. 20240348663), hereafter, referred to as “Crabtree”, in view of Bradski, Gary et al (PGPUB Document No. 20160026253), hereafter, referred to as “Bradski”, in further view of Levinson, David (US Patent No. US 9652809), hereafter, referred to as “Levinson”.
Regarding claim 3 (Original), Bleicher, Montogomery, Andriukhin, Seyeditabari and Crabtree teach all the limitations of claim 1 and Crabtree further teaches wherein generating Al marketing personas comprises: creating Al characters with specific income characteristics(Crabtree, para 0114 discloses using SHAP to AI models for output generation “Feature importance analysis can be applied to these AI models to understand which features contribute the most to their predictions or decisions. Techniques like permutation importance, SHAP (SHapley Additive explanations), or LIME (Local Interpretable Model-agnostic Explanations) can be used to quantify the importance of each feature”; where prior art Montogomery teaches user attributes such as demographical information);
But Bleicher, Montogomery, Andriukhin, Seyeditabari and Crabtree don’t explicitly teach managing Al character deployment through a puppet master role; matching Al characters to primary purchasers based on shared characteristics.
However, in the same field of endeavor of avatar generation Bradski teaches managing Al character deployment through a puppet master role (Bradski, in para 1273 teaches AI character/avatar rendered as puppet-master role “the system may generate an avatar that may lead the user through a variety of options. In one or more embodiments, the avatar may be a representation of the user. In essence, the user may be rendered as a “puppet master” and the user avatar of the AR system”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of managing avatar as a puppet-master role of Bradski into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin, Seyeditabari and Crabtree to produce an expected result of representing users with avatars. The modification would be obvious because one of ordinary skill in the art would be motivated to improve user interactions by storing interaction data (Bradski, para 0702).
But Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Bradski don’t explicitly teach matching Al characters to primary purchasers based on shared characteristics.
However, in the same field of endeavor of avatar generation Levinson teaches matching Al characters to primary purchasers based on shared characteristics(Levinson, col 1:40-41 teaches matching of avatar based on user/purchasers characteristics “The identified avatars may be presented based on how closely each identified avatar matches the accessed user profile information”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of selecting avatar based on characteristics of Levinson into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Bradski to produce an expected result of representing users with avatars. The modification would be obvious because one of ordinary skill in the art would be motivated to accurately represent users via avatar which considers personality or emotional state of the users (Levinson, col 9: 36-39).
Claim 11 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Crabtree, Jason et al (PGPUB Document No. 20240348663), hereafter, referred to as “Crabtree”, in further view of Nadig, Vinaya et al (PGPUB Document No. 12211497), hereafter, referred to as “Nadig”.
Regarding Claim 11 (Previously Presented), Bleicher, Montogomery, Andriukhin, Seyeditabari and Crabtree teach all the limitations of claim 1, and Bleicher further teaches wherein the processing system is further configured to: receive individual user profiles including body scan data (Bleicher, para 0063 discloses scanning of users by smartphone for receiving physical attributes of users “In some embodiments a dedicated or generic application on a smart phone, tablet or other computing device may be used to enable effective photographing or scanning of a user”; where para 0087 further discloses physical measurement data is being used for user profile generation);
But Bleicher, Montogomery, Andriukhin, Seyeditabari and Crabtree don’t explicitly teach aggregate multiple individual profiles into a group profile; analyze collective data to identify shared interests and preferences; implement administrator controls for group access.
However, in the same field of endeavor of user profile generation Nadig teaches aggregate multiple individual profiles into a group profile (Nadig, col 11:36-40 discloses a group profile associated with multiple individual user profile “A group profile may be specific to a group of users. That is, a group profile may be associated with two or more individual user profiles”); analyze collective data to identify shared interests and preferences (Nadig, col 11:41-43 discloses a shared preference for group profile “A group profile may include preferences shared by all the user profiles associated therewith”); implement administrator controls for group access (Nadig, col 18:47-51 discloses privacy/access control for group and users “to maintain user privacy, the first content publisher 135a may not have access to a user identifier and/or group identifier stored in the profile storage 270. In these embodiments, the metadata may include an identifier that uniquely corresponds to a particular user identifier and/or group identifier stored in the profile storage 270”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of generating marketing persona of Nadig into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin, Seyeditabari and Crabtree to produce an expected result of generating avatar/persona using user characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to improve the user experiences by controlling receiving of same contents from different sources (Nadig, col 2: 61-65).
Regarding Claim 17(Previously Presented), Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig teach all the limitations of claim 11, and Nadig further teaches wherein the processing system is further configured to:integrate with external platforms via secure API connections; maintain privacy controls for shared data; implement OAuth protocols for data access delegation(Evgenni, para 0045 discloses integrating or interfacing via secure APIs employing OAuth protocol “integration Layer 291 allows system 200 to interface with external systems via secure APIs, employing protocols like OAuth 2.0 for secure access”; where Nadig, col 18:47-51 discloses privacy/access control for group and users “to maintain user privacy, the first content publisher 135a may not have access to a user identifier and/or group identifier stored in the profile storage 270. In these embodiments, the metadata may include an identifier that uniquely corresponds to a particular user identifier and/or group identifier stored in the profile storage 270”).
Claim 12 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Crabtree, Jason et al (PGPUB Document No. 20240348663), hereafter, referred to as “Crabtree”, in further view of Nadig, Vinaya et al (PGPUB Document No. 12211497), hereafter, referred to as “Nadig”, in view of Carmel, David et al (PGPUB Document No. 20140337129), hereafter, referred to as “Carmel”, in further view of Kim, Yongsung (US Patent No. 9116918), hereafter, referred to as “Kim”.
Regarding Claim 12(Currently Amended), Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig teach all the limitations of claim 11, but don’t explicitly teach wherein analyzing collective data comprises: calculating preference weights between individual and group attributes, applying proportional multipliers based on the relative strength of physical, behavioral, or declared style attributes, dynamically adjusting weights based on engagement metrics and group activity over time.
However, in the same field of endeavor of weighing individual and group preferences Carmel teaches wherein analyzing collective data comprises: calculating preference weights between individual and group attributes (Carmel, para 0012 discloses designating/identifying primary purchaser in a group “Once an individual's preferences (e.g., calculated as weight) for one or more features for the collection of items are determined, then the determined preferences are compared to the preferences of a group of other individuals……..”);
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of measuring individual’s preference compared to a group of Carmel into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig to produce an expected result of delivering content based on preferences. The modification would be obvious because one of ordinary skill in the art would be motivated to recommend items which are preferred to user’s(Carmel, para 0033).
But Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree, Nadig and Carmel don’t explicitly teach applying proportional multipliers based on the relative strength of physical, behavioral, or declared style attributes, dynamically adjusting weights based on engagement metrics and group activity over time.
However, in the same field of endeavor of applying weighing multipliers Kim teaches applying proportional multipliers based on the relative strength of physical, behavioral, or declared style attributes (Kim, col 14:11-15 discloses applying weight multipliers “the query interpretation application can assign a low penalty score or not assign a penalty weight (e.g., a 1.0 multiplier). If half of the terms within the voice recognition term are recognized as entities, the query interpretation application can assign a penalty score or penalty weight, such as a 0.5 multiplier ……..”); dynamically adjusting weights based on engagement metrics and group activity over time (Kim, col 14:11-12 further discloses adjusting or modifying weight “the query interpretation application can provide penalty scores or penalty weights that modify the feasibility score”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of measuring individual’s preference compared to a group of Kim into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree, Nadig and Carmel to produce an expected result of identifying closely matched terms. The modification would be obvious because one of ordinary skill in the art would be motivated to recommend items which are closely matched by terms using proper matching weights(Kim, col 14:11-15).
Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Crabtree, Jason et al (PGPUB Document No. 20240348663), hereafter, referred to as “Crabtree”, in further view of Nadig, Vinaya et al (PGPUB Document No. 12211497), hereafter, referred to as “Nadig”, in view of Carmel, David et al (PGPUB Document No. 20140337129), hereafter, referred to as “Carmel”.
Regarding Claim 13(Currently Amended), Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig teach all the limitations of claim 11, and Nadig further teaches track group-wide engagement metrics(Nadig, col 39:12-15 discloses tracking or keeping history of purchase activities “The other data 815 may also include data representing previously provided user feedback indicating the appropriateness of previously output inferred content. The other data 815 may also include social media data associated with the user, system usage history associated with the user, a history of music listened to by the user, a history of books purchased by the user, a general purchasing history of the user”); refine recommendations through a continuous learning feedback loop (Nadig, col 38:51-57 discloses recommending/outputting based on interaction data “as users interacting with devices 110 located in different geographic locations may be susceptible to receiving different amounts and/or kinds of inferred content. If the supplemental content system 130 determines the geographic location of the device 110 indicates inferred content should be output, the supplemental content system 130 may generate request data 845. If the supplemental content system 130 determines the geographic location of the device 110 indicates inferred content should not be output”).
But Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig don’t explicitly teach deliver personalized content to group members based on weighted preferences using vector similarity scores;
However, in the same field of endeavor of weighing individual and group preferences Carmel teaches deliver personalized content to group members based on weighted preferences using vector similarity scores (Carmel, para 0033 discloses recommending delivery based on preference “items that satisfy the user unique preferences are recommended for delivery…..”);
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of measuring individual’s preference compared to a group of Carmel into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig to produce an expected result of delivering content based on preferences. The modification would be obvious because one of ordinary skill in the art would be motivated to recommend items which are preferred to user’s(Carmel, para 0033).
Claim 14 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Crabtree, Jason et al (PGPUB Document No. 20240348663), hereafter, referred to as “Crabtree”, in view of Nadig, Vinaya et al (PGPUB Document No. 12211497), hereafter, referred to as “Nadig”, in further view of Allen, Carl et al (PGPUB Document No. 20120110011), hereafter, referred to as “Allen”.
Regarding Claim 14(Currently Amended), Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig teach all the limitations of claim 11, and Melcher further teaches wherein implementing administrator controls comprises: replacing physical characteristics with fictional avatars for selected members to preserve privacy or reflect user-selected identity; controlling access to group-wide content based on permissions (Nadig, col 18:47-51 discloses privacy/access control for group and users “to maintain user privacy, the first content publisher 135a may not have access to a user identifier and/or group identifier stored in the profile storage 270. In these embodiments, the metadata may include an identifier that uniquely corresponds to a particular user identifier and/or group identifier stored in the profile storage 270”; where Crabtree teaches avatar/persona with characteristics).
But Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig don’t explicitly managing connection requests through administrator approval processes.
However, in the same field of endeavor of user access control Allen teaches managing connection requests through administrator approval processes (Allen, para 0047 discloses managing access/connection request by administrator approval “a user may request access to application A. The access management application may receive this request and send a request for approval to an administrator”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of managing requests by administrator approval of Allen into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig to produce an expected result of controlling user access. The modification would be obvious because one of ordinary skill in the art would be motivated to improve access security by introducing administrator approval requirement(Allen, para 0047).
Claim 15 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Crabtree, Jason et al (PGPUB Document No. 20240348663), hereafter, referred to as “Crabtree”, in view of Nadig, Vinaya et al (PGPUB Document No. 12211497), hereafter, referred to as “Nadig”, in further view of Todd, Jason et al (PGPUB Document No. 20170132690), hereafter, referred to as “Todd”,
Regarding Claim 15 (Currently Amended), Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig teach all the limitations of claim 11 and Nadiq further teaches Nadig teaches aggregate shopping preferences of group members (Nadig, col 11:41-43 discloses aggregation of any preferences(i.e. shopping etc.) “A group profile may include preferences shared by all the user profiles associated therewith. Each user profile associated with a group profile may additionally include preferences specific to the user associated therewith”;); track commercial activity across the group in real time to support dynamic personalization (Montgomery, para 0030 facilitating shopping/commercial activity tracking such as conversion (activities end up in purchase) “When a viewer (e.g., via a viewer device 110) downloads content including the tracking pixel, the tracking pixel causes the viewer device 110 to direct a cookie request to the targeting server 102 or another suitable server”; para 0033 further disclose conversion data “ Event data may describe things that the viewer does after placement of the cookie including, for example, being served an advertisement or other content, clicking through the advertisement or other content, converting the advertisement or other content, etc. Time of day and day of week data may indicate the times and dates at which the viewer performs the various activities described herein”).
But Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig don’t explicitly teach wherein the processing system is further configured to: designate a primary purchaser for the group.
However, in the same field of endeavor of group shopping Todd teaches wherein the processing system is further configured to: designate a primary purchaser for the group (Todd, para 0012 discloses designating/identifying primary purchaser in a group “the group shopping server 101 provides customer identity services to identify at least a primary shopper (e.g. customer 1) and members in the social circle 120”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of designating/identifying primary purchaser within a group of Todd into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig to produce an expected result of facilitating shopping experience. The modification would be obvious because one of ordinary skill in the art would be motivated to facilitate group shopping experience(Todd, para 0012).
Claim 16 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Crabtree, Jason et al (PGPUB Document No. 20240348663), hereafter, referred to as “Crabtree”, in view of Nadig, Vinaya et al (PGPUB Document No. 12211497), hereafter, referred to as “Nadig”, in further view of Golightly, Michael et al (PGPUB Document No. 20240012847), hereafter, referred to as “Golightly”.
Regarding Claim 16(Currently Amended), Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig teach all the limitations of claim 11, and Bleicher further teaches wherein analyzing collective data comprises: monitoring social media activity across group members(Nadig, col 39:12-15 discloses tracking/monitoring by keeping users’ activity history “The other data 815 may also include data representing previously provided user feedback indicating the appropriateness of previously output inferred content. The other data 815 may also include social media data associated with the user, system usage history associated with the user, a history of music listened to by the user, a history of books purchased by the user, a general purchasing history of the user”); updating group preferences based on detected changes (Nadig, col 38:51-57 discloses recommending/outputting based on detected changes “as users interacting with devices 110 located in different geographic locations may be susceptible to receiving different amounts and/or kinds of inferred content. If the supplemental content system 130 determines the geographic location of the device 110 indicates inferred content should be output, the supplemental content system 130 may generate request data 845. If the supplemental content system 130 determines the geographic location of the device 110 indicates inferred content should not be output”).
But Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig don’t explicitly teach applying negative weights to suppress non-aligned content;
However, in the same field of endeavor of suggesting personalized contents Golightly teaches apply negative weights to suppress non-aligned content (Golightly, para 0083 discloses designating/identifying primary purchaser in a group “each media item or interaction with the media item is assigned a positive or negative weight, wherein positive represents the user liking the media item and negative represents the user disliking the media item ……..”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of assigning negative weight to contents of Golightly into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin, Seyeditabari, Crabtree and Nadig to produce an expected result of eliminating non-aligned contents from suggestion consideration. The modification would be obvious because one of ordinary skill in the art would be motivated to provide user’s favorite/preferred contents higher precedence(Golightly, para 0083).
Claim 4 and 21 are rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Rajaram; Santoash et al (PGPUB Document No. 20250124044), hereafter, referred to as “Rajaram”.
Regarding claim 4 (Currently Amended), Bleicher A method for facilitating digital marketing interactions, comprising (Bleicher, abstract discloses a method for marketing using shopping avatar “system, and method to facilitate personalized online shopping……. personalized shopping assistant application to match one or more products from the one or more product data sources to one or more personalized user shopping avatars”): receiving body scan images from a user's smartphone (Bleicher, para 0063 discloses scanning of users by smartphone for receiving physical attributes of users “In some embodiments a dedicated or generic application on a smart phone, tablet or other computing device may be used to enable effective photographing or scanning of a user” ); processing the images to extract physical measurements by. generating a volume-metric reflectance map of the user's body; calculating precise body contours using computer vision algorithms to identify multiple points of interest including shoulders, elbows, wrists, and feet; creating a three-dimensional digital model indicating volume and depth measurements of each region, wherein the three-dimensional digital model incorporates shadow analysis to determine body curvature (Bleicher, para 0086 discloses extracting anatomical attributes of users using machine learning using 3D analysis “scanned or graphic data may be acquired for a user, for example, from standard photographs, 2D and/or 3D image scanning etc…….computer vision and/or machine learning algorithms may be applied, for example, to identify the Eyes, nose, nose bridge, temples, ears etc” );
matching the user with relevant influencers and associated content based on the profile by(Bleicher, para 0088 discloses matching products/influencers based on user profile which has physical measurement data along with historical and preference data “At step 730 a matchmaking of the user glasses shopping profile to the glasses …..the advanced matching up of appropriate products, in accordance with the specific user's personal shopping profile and preferences”),
But Bleicher does not explicitly teach receiving user-submitted preference data via an onboarding questionnaire, the data including style categories, values, and content interests; creating a user profile incorporating the physical measurements, user-submitted preference data, and additional user data by:
synchronizing the extracted measurements with social platform data through OAuth protocols; implementing pixel tracking to capture user engagement metrics; analyzing social media activity using TensorFlow models to identify style preferences; calculating vector similarity scores using Sentence-BERT embeddings stored in a Milvus vector database, the scores weighted by contextual relevance factors, including physical congruence and behavioral similarity;
facilitating commercial transactions based on the matches through: integrating with external commerce platforms via secure APIs;
tracking conversion events across user interactions;
wherein the matching incorporates a hybrid engine comprising collaborative filtering such as SVD matrix factorization and content-based filtering techniques to refine match accuracy in real time;
However, in the same field of endeavor of user profile creation Montgomery teaches receiving user-submitted preference data via an onboarding questionnaire, the data including style categories, values, and content interests(Montgomery, para 0040 disclose collecting data from questionnaire/surveys “the surveys include questions that gather demographic information about the surveyed individual as well as information about elements that are of interest to the individual. Demographic questions may include, for example, questions regarding age, gender, zip code, income level, education level, etc. Element questions may include questions about what types of movies, sports and other recreational activities are of interest to the individual, what types of products or services the individual purchases, etc. Market research data 703 may include data collected about individual consumers”);
creating a user profile incorporating the physical measurements, user-submitted preference data, and additional user data by(Montgomery, Fig. 4 and para 0033 disclose creation of user profile using user preference and social media data (any tracking data) “profile generation illustrated in the process flow 300, including examples of data types. Data inputs box 402 shows example viewer data that may be provided to generate viewer profiles. This data may include the data received from tracking viewers as described above. For example, viewer element data may indicate elements that a viewer "likes" or otherwise indicates an interest. Offline zip or zip code-based data may indicate general demographic data describing people who live in the same geographic area as a viewer (e.g., as derived from the viewers network address). Site data may describe web sites that the viewer visits after placement of the cookie, as described herein”; where prior art Andriukhin to be discussed next discloses accessing data via OAuth protocols); implementing pixel tracking to capture user engagement metrics(Montgomery, para 0030 discloses implementing pixel tracking (can be used for social media) and conversion “When a viewer (e.g., via a viewer device 110) downloads content including the tracking pixel, the tracking pixel causes the viewer device 110 to direct a cookie request to the targeting server 102 or another suitable server”; para 0033 further disclose conversion data “ Event data may describe things that the viewer does after placement of the cookie including, for example, being served an advertisement or other content, clicking through the advertisement or other content, converting the advertisement or other content, etc. Time of day and day of week data may indicate the times and dates at which the viewer performs the various activities described herein”); facilitating commercial transactions based on the matches through: integrating with external commerce platforms via secure APIs; tracking conversion events across user interactions (Montgomery, para 0030 facilitating shopping/commercial activity such as conversion (activities end up in purchase) “When a viewer (e.g., via a viewer device 110) downloads content including the tracking pixel, the tracking pixel causes the viewer device 110 to direct a cookie request to the targeting server 102 or another suitable server”; para 0033 further disclose conversion data “ Event data may describe things that the viewer does after placement of the cookie including, for example, being served an advertisement or other content, clicking through the advertisement or other content, converting the advertisement or other content, etc. Time of day and day of week data may indicate the times and dates at which the viewer performs the various activities described herein”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of tracking and extracting user activities data of Montgomery into feature of extracting user characteristics from image of Bleicher and Melcher to produce an expected result of profiling users based on user characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to track user activities and provide contents accordingly (Montgomery, para 0030).
But Bleicher and Montgomery don’t explicitly teach synchronizing the extracted measurements with social platform data through OAuth protocols; analyzing social media activity using TensorFlow models to identify style preferences; calculating vector similarity scores using Sentence-BERT embeddings stored in a Milvus vector database, the scores weighted by contextual relevance factors, including physical congruence and behavioral similarity; wherein the matching incorporates a hybrid engine comprising collaborative filtering such as SVD matrix factorization and content-based filtering techniques to refine match accuracy in real time;
However, in the same field of endeavor of securing access Andriukhin teaches synchronizing the extracted measurements with social platform data through OAuth protocols(Andriukhin, para 0045 discloses integrating or interfacing via secure APIs employing OAuth protocol “integration Layer 291 allows system 200 to interface with external systems via secure APIs, employing protocols like OAuth 2.0 for secure access”);
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of securing access of Andriukhin into feature of extracting user characteristics from image of Bleicher and Montgomery to produce an expected result of securing access. The modification would be obvious because one of ordinary skill in the art would be motivated to access data from third-party applications (social media) securely using OAuth protocol(Andriukhin, para 0045).
But Bleicher, Montgomery and Andriukhin don’t explicitly teach analyzing social media activity using TensorFlow models to identify style preferences;
calculating vector similarity scores using Sentence-BERT embeddings stored in a Milvus vector database, the scores weighted by contextual relevance factors, including physical congruence and behavioral similarity; wherein the matching incorporates a hybrid engine comprising collaborative filtering such as SVD matrix factorization and content-based filtering techniques to refine match accuracy in real time.
However, in the same field of endeavor of user profile creation Montgomery teaches the scores weighted by contextual relevance factors, including physical congruence and behavioral similarity; wherein the matching incorporates a hybrid engine comprising collaborative filtering such as SVD matrix factorization and content-based filtering techniques to refine match accuracy in real time(Seyeditabari, col 12:8-16 discloses using BERT for contextual/semantic similarity measurement along with SVD (Singular Value Decomposition) principal “BERT having any number of layers, each of which may be configured to learn different contextual information from the respective representations. Alternatively, the model 485 may be any principal component analysis; singular value decomposition; deep learning system; nearest neighbor method or analysis; factorization method or analysis; generative model; gradient boosted decision tree; support vector machine; similarity measure”)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of measuring similarity of Seyeditabari into feature of extracting user characteristics from image of Bleicher, Montgomery and Andriukhin to produce an expected result of matching with related contents. The modification would be obvious because one of ordinary skill in the art would be motivated to accurately predict the match accurately using trained machine learning model(Seyeditabari, abstract).
But, Bleicher, Montogomery, Andriukhin and Seyeditabari don’t explicitly teach analyzing social media activity using TensorFlow models to identify style preferences; calculating vector similarity scores using Sentence-BERT embeddings stored in a Milvus vector database,
However, in the same field of endeavor of user activity analysis Rajaram teaches analyzing social media activity using TensorFlow models to identify style preferences(Rajaram, para 0093 discloses using TensorFlow platform to analyze and identify and analyze user data (which can similarly be applied for social media data to identify certain type of data) “a context matching algorithm widget may use LLMs, Machine Learning libraries (e.g., scikit-learn, TensorFlow), and Vector databases (e.g., Faiss, Annoy, Pinecone) to identify information relevant to identify the document processed by an AI assistant 137……”); calculating vector similarity scores using Sentence-BERT embeddings stored in a Milvus vector database(Rajaram, para 0040 discloses similarity is being calculated for BERT embeddings/vector “Behavior Instruction is implemented by converting each defined situation in the behavior instructions into a vector embedding using a pre-trained language model (e.g., BERT, a Robustly Optimized BERT Pretraining Approach (ROBERTa), or a fine-tuned domain-specific model). These situation embeddings are stored along with their corresponding desired responses in the AssistantConfig. Then, when processing a user query, the user query is converted into a vector embedding using the same embedding model. The cosine similarity between the query embedding and all stored situation embeddings is computed. Then, if the highest similarity score exceeds a predetermined threshold (e.g., 0.85), the corresponding desired response is retrieved”; where para 0093 discloses a vector database for similarity calculation “Vector databases (e.g., Faiss, Annoy, Pinecone) to identify information relevant to identify the document processed by an AI assistant 137”),
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of using vector database of Rajaram into feature of extracting user characteristics from image of Bleicher, Montogomery, Andriukhin and Seyeditabari to produce an expected result of generating avatar/persona using user characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to make the similarity searches faster by storing vectors in vector database (Rajaram, para 0093).
Regarding Claim 21(New), Bleicher, Montogomery, Andriukhin, Seyeditabari and Rajaram teach all the limitations of claim 4, and Rajaram further teaches wherein analyzing social media activity using TensorFlow models comprises: deploying TensorFlow neural networks trained on fashion and style classification datasets; processing image and text content through TensorFlow convolutional and recurrent neural network architectures(Rajaram, para 0093 discloses using TensorFlow platform to analyze and identify and analyze user dataset (which can similarly be applied for social media data to identify certain type of data) “a context matching algorithm widget may use LLMs, Machine Learning libraries (e.g., scikit-learn, TensorFlow), and Vector databases (e.g., Faiss, Annoy, Pinecone) to identify information relevant to identify the document processed by an AI assistant 137……”); generating style preference vectors through TensorFlow model inference that are subsequently encoded using Sentence-BERT for semantic similarity matching (Rajaram, para 0040 discloses similarity is being calculated for BERT embeddings/vector “Behavior Instruction is implemented by converting each defined situation in the behavior instructions into a vector embedding using a pre-trained language model (e.g., BERT, a Robustly Optimized BERT Pretraining Approach (ROBERTa), or a fine-tuned domain-specific model). These situation embeddings are stored along with their corresponding desired responses in the AssistantConfig. Then, when processing a user query, the user query is converted into a vector embedding using the same embedding model. The cosine similarity between the query embedding and all stored situation embeddings is computed. Then, if the highest similarity score exceeds a predetermined threshold (e.g., 0.85), the corresponding desired response is retrieved”; where para 0093 discloses a vector database for similarity calculation “Vector databases (e.g., Faiss, Annoy, Pinecone) to identify information relevant to identify the document processed by an AI assistant 137”).
Claim 5 and 22 are rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in further view of Nadig, Vinaya et al (PGPUB Document No. 12211497), hereafter, referred to as “Nadig”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in further view of Kim, Yongsung (US Patent No. 9116918), hereafter, referred to as “Kim”, in further view of Hudetz, Casey et al (PGPUB Document No. 20240370479 ), hereafter, referred to as “Hudetz”.
Regarding Claim 5, Bleicher teaches A system for managing group profiles, comprising: a processing system configured to: receive individual user profiles including body scan data (Bleicher, para 0063 discloses scanning of users by smartphone for receiving physical attributes of users “In some embodiments a dedicated or generic application on a smart phone, tablet or other computing device may be used to enable effective photographing or scanning of a user” );
But Bleicher does not explicitly teach and user-submitted preference data collected via onboarding questionnaires; aggregate multiple individual profiles into a group profile by: processing comprehensive body scans for each group member to extract physical measurements; synchronizing individual profile data through secure API integrations;
analyze collective data to identify shared preferences by applying machine learning algorithms to analyze group interaction patterns; calculating preference weights using proportional multipliers based on attribute strength, wherein the preference weights dynamically adjust based on recent group member interactions and purchase history; provide group-based recommendations through: implementing vector similarity algorithms to analyze attribute differences across group members;
matching group members with relevant influencers based on collective physical congruence, shared style preferences, and social behavior patterns using Sentence-BERT embeddings stored in a Milvus vector database; utilizing a hybrid recommendation engine combining collaborative filtering (SVD matrix factorization) and content-based filtering models to optimize content alignment in real time; delivering personalized content based on weighted collective preferences; implement privacy controls for group members by. utilizing OAuth protocols for secure data access delegation; maintaining on-device processing of recommendation insights to limit external exposure.
However, in the same field of endeavor of user profile creation Montgomery teaches and user-submitted preference data collected via onboarding questionnaires (Montgomery, para 0040 disclose collecting data from questionnaire/surveys “the surveys include questions that gather demographic information about the surveyed individual as well as information about elements that are of interest to the individual. Demographic questions may include, for example, questions regarding age, gender, zip code, income level, education level, etc. Element questions may include questions about what types of movies, sports and other recreational activities are of interest to the individual, what types of products or services the individual purchases, etc. Market research data 703 may include data collected about individual consumers”);
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of tracking and extracting user activities data of Montgomery into feature of extracting user characteristics from image of Bleicher to produce an expected result of profiling users based on user characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to track user activities and provide contents accordingly (Montgomery, para 0030).
But Bleicher and Montgomery don’t explicitly teach aggregate multiple individual profiles into a group profile by: processing comprehensive body scans for each group member to extract physical measurements; synchronizing individual profile data through secure API integrations;
analyze collective data to identify shared preferences by applying machine learning algorithms to analyze group interaction patterns; calculating preference weights using proportional multipliers based on attribute strength, wherein the preference weights dynamically adjust based on recent group member interactions and purchase history; provide group-based recommendations through: implementing vector similarity algorithms to analyze attribute differences across group members;
matching group members with relevant influencers based on collective physical congruence, shared style preferences, and social behavior patterns using Sentence-BERT embeddings stored in a Milvus vector database; utilizing a hybrid recommendation engine combining collaborative filtering (SVD matrix factorization) and content-based filtering models to optimize content alignment in real time; delivering personalized content based on weighted collective preferences; implement privacy controls for group members by. utilizing OAuth protocols for secure data access delegation; maintaining on-device processing of recommendation insights to limit external exposure.
But Bleicher and Montgomery don’t explicitly teach aggregate multiple individual profiles into a group profile by: processing comprehensive body scans for each group member to extract physical measurements; synchronizing individual profile data through secure API integrations;
analyze collective data to identify shared preferences by applying machine learning algorithms to analyze group interaction patterns; calculating preference weights using proportional multipliers based on attribute strength, wherein the preference weights dynamically adjust based on recent group member interactions and purchase history;
provide group-based recommendations through: implementing vector similarity algorithms to analyze attribute differences across group members;
matching group members with relevant influencers based on collective physical congruence, shared style preferences, and social behavior patterns using Sentence-BERT embeddings stored in a Milvus vector database; utilizing a hybrid recommendation engine combining collaborative filtering (SVD matrix factorization) and content-based filtering models to optimize content alignment in real time; delivering personalized content based on weighted collective preferences; implement privacy controls for group members by utilizing OAuth protocols for secure data access delegation; maintaining on-device processing of recommendation insights to limit external exposure.
However, in the same field of endeavor of user profile generation Nadig teaches aggregate multiple individual profiles into a group profile by: processing comprehensive body scans for each group member to extract physical measurements(Nadig, col 11:36-40 discloses a group profile associated with multiple individual user profile “A group profile may be specific to a group of users. That is, a group profile may be associated with two or more individual user profiles”; where Bleicher in para 0087 discloses generation of user profile using physical measurement data and additional data such as user’s history and preference data “At step 720 the processed user scanned data may be parameterized, for example, to extract the precise face length, width, height, proportions, nose width and volume….together with the user history data from 705 and user preferences data at 715, to generate a user's glasses shopping profile based on the user's face profile and other physical properties, as well as user behavior and user preference data”); analyze collective data to identify shared preferences by applying machine learning algorithms to analyze group interaction patterns(Nadig, col 11:41-43 discloses a shared preference for group profile “A group profile may include preferences shared by all the user profiles associated therewith”); delivering personalized content based on weighted collective preferences (Nadig, para 0063 discloses delivering/outputting contents based on group/collective preferences “The delivery management component 425 may determine how to indicate the requested content 515 b is available for output based on a user and/or group preference(s) corresponding to the user and/or group identifier associated with the requested content 515 b in the requested content storage 520”);
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of generating marketing persona of Nadig into feature of extracting user characteristics from image of Bleicher and Montgomery to produce an expected result of generating avatar/persona using user characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to improve the user experiences by controlling receiving of same contents from different sources (Nadig, col 2: 61-65).
But Bleicher, Montgomery and Nadig don’t explicitly teach synchronizing individual profile data through secure API integrations; calculating preference weights using proportional multipliers based on attribute strength, wherein the preference weights dynamically adjust based on recent group member interactions and purchase history;
provide group-based recommendations through: implementing vector similarity algorithms to analyze attribute differences across group members;
matching group members with relevant influencers based on collective physical congruence, shared style preferences, and social behavior patterns using Sentence-BERT embeddings stored in a Milvus vector database; utilizing a hybrid recommendation engine combining collaborative filtering (SVD matrix factorization) and content-based filtering models to optimize content alignment in real time; implement privacy controls for group members by. utilizing OAuth protocols for secure data access delegation; maintaining on-device processing of recommendation insights to limit external exposure.
However, in the same field of endeavor of securing access Andriukhin teaches synchronizing individual profile data through secure API integrations; implement privacy controls for group members by. utilizing OAuth protocols for secure data access delegation; maintaining on-device processing of recommendation insights to limit external exposure(Andriukhin, para 0045 discloses integrating or interfacing via secure APIs employing OAuth protocol “integration Layer 291 allows system 200 to interface with external systems via secure APIs, employing protocols like OAuth 2.0 for secure access”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of securing access of Andriukhin into feature of extracting user characteristics from image of Bleicher, Montgomery and Nadig to produce an expected result of securing access. The modification would be obvious because one of ordinary skill in the art would be motivated to access data from third-party applications (social media) securely using OAuth protocol(Andriukhin, para 0045).
But Bleicher, Montgomery, Nadig and Andriukhin don’t explicitly teach calculating preference weights using proportional multipliers based on attribute strength, wherein the preference weights dynamically adjust based on recent group member interactions and purchase history;
provide group-based recommendations through: implementing vector similarity algorithms to analyze attribute differences across group members;
matching group members with relevant influencers based on collective physical congruence, shared style preferences, and social behavior patterns using Sentence-BERT embeddings stored in a Milvus vector database; utilizing a hybrid recommendation engine combining collaborative filtering (SVD matrix factorization) and content-based filtering models to optimize content alignment in real time;
However, in the same field of endeavor of applying weighing multipliers Kim teaches calculating preference weights using proportional multipliers based on attribute strength, wherein the preference weights (Kim, col 14:11-15 discloses applying weight multipliers “the query interpretation application can assign a low penalty score or not assign a penalty weight (e.g., a 1.0 multiplier). If half of the terms within the voice recognition term are recognized as entities, the query interpretation application can assign a penalty score or penalty weight, such as a 0.5 multiplier ……..”); dynamically adjust based on recent group member interactions and purchase history(Kim, col 14:11-12 further discloses adjusting or modifying weight “the query interpretation application can provide penalty scores or penalty weights that modify the feasibility score”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of measuring individual’s preference compared to a group of Kim into feature of extracting user characteristics from image of Bleicher, Montgomery, Nadig and Andriukhin to produce an expected result of identifying closely matched terms. The modification would be obvious because one of ordinary skill in the art would be motivated to recommend items which are closely matched by terms using proper matching weights(Kim, col 14:11-15).
But Bleicher, Montgomery, Nadig, Andriukhin and Kim don’t explicitly teach
provide group-based recommendations through: implementing vector similarity algorithms to analyze attribute differences across group members;
matching group members with relevant influencers based on collective physical congruence, shared style preferences, and social behavior patterns using Sentence-BERT embeddings stored in a Milvus vector database; utilizing a hybrid recommendation engine combining collaborative filtering (SVD matrix factorization) and content-based filtering models to optimize content alignment in real time;
However, in the same field of endeavor of user activity analysis Hudetz teaches
provide group-based recommendations through: implementing vector similarity algorithms to analyze attribute differences across group members;
matching group members with relevant influencers based on collective physical congruence, shared style preferences, and social behavior patterns using Sentence-BERT embeddings(Hudetz, para 0154 teaches matching contents (can similarly be applied for profile matching) similarity is being calculated for BERT embeddings/vector for content identification of suggestion “Elasticsearch also provides a plugin called Elasticsearch Vector Scoring, which enables the use of dense vector embeddings for similarity search. This plugin can be used to index and search documents based on their dense vector embeddings, which can be generated using BERT or other contextualized embedding models. To use Elasticsearch for content-based search with dense vectors, the search model 704 indexes the documents and their embeddings using the Elasticsearch Vector Scoring plugin. The search manager 124 can then search for similar documents by specifying a query embedding and using the cosine similarity as the similarity metric. Elasticsearch will return the top matching documents based on their similarity scores”) stored in a Milvus vector database(Hudetz, para 0142 teaches storing vectors in any vector database “The search manager 124 may store the document vectors 726 in a database 708” ); utilizing a hybrid recommendation engine combining collaborative filtering (SVD matrix factorization) and content-based filtering models to optimize content alignment in real time(Hudetz, para 0226 discloses for similarity measurement SVD (Singular Value Decomposition) principal is being used “ These features are then used to compute a semantic similarity score between the query and each document, which is used to rank the results. One popular example of a semantic ranking algorithm is the Latent Semantic Analysis (LSA) algorithm, which uses singular value decomposition (SVD) to identify latent semantic relationships between words and documents”);
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of measuring similarity of Hudetz into feature of extracting user characteristics from image of Bleicher, Montgomery, Nadig, Andriukhin and Kim to produce an expected result of matching with related contents. The modification would be obvious because one of ordinary skill in the art would be motivated to extracting recommended contents based on computed semantic similarity scores using SVD (Hudetz, para 0226).
Regarding Claim 22 (New), Bleicher, Montgomery, Nadig, Andriukhin, Kim and Hudetz teach all the limitations of claim 5, and Andriukhin further teaches wherein utilizing OAuth protocols for secure data access delegation comprises: implementing OAuth 2.0 client credentials flow for server-to-server authentication; managing OAuth 2.0 refresh token rotation to maintain persistent authorized access to group member social media data; enforcing OAuth 2.0 scope limitations to restrict data access to only profile information necessary for physical characteristic and preference matching (Andriukhin, para 0045 discloses integrating or interfacing via secure APIs employing OAuth protocol to external system (social media) “integration Layer 291 allows system 200 to interface with external systems via secure APIs, employing protocols like OAuth 2.0 for secure access”);
Claim 6 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in further view of Nadig, Vinaya et al (PGPUB Document No. 12211497), hereafter, referred to as “Nadig”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in further view of Kim, Yongsung (US Patent No. 9116918), hereafter, referred to as “Kim”, in further view of Hudetz, Casey et al (PGPUB Document No. 20240370479 ), hereafter, referred to as “Hudetz”, in view of Carmel, David et al (PGPUB Document No. 20140337129), hereafter, referred to as “Carmel”, in further view of Golightly, Michael et al (PGPUB Document No. 20240012847), hereafter, referred to as “Golightly”.
Regarding Claim 6 (Original), Bleicher, Montgomery, Nadig, Andriukhin, Kim and Hudetz teach all the limitations of claim 5, and Nadig further teaches wherein the processing system is further configured to: track group-wide commercial activity (Nadig, col 39:12-15 discloses tracking or keeping history of purchase activities “The other data 815 may also include data representing previously provided user feedback indicating the appropriateness of previously output inferred content. The other data 815 may also include social media data associated with the user, system usage history associated with the user, a history of music listened to by the user, a history of books purchased by the user, a general purchasing history of the user”);
But Bleicher, Montgomery, Nadig, Andriukhin, Kim and Hudetz don’t explicitly teach weight individual versus group preferences; apply negative weights to suppress non-aligned content;
However, in the same field of endeavor of weighing individual and group preferences Carmel teaches weight individual versus group preferences(Carmel, para 0012 discloses designating/identifying primary purchaser in a group “Once an individual's preferences (e.g., calculated as weight) for one or more features for the collection of items are determined, then the determined preferences are compared to the preferences of a group of other individuals……..”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of measuring individual’s preference compared to a group of Carmel into feature of extracting user characteristics from image of Bleicher, Montgomery, Nadig, Andriukhin, Kim and Hudetz to produce an expected result of delivering content based on preferences. The modification would be obvious because one of ordinary skill in the art would be motivated to recommend items which are preferred to user’s(Carmel, para 0033).
But Bleicher, Montgomery, Nadig, Andriukhin, Kim, Hudetz and Carmel don’t explicitly teach apply negative weights to suppress non-aligned content;
However, in the same field of endeavor of suggesting personalized contents Golightly teaches apply negative weights to suppress non-aligned content (Golightly, para 0083 discloses designating/identifying primary purchaser in a group “each media item or interaction with the media item is assigned a positive or negative weight, wherein positive represents the user liking the media item and negative represents the user disliking the media item ……..”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of assigning negative weight to contents of Golightly into feature of extracting user characteristics from image of Bleicher, Montgomery, Nadig, Andriukhin, Kim, Hudetz and Carmel to produce an expected result of eliminating non-aligned contents from suggestion consideration. The modification would be obvious because one of ordinary skill in the art would be motivated to provide user’s favorite/preferred contents higher precedence(Golightly, para 0083).
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in further view of Nadig, Vinaya et al (PGPUB Document No. 12211497), hereafter, referred to as “Nadig”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in further view of Kim, Yongsung (US Patent No. 9116918), hereafter, referred to as “Kim”, in further view of Hudetz, Casey et al (PGPUB Document No. 20240370479 ), hereafter, referred to as “Hudetz”, in further view of Todd, Jason et al (PGPUB Document No. 20170132690), hereafter, referred to as “Todd”.
Regarding Claim 7 (Original), Bleicher, Montgomery, Nadig, Andriukhin, Kim and Hudetz teach all the limitations of claim 5, and Nadig further teaches wherein the processing system is further configured to: aggregate shopping preferences of group members(Nadig, col 11:41-43 discloses a shared/aggregated preference for group profile “A group profile may include preferences shared by all the user profiles associated therewith”); deliver targeted content based on collective preferences (Nadig, para 0063 discloses delivering/outputting contents based on group/collective preferences “The delivery management component 425 may determine how to indicate the requested content 515 b is available for output based on a user and/or group preference(s) corresponding to the user and/or group identifier associated with the requested content 515 b in the requested content storage 520”);
But Bleicher, Montgomery, Nadig, Andriukhin, Kim and Hudetz don’t explicitly teach designate a primary purchaser for the group;.
However, in the same field of endeavor of group shopping Todd teaches designate a primary purchaser for the group (Todd, para 0012 discloses designating/identifying primary purchaser in a group “the group shopping server 101 provides customer identity services to identify at least a primary shopper (e.g. customer 1) and members in the social circle 120”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of designating/identifying primary purchaser within a group of Todd into feature of extracting user characteristics from image of Bleicher, Montgomery, Nadig, Andriukhin, Kim and Hudetz to produce an expected result of facilitating shopping experience. The modification would be obvious because one of ordinary skill in the art would be motivated to facilitate group shopping experience(Todd, para 0012).
Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in further view of Nadig, Vinaya et al (PGPUB Document No. 12211497), hereafter, referred to as “Nadig”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Rajaram; Santoash et al (PGPUB Document No. 20250124044), hereafter, referred to as “Rajaram”, in further view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”.
Regarding claim 8 (Currently Amended), Bleicher A method for managing group profiles, comprising: receiving individual user profiles including body scan data (Bleicher, para 0063 discloses scanning of users by smartphone for receiving physical attributes of users “In some embodiments a dedicated or generic application on a smart phone, tablet or other computing device may be used to enable effective photographing or scanning of a user” ), matching group members with relevant influencers based on collective preferences by(Bleicher, para 0088 discloses matching products/influencers based on user profile which has physical measurement data along with historical and preference data “At step 730 a matchmaking of the user glasses shopping profile to the glasses …..the advanced matching up of appropriate products, in accordance with the specific user's personal shopping profile and preferences”);
But Bleicher does not explicitly teach user-submitted preference data collected via onboarding questionnaires; aggregating multiple profiles into a group profile by processing comprehensive body scans using machine learning algorithms to extract physical measurements; integrating social media data through secure API connections with OAuth protocols; analyzing collective data for shared preferences through: implementing TensorFlow models to analyze group social media activity; calculating attribute weights using vector similarity algorithms wherein the similarity scores incorporate both physical congruence and shared user-declared preferences; matching group members with relevant influencers based on collective preferences by; comparing group and influencer profiles using Sentence-BERT embeddings stored in Milvus vector databases;
applying a hybrid recommendation engine comprising collaborative filtering (SVD matrix factorization) and content-based filtering to optimize influencer alignment; implementing administrator controls for group access by deploying pixel tracking across group member interactions; maintaining secure API connections for external platform integration; providing privacy protections for group members through: processing recommendation insights locally on user devices; implementing data access controls with OAuth delegation protocols.
However, in the same field of endeavor of user profile creation Montgomery teaches user-submitted preference data collected via onboarding questionnaires(Montgomery, para 0040 disclose collecting data from questionnaire/surveys “the surveys include questions that gather demographic information about the surveyed individual as well as information about elements that are of interest to the individual. Demographic questions may include, for example, questions regarding age, gender, zip code, income level, education level, etc. Element questions may include questions about what types of movies, sports and other recreational activities are of interest to the individual, what types of products or services the individual purchases, etc. Market research data 703 may include data collected about individual consumers”); implementing administrator controls for group access by deploying pixel tracking across group member interactions(Montgomery, para 0030 discloses implementing pixel tracking (can be used for social media) and conversion “When a viewer (e.g., via a viewer device 110) downloads content including the tracking pixel, the tracking pixel causes the viewer device 110 to direct a cookie request to the targeting server 102 or another suitable server”; para 0033 further disclose conversion data “ Event data may describe things that the viewer does after placement of the cookie including, for example, being served an advertisement or other content, clicking through the advertisement or other content, converting the advertisement or other content, etc. Time of day and day of week data may indicate the times and dates at which the viewer performs the various activities described herein”);
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of tracking and extracting user activities data of Montgomery into feature of extracting user characteristics from image of Bleicher to produce an expected result of profiling users based on user characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to track user activities and provide contents accordingly (Montgomery, para 0030).
But Bleicher and Montgomery don’t explicitly teach aggregating multiple profiles into a group profile by processing comprehensive body scans using machine learning algorithms to extract physical measurements; integrating social media data through secure API connections with OAuth protocols;
analyzing collective data for shared preferences through: implementing TensorFlow models to analyze group social media activity; calculating attribute weights using vector similarity algorithms wherein the similarity scores incorporate both physical congruence and shared user-declared preferences; comparing group and influencer profiles using Sentence-BERT embeddings stored in Milvus vector databases; applying a hybrid recommendation engine comprising collaborative filtering (SVD matrix factorization) and content-based filtering to optimize influencer alignment;
maintaining secure API connections for external platform integration; providing privacy protections for group members through: processing recommendation insights locally on user devices; implementing data access controls with OAuth delegation protocols.
However, in the same field of endeavor of user profile generation Nadig teaches aggregating multiple profiles into a group profile by processing comprehensive body scans using machine learning algorithms to extract physical measurements (Nadig, col 11:36-40 discloses a group profile associated with multiple individual user profile “A group profile may be specific to a group of users. That is, a group profile may be associated with two or more individual user profiles”; where Bleicher in para 0087 discloses generation of user profile using physical measurement data and additional data such as user’s history and preference data “At step 720 the processed user scanned data may be parameterized, for example, to extract the precise face length, width, height, proportions, nose width and volume….together with the user history data from 705 and user preferences data at 715, to generate a user's glasses shopping profile based on the user's face profile and other physical properties, as well as user behavior and user preference data”); analyzing collective data for shared preferences(Nadig, col 11:41-43 discloses a shared preference for group profile “A group profile may include preferences shared by all the user profiles associated therewith”); maintaining secure API connections for external platform integration; providing privacy protections for group members through: processing recommendation insights locally on user devices(Nadig, col 18:47-51 discloses privacy/access control for group and users “to maintain user privacy, the first content publisher 135a may not have access to a user identifier and/or group identifier stored in the profile storage 270. In these embodiments, the metadata may include an identifier that uniquely corresponds to a particular user identifier and/or group identifier stored in the profile storage 270”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of generating marketing persona of Nadig into feature of extracting user characteristics from image of Bleicher and Montgomery to produce an expected result of generating avatar/persona using user characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to improve the user experiences by controlling receiving of same contents from different sources (Nadig, col 2: 61-65).
But Bleicher, Montgomery and Nadig don’t explicitly teach integrating social media data through secure API connections with OAuth protocols; analyzing collective data for shared preferences through: implementing TensorFlow models to analyze group social media activity; calculating attribute weights using vector similarity algorithms wherein the similarity scores incorporate both physical congruence and shared user-declared preferences; comparing group and influencer profiles using Sentence-BERT embeddings stored in Milvus vector databases;
applying a hybrid recommendation engine comprising collaborative filtering (SVD matrix factorization) and content-based filtering to optimize influencer alignment; implementing data access controls with OAuth delegation protocols.
However, in the same field of endeavor of securing access Andriukhin teaches integrating social media data through secure API connections with OAuth protocols; implementing data access controls with OAuth delegation protocols (Andriukhin, para 0045 discloses integrating or interfacing via secure APIs employing OAuth protocol “integration Layer 291 allows system 200 to interface with external systems via secure APIs, employing protocols like OAuth 2.0 for secure access”);
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of securing access of Andriukhin into feature of extracting user characteristics from image of Bleicher, Montgomery and Nadig to produce an expected result of securing access. The modification would be obvious because one of ordinary skill in the art would be motivated to access data from third-party applications (social media) securely using OAuth protocol(Andriukhin, para 0045).
But Bleicher, Montgomery, Nadig and Andriukhin don’t explicitly teach analyzing collective data for shared preferences through: implementing TensorFlow models to analyze group social media activity; calculating attribute weights using vector similarity algorithms wherein the similarity scores incorporate both physical congruence and shared user-declared preferences; comparing group and influencer profiles using Sentence-BERT embeddings stored in Milvus vector databases;
applying a hybrid recommendation engine comprising collaborative filtering (SVD matrix factorization) and content-based filtering to optimize influencer alignment;
However, in the same field of endeavor of user activity analysis Rajaram teaches analyzing collective data for shared preferences through: implementing TensorFlow models to analyze group social media activity(Rajaram, para 0093 discloses using TensorFlow platform to analyze and identify and analyze user data (which can similarly be applied for social media data to identify certain type of data) “a context matching algorithm widget may use LLMs, Machine Learning libraries (e.g., scikit-learn, TensorFlow), and Vector databases (e.g., Faiss, Annoy, Pinecone) to identify information relevant to identify the document processed by an AI assistant 137……”); calculating attribute weights using vector similarity algorithms wherein the similarity scores incorporate both physical congruence and shared user-declared preferences; comparing group and influencer profiles using Sentence-BERT embeddings stored in Milvus vector databases(Rajaram, para 0040 discloses similarity is being calculated for BERT embeddings/vector for any data (profiles) “Behavior Instruction is implemented by converting each defined situation in the behavior instructions into a vector embedding using a pre-trained language model (e.g., BERT, a Robustly Optimized BERT Pretraining Approach (ROBERTa), or a fine-tuned domain-specific model). These situation embeddings are stored along with their corresponding desired responses in the AssistantConfig. Then, when processing a user query, the user query is converted into a vector embedding using the same embedding model. The cosine similarity between the query embedding and all stored situation embeddings is computed. Then, if the highest similarity score exceeds a predetermined threshold (e.g., 0.85), the corresponding desired response is retrieved”; where para 0093 discloses a vector database for similarity calculation “Vector databases (e.g., Faiss, Annoy, Pinecone) to identify information relevant to identify the document processed by an AI assistant 137”),
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of using vector database of Rajaram into feature of extracting user characteristics from image of Bleicher, Montgomery, Nadig and Andriukhin to produce an expected result of generating avatar/persona using user characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to make the similarity searches faster by storing vectors in vector database (Rajaram, para 0093).
But Bleicher, Montgomery, Nadig, Andriukhin and Rajaram don’t explicitly teach applying a hybrid recommendation engine comprising collaborative filtering (SVD matrix factorization) and content-based filtering to optimize influencer alignment;
However, in the same field of endeavor of user profile creation Montgomery teaches applying a hybrid recommendation engine comprising collaborative filtering (SVD matrix factorization) and content-based filtering to optimize influencer alignment(Seyeditabari, col 12:8-16 discloses using BERT for contextual/semantic similarity measurement along with SVD (Singular Value Decomposition) principal “BERT having any number of layers, each of which may be configured to learn different contextual information from the respective representations. Alternatively, the model 485 may be any principal component analysis; singular value decomposition; deep learning system; nearest neighbor method or analysis; factorization method or analysis; generative model; gradient boosted decision tree; support vector machine; similarity measure”)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of measuring similarity of Seyeditabari into feature of extracting user characteristics from image of Bleicher, Montgomery, Nadig, Andriukhin and Rajaram to produce an expected result of matching with related contents. The modification would be obvious because one of ordinary skill in the art would be motivated to accurately predict the match accurately using trained machine learning model(Seyeditabari, abstract).
Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in further view of Nadig, Vinaya et al (PGPUB Document No. 12211497), hereafter, referred to as “Nadig”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Rajaram; Santoash et al (PGPUB Document No. 20250124044), hereafter, referred to as “Rajaram”, in further view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Carmel, David et al (PGPUB Document No. 20140337129), hereafter, referred to as “Carmel”.
Regarding Claim 9 (Currently Amended), Bleicher, Montgomery, Nadig, Andriukhin, Rajaram and Seyeditabari teach all the limitations of claim 8, and Nadig further teaches tracking group-wide engagement metrics to refine future recommendations through a feedback optimization loop(Nadig, col 39:12-15 discloses tracking or keeping history of purchase activities “The other data 815 may also include data representing previously provided user feedback indicating the appropriateness of previously output inferred content. The other data 815 may also include social media data associated with the user, system usage history associated with the user, a history of music listened to by the user, a history of books purchased by the user, a general purchasing history of the user”; Nadig, col 38:51-57 discloses recommending/outputting based on interaction data “as users interacting with devices 110 located in different geographic locations may be susceptible to receiving different amounts and/or kinds of inferred content. If the supplemental content system 130 determines the geographic location of the device 110 indicates inferred content should be output, the supplemental content system 130 may generate request data 845. If the supplemental content system 130 determines the geographic location of the device 110 indicates inferred content should not be output”);
But Bleicher, Montgomery, Nadig, Andriukhin, Rajaram and Seyeditabari don’t explicitly teach calculating preference weights between individual and group attributes, wherein the weights incorporate factors including physical congruence, shared self-reported preferences, and historical interaction patterns, and are dynamically updated using real-time behavioral data; delivering personalized content based on the dynamically weighted preferences;
However, in the same field of endeavor of weighing individual and group preferences Carmel teaches calculating preference weights between individual and group attributes, wherein the weights incorporate factors including physical congruence, shared self-reported preferences, and historical interaction patterns, and are dynamically updated using real-time behavioral data (Carmel, para 0012 discloses designating/identifying primary purchaser in a group “Once an individual's preferences (e.g., calculated as weight) for one or more features for the collection of items are determined, then the determined preferences are compared to the preferences of a group of other individuals……..”); personalized content based on the dynamically weighted preferences (Carmel, para 0033 discloses recommending delivery based on preference “items that satisfy the user unique preferences are recommended for delivery…..”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of measuring individual’s preference compared to a group of Carmel into feature of extracting user characteristics from image of Bleicher, Montgomery, Nadig, Andriukhin, Rajaram and Seyeditabari to produce an expected result of delivering content based on preferences. The modification would be obvious because one of ordinary skill in the art would be motivated to recommend items which are preferred to user’s(Carmel, para 0033).
Claim 10 is rejected under 35 U.S.C. 103 as being unpatentable over Bleicher, David et al (PGPUB Document No. 20190188784), hereafter referred as to “Bleicher”, in view of Montgomery, Gerard et al (PGPUB Document No. 20130282817), hereafter, referred to as “Montgomery”, in further view of Nadig, Vinaya et al (PGPUB Document No. 12211497), hereafter, referred to as “Nadig”, in view of Andriukhin, Evgenni et al (PGPUB Document No. 20240154993), hereafter, referred to as “Andriukhin”, in view of Rajaram; Santoash et al (PGPUB Document No. 20250124044), hereafter, referred to as “Rajaram”, in further view of Seyeditabari, Narjessadat et al (US Patent No. 11863643 ), hereafter, referred to as “Seyeditabari”, in further view of Tepmongkol, Warangkana et al (PGPUB Document No. 20120110011), hereafter, referred to as “Tepmongkol”.
Regarding Claim 10 (Currently Amended), Bleicher, Montgomery, Nadig, Andriukhin, Rajaram and Seyeditabari teach all the limitations of claim 8, and Nadig further teaches controlling access to group-wide content based on role-based permissions; managing connection requests through administrator approval workflows (Nadig, col 18:47-51 discloses privacy/access control for group and users “to maintain user privacy, the first content publisher 135a may not have access to a user identifier and/or group identifier stored in the profile storage 270. In these embodiments, the metadata may include an identifier that uniquely corresponds to a particular user identifier and/or group identifier stored in the profile storage 270”);
However, in the same field of endeavor of avatar generation Tepmongkol teaches replacing actual physical characteristics with fictional avatars for selected group members to protect privacy or enhance self-expression (Tepmongkol, abstract discloses avatar to represent a group of users “create a social network to share the appearances of an exclusive social group as represented by avatars dressed in virtual clothing and exchanging social commentary on same. Avatars, virtual clothing, and social commentary, …..and other such information related to personal looks can be exchanged in real time or by messages and data files to be read at a later time. Also included are means to ensure the privacy of an avatar. Also included are a means to estimate physical characteristics of clothing and from a virtual clothing representation”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the feature of representing avatar for a group of Tepmongkol into feature of extracting user characteristics from image of Bleicher, Montgomery, Nadig, Andriukhin, Rajaram and Seyeditabari to produce an expected result of generating avatar/persona using group characteristic data. The modification would be obvious because one of ordinary skill in the art would be motivated to improve the product purchasing experience of users by allowing group user feedback on virtual clothing representation on avatar (Tepmongkol, abstract).
Response to Arguments
I. 35 U.S.C §101
35 U.S.C §101 abstract idea rejection claim 1-17 has been withdrawn in response to claim amendments and argument consideration.
II. 35 U.S.C §103
Applicant’s arguments filed on 9/9/2025 have been fully considered but are
moot because the independent claim 1, 4, 5 and 8 have been amended with newly added features which applicant’s arguments are directed towards. Since claims have been amended with new features, a new ground of rejection is presented.
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 ABDULLAH A DAUD whose telephone number is (469)295-9283. The examiner can normally be reached M~F: 9:30 am~6:30 pm.
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 at 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 published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/ABDULLAH A DAUD/Examiner, Art Unit 2164 /AMY NG/Supervisory Patent Examiner, Art Unit 2164
Prosecution Insights