Patent Application 18195764 - DATA PROCESSING METHOD AND RELATED APPARATUS

Title: DATA PROCESSING METHOD AND RELATED APPARATUS
Application Information

Invention Title: DATA PROCESSING METHOD AND RELATED APPARATUS
Application Number: 18195764
Submission Date: 2025-05-13T00:00:00.000Z
Effective Filing Date: 2023-05-10T00:00:00.000Z
Filing Date: 2023-05-10T00:00:00.000Z
National Class: 705
National Sub-Class: 014420
Examiner Employee Number: 90447
Art Unit: 3621
Tech Center: 3600
Rejection Summary

102 Rejections: 0
103 Rejections: 2
Cited Patents

The following patents were cited in the rejection:
Office Action Text


    DETAILED ACTION
Continued Examination under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection.  Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.  Applicant's submission filed on 04/30/2025 has been entered.

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 .   

Examiner’s Comment
This Action is in response to the Request for Continued Examination filed on 04/30/2025 with Amended Claims and Applicant's Remarks filed on 04/30/2025.
Applicant has amended claims 1, 6, 12, 17, and 20 according to Amendments filed on 04/30/2025. Claims 1-20 are pending and currently under consideration for patentability.

Claim Rejections - 35 USC § 112(b)
The following is a quotation of 35 U.S.C. 112(b):
(b)  CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.


The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.


Claims 1-20 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA  35 U.S.C. 112, the applicant), regards as the invention. Independent Claims 1, 12, and 20 recites the limitation “the bid contract is associated with an advertisement effect and a pre-offered bid”.  There is insufficient antecedent basis for this limitation in the claim. The Examiner interprets “bid contract” to be “bid advertisement”.

Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.

Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter. The claims are directed to a judicial exception (i.e., a law of nature, natural phenomenon, or abstract idea) without significantly more.
Step 1:	In a test for patent subject matter eligibility, claims 1-20 are found to be in accordance with Step 1 (see 2019 Revised Patent Subject Matter Eligibility), as they are related to a process, machine, manufacture, or composition of matter. Claims 1-11 recite a method, claims 12-19 recite a system, and claim 20 recites a computer-readable storage medium. When assessed under Step 2A, Prong I, they are found to be directed towards an abstract idea. The rationale for this finding is explained below: 
Step 2A, Prong I: Under Step 2A, Prong I, claims 1, 12, and 20 are directed to an abstract idea without significantly more, as they all recite a judicial exception. Claims 1, 12, and 20 recite limitations directed to the abstract idea including obtaining an advertisement state of each candidate advertisement of a plurality of candidate advertisements corresponding to a current exposure request, the advertisement state representing a competition condition in response to that the candidate advertisement competes for the current exposure request, and obtaining an overall state of an advertising platform in response to the current exposure request, the overall state representing a current exposure task performance situation of the advertising platform; determining probabilities of each candidate advertisement belonging to different reference advertisement types; determining a competition score of each candidate advertisement for the current exposure request according to the advertisement state corresponding to the candidate advertisement and the overall state based on the probabilities of the candidate advertisement belonging to different reference advertisement types, wherein the plurality of candidate advertisements comprises a contract advertisement and a bid advertisement, the contract advertisement is associated with a contract specifying at least a predetermined playing amount, a selling price, and a targeting condition, the bid contract is associated with an advertisement effect and a pre-offered bid, the contract advertisement and the bid advertisement are mixed in the plurality of candidate advertisements for the determining of the competition score; determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request; and transmitting the target advertisement […] for playback. These further limitations are not seen as any more than the judicial exception: The limitations reciting “obtaining an advertisement state of each candidate advertisement corresponding to a current exposure request, the advertisement state representing a competition condition in response to that the candidate advertisement competes for the current exposure request, and obtaining an overall state of an advertising platform in response to the current exposure request, the overall state representing a current exposure task performance situation of the advertising platform; determining probabilities of each candidate advertisement belonging to different reference advertisement types; wherein the plurality of candidate advertisements comprises a contract advertisement and a bid advertisement, the contract advertisement is associated with a contract specifying at least a predetermined playing amount, a selling price, and a targeting condition, the bid contract is associated with an advertisement effect and a pre-offered bid, the contract advertisement and the bid advertisement are mixed in the plurality of candidate advertisements for the determining of the competition score; and determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request; and transmitting the target advertisement […] for playback” is considered to be an abstract idea, specifically, certain methods of organizing human activity; such as commercial interactions, advertising, marketing, and sales because the claims are directed to advertising interactions between different parties. The limitations reciting “obtaining an advertisement state of each candidate advertisement corresponding to a current exposure request, the advertisement state representing a competition condition in response to that the candidate advertisement competes for the current exposure request, and obtaining an overall state of an advertising platform in response to the current exposure request, the overall state representing a current exposure task performance situation of the advertising platform; determining probabilities of each candidate advertisement belonging to different reference advertisement types; and determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request; and determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request;” is directed to another abstract idea, specifically, mental processes such as concepts performed in the human mind (including an observation, evaluation, judgment, opinion) because the steps in the claims are directed to receiving/obtaining data (i.e. advertisement state and overall state) and determining data (i.e. probabilities and target advertisement). The limitations reciting “determining probabilities of each candidate advertisement belonging to different reference advertisement types; and determining a competition score of each candidate advertisement for the current exposure request according to the advertisement state corresponding to the candidate advertisement and the overall state based on the probabilities of the candidate advertisement belonging to different reference advertisement types;” is directed to another abstract idea, specifically, mathematical concepts such as mathematical relationships / formulas or equations / calculations because the claims are directed to determining probabilities and a score (i.e. competition score). Claims 1, 12, and 20 recite additional limitations including by a classification/scoring network in a scoring model; and “the scoring model comprising multiple scoring networks corresponding to different reference advertisement types, each of the multiple scoring networks being configured to score each candidate advertisement based on a corresponding reference advertisement type of the different reference advertisement types”; and terminal device. Therefore, under Step 2A, Prong I, claims 1, 12, and 20 are directed towards an abstract idea. 
Step 2A, Prong II: Step 2A, Prong II is to determine whether any claim recites any additional element that integrate the judicial exception (abstract idea) into a practical application. Claims 1, 12, and 20 have recited the following additional elements: by a classification/scoring network in a scoring model; and “the scoring model comprising multiple scoring networks corresponding to different reference advertisement types, each of the multiple scoring networks being configured to score each candidate advertisement based on a corresponding reference advertisement type of the different reference advertisement types”; and terminal device. The additional elements reciting – “by a classification/scoring network in a scoring model” and “terminal device” in claims 1, 12, and 20 are not found to integrate the judicial exception into a practical application. Merely adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea and Generally linking the use of the judicial exception to a particular technological environment or field of use – see MPEP 2106.05(f) is not indicative of integration into a practical application. Accordingly, alone, and in combination, these additional elements are seen as using a computer or tool to perform an abstract idea, adding insignificant-extra-solution activity to the judicial exception. They do no more than link the judicial exception to a particular technological environment or field of use, i.e. networks in a model and device, and therefore do not integrate the abstract idea into a practical application. The courts decided that although the additional elements did limit the use of the abstract idea, the court explained that this type of limitation merely confines the use of the abstract idea to a particular technological environment and this fails to add an inventive concept to the claims (See Affinity Labs of Texas v. DirecTV, LLC,). Under Step 2A, Prong II, these claims remain directed towards an abstract idea. 
Step 2B: Claims 1, 12, and 20 recite the following additional limitations including by a classification/scoring network in a scoring model; and “the scoring model comprising multiple scoring networks corresponding to different reference advertisement types, each of the multiple scoring networks being configured to score each candidate advertisement based on a corresponding reference advertisement type of the different reference advertisement types”; and terminal device. The additional elements reciting – “by a classification/scoring network in a scoring model” and “terminal device” do not integrate the judicial exception (abstract idea) into a practical application because of the analysis provided in Step 2A, Prong II. Claims 1, 12, and 20 also recite additional elements – “the scoring model comprising multiple scoring networks corresponding to different reference advertisement types, each of the multiple scoring networks being configured to score each candidate advertisement based on a corresponding reference advertisement type of the different reference advertisement types”. Merely describing a model that has different scoring networks for different advertisement types are not indicative of integration into a practical application because the limitation is adding insignificant extra-solution activity to the judicial exception - see MPEP 2106.05(g). The courts have noted that “performing repetitive calculations” is seen as a well-understood, routine, and conventional computer function (See: Flook, 437 U.S. at 594, 198 USPQ2d at 199 (recomputing or readjusting alarm limit values); Bancorp Services v. Sun Life, 687 F.3d 1266, 1278, 103 USPQ2d 1425, 1433 (Fed. Cir. 2012) ("The computer required by some of Bancorp’s claims is employed only for its most basic function, the performance of repetitive calculations, and as such does not impose meaningful limits on the scope of those claims.")). Claims 1, 12, and 20 do not include additional elements or a combination of elements that result in the claims amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements listed amount to no more than mere instructions to apply an exception using a generic computer component. In addition, the applicant’s specifications describe generic computer-based elements, ¶¶ [0011] [0025], for implementing the “apparatus” or  “processor”, which do not amount to significantly more than the abstract idea of itself, which is not enough to transform an abstract idea into eligible subject matter. Furthermore, there is no improvement in the functioning of the computer or technological field, and there is no transformation of subject matter into a different state. Under Step 2B in a test for patent subject matter eligibility, these claims are not patent eligible. 
Dependent claims 2-11 and 13-19 further recite the method and system of claims 1 and 12, respectively. Dependent claims 2-11 and 13-19 when analyzed as a whole are held to be patent ineligible under 35 U.S.C. 101 because the additional recited limitation fail to establish that the claims are not directed to an abstract idea:
Under Step 2A, Prong I, these additional claims only further narrow the abstract idea set forth in claims 1, 12, and 20. For example, dependent claims 2-11 and 13-20 further describe the limitations for determining a competition score for a candidate advertisement in order to determine an advertisement state – which is only further narrowing the scope of the abstract idea recited in the independent claims.  
Under Step 2A, Prong II, for dependent claims 2-11 and 13-20, there are no additional elements introduced that integrate the claims into a practical application. For example, dependent claims 2-11 and 13-20 recite further describe the environment in which the abstract idea takes place such as an advertising platform with scoring models and scoring networks. Thus, they do not present integration into a practical application, or amount to significantly more. 
Under Step 2B, the dependent claims do not include any additional elements that are sufficient to amount to significantly more than the judicial exception. Additionally, there is no improvement in the functioning of the computer or technological field, and there is no transformation of subject matter into a different state. As discussed above with respect to integration of the abstract idea into a practical application, the additional claims do not provide any additional elements that would amount to significantly more than the judicial exception. Under Step 2B, these claims are not patent eligible.


Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.

Claim(s) 1-9 and 11-20 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Publication 2016/0275554 to Yan in view of U.S. Publication 2023/0004833 to Sahasi.

Claims 1, 12, and 20 are method, system, and computer-readable medium claims, respectively, with substantially indistinguishable features between each group.  For purposes of compact prosecution, the Office has grouped the common method, system and non-transitory computer readable storage medium claims in applying applicable prior art.

With respect to Claim 1:
Yan teaches:
A data processing method, performed by a computing device, the method comprising: obtaining an advertisement state of each candidate advertisement of a plurality of candidate advertisements corresponding to a current exposure request, the advertisement state representing a competition condition in response to that the candidate advertisement competes for the current exposure request, and obtaining an overall state of an advertising platform in response to the current exposure request, the overall state representing a current exposure task performance situation of the advertising platform (i.e. obtaining targeting criteria for candidate advertisements which reads on advertisement state representing a competition condition and obtaining an expected value corresponding to bid value of ad request which reads on overall state representing a current exposure task performance) (Yan: ¶¶ [0033] [0034] “In one embodiment, the targeting criteria may specify actions or types of connections between a user and another user or object of the online system 140. Targeting criteria may also specify interactions between a user and objects performed external to the online system 140, such as on a third party system 130…As another example, targeting criteria identifies users having a connection to another user or object or having a particular type of connection to another user or object. The correlation module 235 determines a correlation between content provided by third party systems 130 accessed by a viewing user of the online system 140 and content provided by third party systems 130 accessed by additional users of the online system 140. For example, the correlation module 235 determines a correlation between web pages provided by one or more third party systems 130 accessed by the viewing user and web pages provided by one or more third party systems 130 accessed by additional users of the online system 140 who interacted with an advertisement. The correlation may be based on any suitable metric for measuring a similarity between content provided by one or more third party systems 130 accessed by the viewing user and content provided by one or more third party systems 130 accessed by additional users. For example, a correlation between web pages from one or more third party systems 130 accessed by a viewing user and web pages from one or more third party systems 130 accessed by additional users who accessed an advertisement is a cosine similarity between web pages from one or more third party systems 130 accessed by the viewing user and web pages from one or more third party systems 130 accessed by the additional users who accessed the advertisement.” Furthermore, as cited in ¶ [0031] “The bid amount is associated with an advertisement by an advertiser and is used to determine an expected value, such as monetary compensation, provided by an advertiser to the online system 140 if advertisement content in the ad request is presented to a user, if the advertisement content in the ad request receives a user interaction when presented, or if any suitable condition is satisfied when advertisement content in the ad request is presented to a user. For example, the bid amount specifies a monetary amount that the online system 140 receives from the advertiser if advertisement content in an ad request is displayed. In some embodiments, the expected value to the online system of presenting the advertisement content may be determined by multiplying the bid amount by a probabilities of the advertisement content being accessed by a user.”); 
determining, by a scoring network in the scoring model, a competition score of each candidate advertisement for the current exposure request according to the advertisement state corresponding to the candidate advertisement and the overall state based on the probabilities of the candidate advertisement belonging to different reference advertisement types (i.e. determining affinity scores for each candidate advertisement/content based on probabilities or cosine similarity of candidate advertisement being similar/different to different advertisements) (Yan: ¶ [0037] “Based at least in part on the correlation between content provided by one or more third party systems 130 accessed by the viewing user and content provided by one or more third party systems 130 accessed by one or more additional users who interacted with an advertisement, the correlation module 235 determines a score for the advertisement. For example, the greater the cosine similarity between content provided by one or more third party systems 13 0 accessed by the viewing user and content provided by one or more third party systems 130 accessed by one or more additional users who interacted with an advertisement, the higher the score computed for the advertisement. In some embodiments, the score for an advertisement may also be determined based content provided by one or more third party systems 130 with which the viewing user accessed and interactions with the advertisement by additional users who also accessed the content provided by the one or more third party systems 130… For example, a score associated with an advertisement is proportional to a percentage of online system users presented with the advertisement who performed a specific type of interaction with the advertisement (e.g., accessed the advertisement, made a purchase of a product identified by the advertisement within a threshold time of interacting with the advertisement), so the greater the percentage of users performing the specific type of interaction, the greater the score for the advertisement. As an additional example, the score for an advertisement is proportional to an amount of revenue generated from purchases made by users in association with accessing the advertisement.”),
wherein the plurality of candidate advertisements comprises a contract advertisement and a bid advertisement, the contract advertisement is associated with a contract specifying at least a predetermined playing amount, a selling price, and a targeting condition, the bid contract is associated with an advertisement effect and a pre-offered bid, the contract advertisement and the bid advertisement are mixed in the plurality of candidate advertisements for the determining of the competition score (i.e. candidate advertisements include ad requests with bid amounts or contract ads and ad requests associated with content stories or bid advertisements, wherein ad requests associated with bid requests include expected value, targeting criteria and bid amount, the ad requests associated with content stories include an advertisement effect or action and associated bid amount, wherein both ad requests associated with bid amounts and ad requests associated with content stories are included in a unified ranking to select candidate advertisements) (Yan: ¶¶ [0039] [0040] “An expected value associated with an ad request or with a content item represents an expected amount of compensation to the online system 140 for presenting an ad request or a content item. For example, the expected value associated with an ad request is a product of the ad request's bid amount and a likelihood of the user interacting with the ad content from the ad request. The content selection module 240 may rank ad requests based on their associated bid amounts and select ad requests having at least a threshold position in the ranking for presentation to the user. In some embodiments, the content selection module 240 ranks both content items not associated with bid amounts and ad requests in a unified ranking based on bid amounts associated with ad requests and measures of relevance associated with content items and ad requests. Based on the unified ranking, the content selection module 240 selects content for presentation to the user. Selecting ad requests and other content items through a unified ranking is further described in U.S. patent application Ser. No. 13/545, 266, filed on Jul. 10, 2012, which is hereby incorporated by reference in its entirety…For example, the content selection module 240 receives a request to present a feed of content to a user of the online system 140. The feed may include one or more advertisements from ad request as well as content items, such as stories describing actions associated with other online system users connected to the user. The content selection module 240 accesses one or more of the user profile store 205, the content store 210, the action log 220, and the edge store 225 to retrieve information about the user. For example, stories or other data associated with users connected to the identified user are retrieved. Additionally, one or more advertisement requests ("ad requests") may be retrieved from the ad request store 230 The retrieved stories, ad requests, or other content items, are analyzed by the content selection module 240 to identify candidate content that is likely to be relevant to the identified user. For example, stories associated with users not connected to the identified user or stories associated with users for which the identified user has less than a threshold affinity are discarded as candidate content. Based on various criteria, the content selection module 240 selects one or more of the content items or ad requests identified as candidate content for presentation to the identified user. The selected content items or ad requests are included in a feed of content that is presented to the user. For example, the feed of content includes at least a threshold number of content items describing actions associated with users connected to the user via the online system 140.” Furthermore, as cited in ¶¶ [0031] [0032] “The bid amount is associated with an advertisement by an advertiser and is used to determine an expected value, such as monetary compensation, provided by an advertiser to the online system 140 if advertisement content in the ad request is presented to a user, if the advertisement content in the ad request receives a user interaction when presented, or if any suitable condition is satisfied when advertisement content in the ad request is presented to a user. For example, the bid amount specifies a monetary amount that the online system 140 receives from the advertiser if advertisement content in an ad request is displayed…Targeting criteria included in an advertisement request specify one or more characteristics of users eligible to be presented with advertisement content in the advertisement request. For example, targeting criteria are used to identify users having user profile information, edges, or actions satisfying at least one of the targeting criteria. Hence, targeting criteria allow an advertiser to identify users having specific characteristics, simplifying subsequent distribution of content to different users.”); and 
determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request (i.e. determining/selecting content for current advertisement request based on a rank/score) (Yan: ¶¶ [0038] [0039] “The content selection module 240 selects one or more content items for communication to a client device 110 to be presented to a user. Content items eligible for presentation to the user are retrieved from the content store 210, from the ad request store 230, or from another source by the content selection module 235, which selects one or more of the content items for presentation to the viewing user. A content item eligible for presentation to the user is a content item associated with at least a threshold number of targeting criteria satisfied by characteristics of the user or is a content item that is not associated with targeting criteria… The content selection module 240 may rank ad requests based on their associated bid amounts and select ad requests having at least a threshold position in the ranking for presentation to the user. In some embodiments, the content selection module 240 ranks both content items not associated with bid amounts and ad requests in a unified ranking based on bid amounts associated with ad requests and measures of relevance associated with content items and ad requests. Based on the unified ranking, the content selection module 240 selects content for presentation to the user.”); and
transmitting the target advertisement to a terminal device for playback (i.e. target ad is transmitted to client device for presentation) (Yan: ¶ [0059] “Based on the one or more selection processes, the online system 140 selects one or more ad requests for presentation to the viewing user and communicates advertisements from the selected one or more ad requests to the client device 110 for presentation to the viewing user.”).
Yan does not explicitly disclose determining, by a classification network in a scoring model, probabilities of each candidate advertisement belonging to different reference advertisement types; and the scoring model comprising multiple scoring networks corresponding to different reference advertisement types, each of the multiple scoring networks being configured to score each candidate advertisement based on a corresponding reference advertisement type of the different reference advertisement types.
However, Sahasi further discloses:
determining, by a classification network in a scoring model, probabilities of each candidate advertisement belonging to different reference advertisement types (i.e. determining/predicting labels according to features or different reference advertisement types, wherein the predicted labels reads on probabilities of candidate ad belonging to different reference types because the media asset is predicted to correspond to labeled ad feature) (Sahasi: ¶¶ [0151] [0152] “described herein may be referred to as "at least one machine learning model 1630" or simply the "machine learning model 1630", as shown in FIG. 16 The at least one machine learning model 1630 may be trained by a system 1600 shown in FIG. 16. The system 1600 may be configured to use machine learning techniques to train, based on an analysis of one or more training datasets 1610A-1610B by a training module 1620, the at least one machine learning model 1630. The at least one machine learning model 1630, once trained, may be configured to determine a prediction that a media asset is of interest to a particular user or not of interest to the particular user. A dataset indicative of a plurality of media assets and a labeled ( e.g., predetermined/known) prediction indicating whether the corresponding media assets are of interest to a particular user or not may be used by the training module 1620 to train the at least one machine learning model 1630. Each of the plurality of media assets in the dataset may be associated with a plurality of features that are present within each corresponding media asset. The plurality of features and the labeled predictions may be used to train the at least one machine learning model 1630…The training dataset 1610A may comprise a first portion of the plurality of media assets in the dataset. Each media asset in the first portion may have a labeled (e.g., predetermined) prediction and one or more labeled features. The training dataset 1610B may comprise a second portion of the plurality of media assets in the dataset. Each media asset in the second portion may have a labeled (e.g., predetermined) prediction and one or more labeled features. The plurality of media assets may be randomly assigned to the training dataset 1610A, the training dataset 1610B, and/or to a testing dataset.”); 
the scoring model comprising multiple scoring networks corresponding to different reference advertisement types (i.e. different models correspond to different ad feature sets, wherein the feature sets reads on reference advertisement types) (Sahasi: ¶¶ [0154] [0155] “The training module 1620 may extract a feature set from the training dataset 1610A and/or the training dataset 1610B in a variety of ways. For example, the training module 1620 may extract a feature set from the training dataset 1610A and/or the training dataset 1610B using a classification model (e.g., a machine learning model). The training module 1620 may perform feature extraction multiple times, each time using a different feature-extraction technique. In one example, the feature sets generated using the different techniques may each be used to generate different machine learning models 1640. For example, the feature set with the highest quality features (e.g., most indicative of interest or not of interest to a particular user(s)) may be selected for use in training… The training dataset 1610A and/or the training dataset 1610B may be analyzed to determine any dependencies, associations, and/or correlations between features and the labeled predictions in the training dataset 1610A and/or the training dataset 1610B. The identified correlations may have the form of a list of features that are associated with different labeled predictions (e.g., of interest to a particular user vs. not of interest to a particular user). The term "feature", as used herein, may refer to any characteristic of an item of data that may be used to determine whether the item of data falls within one or more specific categories or within a range. By way of example, the features described herein may comprise one or more features present within each of the media assets that may be correlative (or not correlative as the case may be) with a particular media asset being of interest to a particular user or not.”),
each of the multiple scoring networks being configured to score each candidate advertisement based on a corresponding reference advertisement type of the different reference advertisement types (i.e. multiple different scoring models correspond to multiple different advertisement features sets) (Sahasi: ¶¶ [0160] [0161] “After the training module 1620 has generated a feature set(s), the training module 1620 may generate the one or more machine learning models 1640A-1640N based on the feature set(s). A machine learning model (e.g., any of the one or more machine learning models 1640A-1640N) may refer to a complex mathematical model for data classification that is generated using machine-learning techniques as described herein. In one example, a machine learning model may include a map of support vectors that represent boundary features. By way of example, boundary features may be selected from, and/or represent the highest-ranked features in, a feature set…The training module 1620 may use the feature sets extracted from the training dataset 161 0A and/ or the training dataset 1610B to build the one or more machine learning models 1640A-1640N for each classification category (e.g., "of interest to a particular user media asset" and "not of interest to the particular user media asset"). In some examples, the one or more machine learning models 1640A- 340N may be combined into a single machine learning model 1640 (e.g., an ensemble model). Similarly, the machine learning model 1630 may represent a single classifier containing a single or a plurality of machine learning models 1640 and/or multiple classifiers containing a single or a plurality of machine learning models 1640 (e.g., an ensemble classifier).”).
Therefore, it would have been obvious to one of ordinary skill in the art, at the time the invention was made, to add Sahasi’s determining, by a classification network in a scoring model, probabilities of each candidate advertisement belonging to different reference advertisement types; and the scoring model comprising multiple scoring networks corresponding to different reference advertisement types, each of the multiple scoring networks being configured to score each candidate advertisement based on a corresponding reference advertisement type of the different reference advertisement types to Yan’s determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request.  One of ordinary skill in the art would have been motivated to do so in order to allow “for improved model selection and content recommendations” (Sahasi: ¶ [0002]).
With respect to Claims 12 and 20:
All limitations as recited have been analyzed and rejected to claim 1. Claim 12 recites “A data processing apparatus, deployed on a computing device, the apparatus comprising: a memory storing computer program instructions; and a processor coupled to the memory and configured to execute the computer program instructions and perform:” (Yan: ¶¶ [0062] [0063]) the steps of method claim 1. Claim 20 recites “A non-transitory computer-readable storage medium storing computer program instructions executable by at least one processor to perform:” (Yan: ¶¶ [0062] [0063]) the steps of method claim 1. Claims 12 and 20 do not teach or define any new limitations beyond claim 1. Therefore they are rejected under the same rationale.

With respect to Claim 2:
Yan teaches:
The method according to claim 1, wherein determining the competition score of the candidate advertisement comprises: determining an input feature of the candidate advertisement according to the advertisement state corresponding to the candidate advertisement and the overall state (i.e. advertiser specifies/determines targeting criteria for advertisement slot) (Yan: ¶¶ [0032] [0033] “Additionally, an advertisement request may include one or more targeting criteria specified by the advertiser. Targeting criteria included in an advertisement request specify one or more characteristics of users eligible to be presented with advertisement content in the advertisement request. For example, targeting criteria are used to identify users having user profile information, edges, or actions satisfying at least one of the targeting criteria. Hence, targeting criteria allow an advertiser to identify users having specific characteristics, simplifying subsequent distribution of content to different users…In one embodiment, the targeting criteria may specify actions or types of connections between a user and another user or object of the online system 140. Targeting criteria may also specify interactions between a user and objects performed external to the online system 140, such as on a third party system 130.”); 
performing […] weighted processing on the input feature of the candidate advertisement to obtain an input feature of the candidate advertisement under each reference advertisement type (i.e. weighted processing is applied to the targeting criteria of the advertisements) (Yan: ¶ [0036] “Additionally, when determining correlation between content provided by one or more third party systems 130 accessed by the viewing user and content provided by one or more third party systems 130 accessed by additional users who interacted with an advertisement, the correlation module 235 accounts for a time period associated with the accesses to content provided by one or more third party systems 130 or associated with the interactions with the advertisement. For example, content provided by one or more third party systems 130 accessed by additional users within a threshold time interval of a current time have higher weights than content provided by one or more party systems 130 greater than the threshold time interval.” Furthermore, as cited in ¶ [0055] “The average set of content accessed by the additional users may be based at least in part on weights associated with content provided by one or more third party systems 130. In some embodiments, weights associated with content provided by one or more third party systems 130 are determined based at least in part on a number of times, or a frequency with which, additional users accessed different content provided by third party systems 130. For example, content provided by a third party system 130 that was accessed by additional users who interacted with more than a threshold amount of different content provided by one or more third party systems 130 has a higher weight than content provided by a third party system 130 that was accessed by additional users who interacted with less than the threshold amount of different content provided by the one or more third party systems 130.”); 
configuring, by each of the scoring networks in the scoring model, a competition score for the candidate advertisement according to the input feature of the candidate advertisement under the reference advertisement type corresponding to the scoring network (i.e. configuring the score for the advertisement according to features of the candidate advertisement and other different advertisement types corresponding to different scoring rules) (Yan: ¶ [0037] “Based at least in part on the correlation between content provided by one or more third party systems 130 accessed by the viewing user and content provided by one or more third party systems 130 accessed by one or more additional users who interacted with an advertisement, the correlation module 235 determines a score for the advertisement. For example, the greater the cosine similarity between content provided by one or more third party systems 13 0 accessed by the viewing user and content provided by one or more third party systems 130 accessed by one or more additional users who interacted with an advertisement, the higher the score computed for the advertisement. In some embodiments, the score for an advertisement may also be determined based content provided by one or more third party systems 130 with which the viewing user accessed and interactions with the advertisement by additional users who also accessed the content provided by the one or more third party systems 130. For example, advertisements with which at least a threshold number of additional users who accessed content provided by one or more third party systems 130 with which the viewing user also accessed at least a threshold number of times have higher scores than advertisements with which additional users who accessed content provided by one or more third party systems 130 with which the viewing user accessed less than the threshold number of times. In some embodiments, the correlation module 235 accounts for overall interaction with an advertisement by users of the online system 140 when computing a score for the advertisement. For example, a score associated with an advertisement is proportional to a percentage of online system users presented with the advertisement who performed a specific type of interaction with the advertisement ( e.g., accessed the advertisement, made a purchase of a product identified by the advertisement within a threshold time of interacting with the advertisement), so the greater the percentage of users performing the specific type of interaction, the greater the score for the advertisement. As an additional example, the score for an advertisement is proportional to an amount of revenue generated from purchases made by users in association with accessing the advertisement.”); and 
determining a competition score of the candidate advertisement for the current exposure request through the competition scores configured for the candidate advertisement by each of the scoring networks in the scoring model  (i.e. determining/selecting content for current advertisement request based on a rank/score) (Yan: ¶¶ [0038] [0039] “The content selection module 240 selects one or more content items for communication to a client device 110 to be presented to a user. Content items eligible for presentation to the user are retrieved from the content store 210, from the ad request store 230, or from another source by the content selection module 235, which selects one or more of the content items for presentation to the viewing user. A content item eligible for presentation to the user is a content item associated with at least a threshold number of targeting criteria satisfied by characteristics of the user or is a content item that is not associated with targeting criteria… The content selection module 240 may rank ad requests based on their associated bid amounts and select ad requests having at least a threshold position in the ranking for presentation to the user. In some embodiments, the content selection module 240 ranks both content items not associated with bid amounts and ad requests in a unified ranking based on bid amounts associated with ad requests and measures of relevance associated with content items and ad requests. Based on the unified ranking, the content selection module 240 selects content for presentation to the user.” Furthermore, as cited in ¶ [0057] “Additional factors may be used by the online system 140 when generating 355 the score for an advertisement. For example, content provided by third party systems 130 accessed by the viewing user is used when generating 355 the score for an advertisement. For example, if the viewing user accesses content from a product page of an online merchant more frequently than content from other third party systems 130, a score of an ad request including an advertisement with which additional users who accessed product page interacted is generated 355 in part based on a percentage of the additional users who accessed the product page who also interacted with the advertisement. As an additional example, if the viewing user accesses particular web page, scores for ad requests including advertisements with which at least a threshold number or percentage of additional users who accessed the web page interacted are generated 355 based at least in part on the number of times or frequency with which the viewing user accessed the particular web page. A score generated 355 for an ad request including an advertisement may additionally be based on an overall performance of the advertisement when presented to online system users as a whole. For example, a score generated 355 for an ad request is proportional to a number or a percentage of online system users presented with an advertisement in the ad request who performed one or more types of interactions with the advertisement, so ad requests including advertisements with which a larger number or percentage of online system users presented with the advertisements interacted have higher scores.”).
Yan does not explicitly disclose performing, based on the probabilities of the candidate advertisement belonging to different reference advertisement types, weighted processing on the input feature of the candidate advertisement to obtain an input feature of the candidate advertisement under each reference advertisement type.
However, Sahasi further discloses performing, based on the probabilities of the candidate advertisement belonging to different reference advertisement types, weighted processing on the input feature of the candidate advertisement to obtain an input feature of the candidate advertisement under each reference advertisement type (i.e. performing, based on predicted similar content features, weighted processing on advertisement features to determine recommended advertisement features) (Sahasi: ¶ [0059] “A feature vector may be associated with a particular user device(s). A feature vector may comprise a quantification of a level/amount of engagement with each media asset and/or a numerical weight associated with an engagement feature as described herein. The number and arrangement of items in such a feature vector may be the same as those of features vectors used during training of the scoring model 248. The scoring unit 230 can then apply the scoring model 248 to the feature vector to generate an interest attribute representing a level of interest on the media asset. The interest attribute can be a numerical value (e.g., an integer number) or textual label that indicates the level of interest ( e.g., "high", "moderate", and "low").” Furthermore, as cited in ¶ [0184] “At step 2020, the computing device may determine a feature vector associated with each user device of the plurality of user devices. For example, the computing device may determine the feature vector based on the plurality of activity data and the plurality of engagements. Each feature vector associated with each user device of the plurality of user devices may comprise at least one content feature and at least one engagement feature associated with at least one media asset of the plurality media assets. The at least one content feature may comprise at least one: a content type, a content rating, content metadata, a date of creation, a content tag, a content category, a content filter, a language, or one or more words of a content description. The at least one engagement feature of each feature vector may comprise at least one of: a quantification of an engagement with the at least one media asset for a numerical weight associated with an engagement type.”).
Therefore, it would have been obvious to one of ordinary skill in the art, at the time the invention was made, to add Sahasi’s performing, based on the probabilities of the candidate advertisement belonging to different reference advertisement types, weighted processing on the input feature of the candidate advertisement to obtain an input feature of the candidate advertisement under each reference advertisement type to Yan’s determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request.  One of ordinary skill in the art would have been motivated to do so in order to allow “for improved model selection and content recommendations” (Sahasi: ¶ [0002]).
With respect to Claim 13:
All limitations as recited have been analyzed and rejected to claim 2. Claim 13 does not teach or define any new limitations beyond claim 2. Therefore it is rejected under the same rationale.

With respect to Claim 3:
Yan teaches:
The method according to claim 1, wherein determining the competition score of the candidate advertisement comprises: determining an input feature of the candidate advertisement according to the advertisement state corresponding to the candidate advertisement and the overall state (i.e. advertiser specifies/determines targeting criteria for advertisement slot) (Yan: ¶¶ [0032] [0033] “Additionally, an advertisement request may include one or more targeting criteria specified by the advertiser. Targeting criteria included in an advertisement request specify one or more characteristics of users eligible to be presented with advertisement content in the advertisement request. For example, targeting criteria are used to identify users having user profile information, edges, or actions satisfying at least one of the targeting criteria. Hence, targeting criteria allow an advertiser to identify users having specific characteristics, simplifying subsequent distribution of content to different users…In one embodiment, the targeting criteria may specify actions or types of connections between a user and another user or object of the online system 140. Targeting criteria may also specify interactions between a user and objects performed external to the online system 140, such as on a third party system 130.”); 
configuring a competition score for the candidate advertisement by each of the scoring networks in the scoring model according to the input feature of the candidate advertisement (i.e. configuring the affinity score according to target criteria corresponding to cosine similarity of candidate advertisement being similar/different to different advertisements, wherein the scoring includes different scoring network or rules corresponding to different advertisement features/type to determine affinity score) (Yan: ¶ [0037] “Based at least in part on the correlation between content provided by one or more third party systems 130 accessed by the viewing user and content provided by one or more third party systems 130 accessed by one or more additional users who interacted with an advertisement, the correlation module 235 determines a score for the advertisement. For example, the greater the cosine similarity between content provided by one or more third party systems 130 accessed by the viewing user and content provided by one or more third party systems 130 accessed by one or more additional users who interacted with an advertisement, the higher the score computed for the advertisement. In some embodiments, the score for an advertisement may also be determined based content provided by one or more third party systems 130 with which the viewing user accessed and interactions with the advertisement by additional users who also accessed the content provided by the one or more third party systems 130… For example, a score associated with an advertisement is proportional to a percentage of online system users presented with the advertisement who performed a specific type of interaction with the advertisement (e.g., accessed the advertisement, made a purchase of a product identified by the advertisement within a threshold time of interacting with the advertisement), so the greater the percentage of users performing the specific type of interaction, the greater the score for the advertisement. As an additional example, the score for an advertisement is proportional to an amount of revenue generated from purchases made by users in association with accessing the advertisement.” Furthermore, as cited in ¶ [0057] “Additional factors may be used by the online system 140 when generating 355 the score for an advertisement. For example, content provided by third party systems 130 accessed by the viewing user is used when generating 355 the score for an advertisement. For example, if the viewing user accesses content from a product page of an online merchant more frequently than content from other third party systems 130, a score of an ad request including an advertisement with which additional users who accessed product page interacted is generated 355 in part based on a percentage of the additional users who accessed the product page who also interacted with the advertisement. As an additional example, if the viewing user accesses particular web page, scores for ad requests including advertisements with which at least a threshold number or percentage of additional users who accessed the web page interacted are generated 355 based at least in part on the number of times or frequency with which the viewing user accessed the particular web page. A score generated 355 for an ad request including an advertisement may additionally be based on an overall performance of the advertisement when presented to online system users as a whole. For example, a score generated 355 for an ad request is proportional to a number or a percentage of online system users presented with an advertisement in the ad request who performed one or more types of interactions with the advertisement, so ad requests including advertisements with which a larger number or percentage of online system users presented with the advertisements interacted have higher scores.”).
Yan does not explicitly disclose performing weighted summation processing on the competition score configured for the candidate advertisement by each of the scoring networks based on the probabilities of the candidate advertisement belonging to different reference advertisement types, to obtain a competition score of the candidate advertisement for the current exposure request.
However, Sahasi further discloses performing weighted summation processing on the competition score configured for the candidate advertisement by each of the scoring networks based on the probabilities of the candidate advertisement belonging to different reference advertisement types, to obtain a competition score of the candidate advertisement for the current exposure request (i.e. performing weighted summation based on probabilities of content feature belonging to group in order to determine score) (Sahasi: ¶ [0059] “A feature vector may be associated with a particular user device(s). A feature vector may comprise a quantification of a level/amount of engagement with each media asset and/or a numerical weight associated with an engagement feature as described herein. The number and arrangement of items in such a feature vector may be the same as those of features vectors used during training of the scoring model 248. The scoring unit 230 can then apply the scoring model 248 to the feature vector to generate an interest attribute representing a level of interest on the media asset. The interest attribute can be a numerical value (e.g., an integer number) or textual label that indicates the level of interest ( e.g., "high", "moderate", and "low").” Furthermore, as cited in ¶ [0168] “Generally, recall refers to a ratio of true positives to a sum of true positives and false negatives, which quantifies a sensitivity of the machine learning model 1630. Similarly, precision refers to a ratio of true positives a sum of true and false positives. When such a desired accuracy level is reached, the training phase ends and the machine learning model 1630 may be output at step 1790; when the desired accuracy level is not reached, however, then a subsequent iteration of the training method 1700 may be performed starting at step 1610 with variations such as, for example, considering a larger collection of media assets. The machine learning model 1630 may be output at step 1790. The machine learning model 1630 may be configured to determine predicted predictions for media assets that are not within the plurality of media assets used to train the machine learning model.”).
Therefore, it would have been obvious to one of ordinary skill in the art, at the time the invention was made, to add Sahasi’s performing weighted summation processing on the competition score configured for the candidate advertisement by each of the scoring networks based on the probabilities of the candidate advertisement belonging to different reference advertisement types, to obtain a competition score of the candidate advertisement for the current exposure request to Yan’s determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request.  One of ordinary skill in the art would have been motivated to do so in order to allow “for improved model selection and content recommendations” (Sahasi: ¶ [0002]). 
With respect to Claim 14:
All limitations as recited have been analyzed and rejected to claim 3. Claim 14 does not teach or define any new limitations beyond claim 3. Therefore it is rejected under the same rationale.

With respect to Claim 4:
Yan teaches:
The method according to claim 1, wherein determining the competition score of the candidate advertisement comprises: determining an input feature of the candidate advertisement according to the advertisement state corresponding to the candidate advertisement and the overall state (i.e. advertiser specifies/determines targeting criteria for advertisement slot) (Yan: ¶¶ [0032] [0033] “Additionally, an advertisement request may include one or more targeting criteria specified by the advertiser. Targeting criteria included in an advertisement request specify one or more characteristics of users eligible to be presented with advertisement content in the advertisement request. For example, targeting criteria are used to identify users having user profile information, edges, or actions satisfying at least one of the targeting criteria. Hence, targeting criteria allow an advertiser to identify users having specific characteristics, simplifying subsequent distribution of content to different users…In one embodiment, the targeting criteria may specify actions or types of connections between a user and another user or object of the online system 140. Targeting criteria may also specify interactions between a user and objects performed external to the online system 140, such as on a third party system 130.”); and 
determining a competition score of the candidate advertisement for the current exposure request by the scoring network corresponding to the candidate advertisement according to the input feature of the candidate advertisement (i.e. determining the affinity score according to target criteria corresponding to cosine similarity of candidate advertisement being similar/different to different advertisements, wherein the scoring includes different scoring network or rules corresponding to different advertisement features/type to determine affinity score) (Yan: ¶ [0037] “Based at least in part on the correlation between content provided by one or more third party systems 130 accessed by the viewing user and content provided by one or more third party systems 130 accessed by one or more additional users who interacted with an advertisement, the correlation module 235 determines a score for the advertisement. For example, the greater the cosine similarity between content provided by one or more third party systems 130 accessed by the viewing user and content provided by one or more third party systems 130 accessed by one or more additional users who interacted with an advertisement, the higher the score computed for the advertisement. In some embodiments, the score for an advertisement may also be determined based content provided by one or more third party systems 130 with which the viewing user accessed and interactions with the advertisement by additional users who also accessed the content provided by the one or more third party systems 130… For example, a score associated with an advertisement is proportional to a percentage of online system users presented with the advertisement who performed a specific type of interaction with the advertisement (e.g., accessed the advertisement, made a purchase of a product identified by the advertisement within a threshold time of interacting with the advertisement), so the greater the percentage of users performing the specific type of interaction, the greater the score for the advertisement. As an additional example, the score for an advertisement is proportional to an amount of revenue generated from purchases made by users in association with accessing the advertisement.” Furthermore, as cited in ¶ [0057] “Additional factors may be used by the online system 140 when generating 355 the score for an advertisement. For example, content provided by third party systems 130 accessed by the viewing user is used when generating 355 the score for an advertisement. For example, if the viewing user accesses content from a product page of an online merchant more frequently than content from other third party systems 130, a score of an ad request including an advertisement with which additional users who accessed product page interacted is generated 355 in part based on a percentage of the additional users who accessed the product page who also interacted with the advertisement. As an additional example, if the viewing user accesses particular web page, scores for ad requests including advertisements with which at least a threshold number or percentage of additional users who accessed the web page interacted are generated 355 based at least in part on the number of times or frequency with which the viewing user accessed the particular web page. A score generated 355 for an ad request including an advertisement may additionally be based on an overall performance of the advertisement when presented to online system users as a whole. For example, a score generated 355 for an ad request is proportional to a number or a percentage of online system users presented with an advertisement in the ad request who performed one or more types of interactions with the advertisement, so ad requests including advertisements with which a larger number or percentage of online system users presented with the advertisements interacted have higher scores.”).
Yan does not explicitly disclose determining a scoring network corresponding to the candidate advertisement in the scoring model based on the probabilities of the candidate advertisement belonging to different reference advertisement types.
However, Sahasi further discloses determining a scoring network corresponding to the candidate advertisement in the scoring model based on the probabilities of the candidate advertisement belonging to different reference advertisement types (i.e. determine different models that correspond to different ad feature sets, wherein the feature sets reads on reference advertisement types) (Sahasi: ¶¶ [0154] [0155] “The training module 1620 may extract a feature set from the training dataset 1610A and/or the training dataset 1610B in a variety of ways. For example, the training module 1620 may extract a feature set from the training dataset 1610A and/or the training dataset 1610B using a classification model (e.g., a machine learning model). The training module 1620 may perform feature extraction multiple times, each time using a different feature-extraction technique. In one example, the feature sets generated using the different techniques may each be used to generate different machine learning models 1640. For example, the feature set with the highest quality features (e.g., most indicative of interest or not of interest to a particular user(s)) may be selected for use in training… The training dataset 1610A and/or the training dataset 1610B may be analyzed to determine any dependencies, associations, and/or correlations between features and the labeled predictions in the training dataset 1610A and/or the training dataset 1610B. The identified correlations may have the form of a list of features that are associated with different labeled predictions (e.g., of interest to a particular user vs. not of interest to a particular user). The term "feature", as used herein, may refer to any characteristic of an item of data that may be used to determine whether the item of data falls within one or more specific categories or within a range. By way of example, the features described herein may comprise one or more features present within each of the media assets that may be correlative (or not correlative as the case may be) with a particular media asset being of interest to a particular user or not.”).
Therefore, it would have been obvious to one of ordinary skill in the art, at the time the invention was made, to add Sahasi’s determining a scoring network corresponding to the candidate advertisement in the scoring model based on the probabilities of the candidate advertisement belonging to different reference advertisement types to Yan’s determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request.  One of ordinary skill in the art would have been motivated to do so in order to allow “for improved model selection and content recommendations” (Sahasi: ¶ [0002]).
With respect to Claim 15:
All limitations as recited have been analyzed and rejected to claim 4. Claim 15 does not teach or define any new limitations beyond claim 4. Therefore it is rejected under the same rationale.

With respect to Claim 5:
Yan teaches the method according to claim 1, wherein determining the probabilities of the candidate advertisement comprises one or more of: determining, by the classification network, probabilities of the candidate advertisement belonging to different reference advertisement types according to the advertisement state corresponding to the candidate advertisement and the overall state (i.e. determining cosine similarity of candidate advertisement being similar/different to different advertisements based on advertisement features and/or revenue generated by advertisements) (Yan: ¶ [0037] “Based at least in part on the correlation between content provided by one or more third party systems 130 accessed by the viewing user and content provided by one or more third party systems 130 accessed by one or more additional users who interacted with an advertisement, the correlation module 235 determines a score for the advertisement. For example, the greater the cosine similarity between content provided by one or more third party systems 13 0 accessed by the viewing user and content provided by one or more third party systems 130 accessed by one or more additional users who interacted with an advertisement, the higher the score computed for the advertisement. In some embodiments, the score for an advertisement may also be determined based content provided by one or more third party systems 130 with which the viewing user accessed and interactions with the advertisement by additional users who also accessed the content provided by the one or more third party systems 130… For example, a score associated with an advertisement is proportional to a percentage of online system users presented with the advertisement who performed a specific type of interaction with the advertisement (e.g., accessed the advertisement, made a purchase of a product identified by the advertisement within a threshold time of interacting with the advertisement), so the greater the percentage of users performing the specific type of interaction, the greater the score for the advertisement. As an additional example, the score for an advertisement is proportional to an amount of revenue generated from purchases made by users in association with accessing the advertisement.”).
Yan does not explicitly disclose determining, by the classification network, probabilities of the candidate advertisement belonging to different reference advertisement types according to the advertisement state corresponding to the candidate advertisement; and determining, by the classification network, probabilities of the candidate advertisement belonging to different reference advertisement types according to an advertisement feature corresponding to the candidate advertisement.
However, Sahasi further discloses:
determining, by the classification network, probabilities of the candidate advertisement belonging to different reference advertisement types according to the advertisement state corresponding to the candidate advertisement (i.e. determining/predicting label according to features or different reference advertisement types, wherein the predicted labels reads on probabilities of candidate ad belonging to different reference types because the media asset is predicted to correspond to labeled ad feature or advertisement state) (Sahasi: ¶¶ [0151] [0152] “described herein may be referred to as "at least one machine learning model 1630" or simply the "machine learning model 1630", as shown in FIG. 16 The at least one machine learning model 1630 may be trained by a system 1600 shown in FIG. 16. The system 1600 may be configured to use machine learning techniques to train, based on an analysis of one or more training datasets 1610A-1610B by a training module 1620, the at least one machine learning model 1630. The at least one machine learning model 1630, once trained, may be configured to determine a prediction that a media asset is of interest to a particular user or not of interest to the particular user. A dataset indicative of a plurality of media assets and a labeled ( e.g., predetermined/known) prediction indicating whether the corresponding media assets are of interest to a particular user or not may be used by the training module 1620 to train the at least one machine learning model 1630. Each of the plurality of media assets in the dataset may be associated with a plurality of features that are present within each corresponding media asset. The plurality of features and the labeled predictions may be used to train the at least one machine learning model 1630…The training dataset 1610A may comprise a first portion of the plurality of media assets in the dataset. Each media asset in the first portion may have a labeled (e.g., predetermined) prediction and one or more labeled features. The training dataset 1610B may comprise a second portion of the plurality of media assets in the dataset. Each media asset in the second portion may have a labeled (e.g., predetermined) prediction and one or more labeled features. The plurality of media assets may be randomly assigned to the training dataset 1610A, the training dataset 1610B, and/or to a testing dataset.”); and 
determining, by the classification network, probabilities of the candidate advertisement belonging to different reference advertisement types according to an advertisement feature corresponding to the candidate advertisement (i.e. determining/predicting label according to features or different reference advertisement types, wherein the predicted labels reads on probabilities of candidate ad belonging to different reference types because the media asset is predicted to correspond to labeled ad feature or advertisement state) (Sahasi: ¶¶ [0151] [0152] “described herein may be referred to as "at least one machine learning model 1630" or simply the "machine learning model 1630", as shown in FIG. 16 The at least one machine learning model 1630 may be trained by a system 1600 shown in FIG. 16. The system 1600 may be configured to use machine learning techniques to train, based on an analysis of one or more training datasets 1610A-1610B by a training module 1620, the at least one machine learning model 1630. The at least one machine learning model 1630, once trained, may be configured to determine a prediction that a media asset is of interest to a particular user or not of interest to the particular user. A dataset indicative of a plurality of media assets and a labeled ( e.g., predetermined/known) prediction indicating whether the corresponding media assets are of interest to a particular user or not may be used by the training module 1620 to train the at least one machine learning model 1630. Each of the plurality of media assets in the dataset may be associated with a plurality of features that are present within each corresponding media asset. The plurality of features and the labeled predictions may be used to train the at least one machine learning model 1630…The training dataset 1610A may comprise a first portion of the plurality of media assets in the dataset. Each media asset in the first portion may have a labeled (e.g., predetermined) prediction and one or more labeled features. The training dataset 1610B may comprise a second portion of the plurality of media assets in the dataset. Each media asset in the second portion may have a labeled (e.g., predetermined) prediction and one or more labeled features. The plurality of media assets may be randomly assigned to the training dataset 1610A, the training dataset 1610B, and/or to a testing dataset.”).
Therefore, it would have been obvious to one of ordinary skill in the art, at the time the invention was made, to add Sahasi’s determining, by the classification network, probabilities of the candidate advertisement belonging to different reference advertisement types according to the advertisement state corresponding to the candidate advertisement and the overall state; determining, by the classification network, probabilities of the candidate advertisement belonging to different reference advertisement types according to the advertisement state corresponding to the candidate advertisement; and determining, by the classification network, probabilities of the candidate advertisement belonging to different reference advertisement types according to an advertisement feature corresponding to the candidate advertisement to Yan’s determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request.  One of ordinary skill in the art would have been motivated to do so in order to allow “for improved model selection and content recommendations” (Sahasi: ¶ [0002]).
With respect to Claim 16:
All limitations as recited have been analyzed and rejected to claim 5. Claim 16 does not teach or define any new limitations beyond claim 5. Therefore it is rejected under the same rationale.

With respect to Claim 6:
Yan teaches:
The method according to claim 1, wherein an advertisement state corresponding to the contract advertisement comprises a competition environment in response to that the contract advertisement competes for the current exposure request, which is determined according to advertisement features of other advertisements in the plurality of candidate advertisements except the contract advertisement (i.e. advertisement features includes targeting criteria in response to ad request, wherein the advertisement features are included from other advertisement than the recommended ad) (Yan: ¶ [0047] “For example, the online system 140 maintains a table identifying different users of the online system 140 and different content provided by one or more third party systems 130 and stores information in the table corresponding to a pairing of the viewing user and the content from the third party system 130 accessed 305 by the viewing user. The information in the table may be a binary value indicating whether the viewing user accessed 305 the content from the third party system 130 or may be information identifying a number of times the viewing user accessed 305 the content from the third party system 130. As another example, the online system 140 stores a vector on a graph associated with the viewing user, where different dimensions of the vector correspond to content from one or more third party systems 130 accessed 305 by the viewing user.” Furthermore, as cited in ¶ [0050] “Each column of the table is associated different content 430 provided by one or more third party systems 130 or is associated with a different advertisement 440. While FIG. 4 shows an example table including content 430 provided by third party systems 130 as well as advertisements 440, in other embodiments, different tables are maintained for content 430 provided by third party systems 130 and for advertisements 440. In other embodiments, different vectors are associated with each user describing each user's interactions with advertisements 440, where the dimensions of a vector correspond to advertisements with which a user interacted and a number of interactions with the advertisements by the user.”); 
the advertisement state corresponding to the contract advertisement further comprises one or more of: a playing amount, a shortage, the predetermined playing amount, the selling price, a playing control parameter, and the targeting condition of the contract advertisement (i.e. advertisement state/information corresponds to how many times the advertisement was played and a specified number of interactions, the amount of money the advertiser will be compensated for displaying the ad and represents minimum amount of compensation, parameters that control how the advertisement is played such as specified interactions, and targeting conditions of the advertisement) (Yan: ¶ [0051] “In the example of FIG. 4, an integer in the table at an intersection of a column and row represents a number of times that a user corresponding to the row accessed content from a third party system 130 corresponding to the column or interacted with an advertisement corresponding to the column. For example, in FIG. 4, columns 435A-435G each correspond to different content from one or more third party systems 130…For example, a table includes user interactions with advertisements and content provided by one or more third party systems 130 users accessed within three months of a current date. As another example, the table includes a specified amount of content accessed by online system users ( e.g., the most recent 1000 web pages accessed by different online system users) or includes a specified number of interactions with advertisements by online system users (e.g., 500 most recent interactions with advertisements).” Furthermore, as cited in ¶ [0031] “The bid amount is associated with an advertisement by an advertiser and is used to determine an expected value, such as monetary compensation, provided by an advertiser to the online system 140 if advertisement content in the ad request is presented to a user, if the advertisement content in the ad request receives a user interaction when presented, or if any suitable condition is satisfied when advertisement content in the ad request is presented to a user. For example, the bid amount specifies a monetary amount that the online system 140 receives from the advertiser if advertisement content in an ad request is displayed.” Furthermore, as cited in ¶ [0037] “For example, a score associated with an advertisement is proportional to a percentage of online system users presented with the advertisement who performed a specific type of interaction with the advertisement (e.g., accessed the advertisement, made a purchase of a product identified by the advertisement within a threshold time of interacting with the advertisement), so the greater the percentage of users performing the specific type of interaction, the greater the score for the advertisement. As an additional example, the score for an advertisement is proportional to an amount of revenue generated from purchases made by users in association with accessing the advertisement. Scoring advertisements is further described below in conjunction with FIG. 3.” Furthermore, as cited in ¶ [0032] “Targeting criteria included in an advertisement request specify one or more characteristics of users eligible to be presented with advertisement content in the advertisement request. For example, targeting criteria are used to identify users having user profile information, edges, or actions satisfying at least one of the targeting criteria. Hence, targeting criteria allow an advertiser to identify users having specific characteristics, simplifying subsequent distribution of content to different users.”); and 
an advertisement state corresponding to the bid advertisement comprises a competition environment in response to that the bid advertisement competes for the current exposure request, which is determined according to advertisement features of other advertisements in the plurality of candidate advertisements except the bid advertisement (i.e. advertisement includes bid amount in response to ad request, wherein the bid amounts are included from other advertisement than the recommended ad) (Yan: ¶ [0039] “Content items selected for presentation to the user may include ad requests or other content items associated with bid amounts. The content selection module 240 uses the bid amounts associated with ad requests when selecting content for presentation to the viewing user. In various embodiments, the content selection module 240 determines an expected value associated with various ad requests (or other content items) based on their bid amounts and selects content items associated with a maximum expected value or associated with at least a threshold expected value for presentation. An expected value associated with an ad request or with a content item represents an expected amount of compensation to the online system 140 for presenting an ad request or a content item. For example, the expected value associated with an ad request is a product of the ad request's bid amount and a likelihood of the user interacting with the ad content from the ad request. The content selection module 240 may rank ad requests based on their associated bid amounts and select ad requests having at least a threshold position in the ranking for presentation to the user. In some embodiments, the content selection module 240 ranks both content items not associated with bid amounts and ad requests in a unified ranking based on bid amounts associated with ad requests and measures of relevance associated with content items and ad requests.” Furthermore, as cited in ¶ [0050] “Each column of the table is associated different content 430 provided by one or more third party systems 130 or is associated with a different advertisement 440. While FIG. 4 shows an example table including content 430 provided by third party systems 130 as well as advertisements 440, in other embodiments, different tables are maintained for content 430 provided by third party systems 130 and for advertisements 440. In other embodiments, different vectors are associated with each user describing each user's interactions with advertisements 440, where the dimensions of a vector correspond to advertisements with which a user interacted and a number of interactions with the advertisements by the user.”).
With respect to Claim 17:
All limitations as recited have been analyzed and rejected to claim 6. Claim 17 does not teach or define any new limitations beyond claim 6. Therefore it is rejected under the same rationale.

With respect to Claim 7:
Yan does not explicitly disclose the method according to claim 1, wherein the scoring model is trained by: simulating a virtual advertising platform based on historical data of the advertising platform; determining a training candidate advertisement corresponding to a training exposure request on the virtual advertising platform; determining, by an initial scoring model to be trained, a training competition score of each training candidate advertisement for the training exposure request according to an advertisement state corresponding to the training candidate advertisement and an overall state of the virtual advertising platform, the initial scoring model comprising an initial classification network and multiple initial scoring networks corresponding to different reference advertisement types; determining a training target advertisement exposed by the training exposure request according to the training competition score of each training candidate advertisement for the training exposure request, and simulating a training reward generated by the virtual advertising platform exposing the training target advertisement; determining, by a judgment model, feedback information corresponding to a current round of scoring operation of the initial scoring model according to the overall state of the virtual advertising platform after exposing the training target advertisement and the training reward, and inputting the feedback information into the initial scoring model as reference information in response to that the initial scoring model scores the training candidate advertisement corresponding to the training exposure request in a next round, so as to assist in adjusting a model parameter of the initial scoring model; and determining the initial scoring model as the scoring model in response to confirming that a training end condition is satisfied.
However, Sahasi further discloses:
simulating a virtual advertising platform based on historical data of the advertising platform (i.e. training model based on historical data of advertising platform) (Sahasi: ¶ [0142] “The client model 1550 may have been trained using historical user activity data, as further described herein. The client model 1550 may determine the at least one content recommendation 1560 based on activity/interest data 1540 associated with other users. For example, the activity/interest data 1540 may comprise activity data for one or more other users who engaged with one or more media assets that the user 1510 engaged with.”); 
determining a training candidate advertisement corresponding to a training exposure request on the virtual advertising platform (i.e. determining content to recommend based on training data set being performed via model) (Sahasi: ¶ [0150] “The training datasets may comprise UICs associated with users who interacted with (e.g., engaged with) a plurality of media assets. The UICs that are used during training and/or retraining may comprise interest attributes, interest levels, functionality features, a content features, a combination thereof, and/or the like. A training module, such as the training module 1620 shown in FIG. 16, may then determine which features in the UICs correlate with the particular features of the plurality of media assets. The machine learning models, once trained ( or retrained as the case may be), may provide a recommendation for a user(s) and a media asset(s) based on the corresponding UIC(s) and the features of that media asset.”); 
determining, by an initial scoring model to be trained, a training competition score of each training candidate advertisement for the training exposure request according to an advertisement state corresponding to the training candidate advertisement and an overall state of the virtual advertising platform, the initial scoring model comprising an initial classification network and multiple initial scoring networks corresponding to different reference advertisement types (i.e. training model determines a score according to advertisement features which includes content features and features of the advertiser such as revenue or business type) (Sahasi: ¶¶ [0154] [0155] “The training module 1620 may extract a feature set from the training dataset 1610A and/or the training dataset 1610B in a variety of ways. For example, the training module 1620 may extract a feature set from the training dataset 1610A and/or the training dataset 1610B using a classification model (e.g., a machine learning model). The training module 1620 may perform feature extraction multiple times, each time using a different feature-extraction technique. In one example, the feature sets generated using the different techniques may each be used to generate different machine learning models 1640. For example, the feature set with the highest quality features (e.g., most indicative of interest or not of interest to a particular user(s)) may be selected for use in training… The training dataset 1610A and/or the training dataset 1610B may be analyzed to determine any dependencies, associations, and/or correlations between features and the labeled predictions in the training dataset 1610A and/or the training dataset 1610B. The identified correlations may have the form of a list of features that are associated with different labeled predictions (e.g., of interest to a particular user vs. not of interest to a particular user). The term "feature", as used herein, may refer to any characteristic of an item of data that may be used to determine whether the item of data falls within one or more specific categories or within a range. By way of example, the features described herein may comprise one or more features present within each of the media assets that may be correlative (or not correlative as the case may be) with a particular media asset being of interest to a particular user or not.” Furthermore, as cited in ¶ [0107] “To that end, the content selection unit 1210 can direct an ingestion unit 1220 to obtain a group of directed content assets from directed content storage 1280 retaining a corpus of directed content assets 1284. In some cases, the corpus of directed content assets 1264 can be categorized according to attributes of an end-user. The attributes can include, for example, market type, market segment, geography, business size, business type, revenue, profits, and similar. Accordingly, for a particular user device for which the personalization is being implemented, the content selection unit 1210 can direct the ingestions unit 1220 to obtain directed content assets having a particular set of attributes.”); 
determining a training target advertisement exposed by the training exposure request according to the training competition score of each training candidate advertisement for the training exposure request, and simulating a training reward generated by the virtual advertising platform exposing the training target advertisement (i.e. determining a recommended content feature based on training model score, wherein the training model simulates or predicts a value of the advertisement) (Sahasi: ¶¶ [0165]-[0167] “The training method 1700 may determine (e.g., extract, select, etc.), at step 1730, one or more features that may be used by, for example, a classifier to differentiate among different classifications ( e.g., predictions/recommendations). The one or more features may comprise a set of features. As an example, the training method 1700 may determine a set features from the first media assets. As another example, the training method 1700 may determine a set of features from the second media assets. In a further example, a set of features may be determined from other media assets of the plurality of media assets ( e.g., a third portion) associated with a specific feature(s) and/or range(s) of predetermined predictions that may be different than the specific feature(s) and/or range(s) of predetermined predictions associated with the media assets of the training dataset and the testing dataset.…The machine learning models trained at step 17 40 may be selected based on different criteria depending on the problem to be solved and/or data available in the training dataset. For example, machine learning models may suffer from different degrees of bias. Accordingly, more than one machine learning model may be trained at 1740, and then optimized, improved, and cross-validated at step 1750...The training method 1700 may select one or more machine learning models to build the machine learning model 1630 at step 1760. The machine learning model 1630 may be evaluated using the testing dataset. The machine learning model 1630 may analyze the testing dataset and generate classification values and/or predicted values (e.g., predictions) at step 1770. Classification and/or prediction values may be evaluated at step 1780 to determine whether such values have achieved a desired accuracy level. Performance of the machine learning model 1630 may be evaluated in a number of ways based on a number of true positives, false positives, true negatives, and/or false negatives classifications of the plurality of data points indicated by the machine learning model 1630.”); 
determining, by a judgment model, feedback information corresponding to a current round of scoring operation of the initial scoring model according to the overall state of the virtual advertising platform after exposing the training target advertisement and the training reward, and inputting the feedback information into the initial scoring model as reference information in response to that the initial scoring model scores the training candidate advertisement corresponding to the training exposure request in a next round, so as to assist in adjusting a model parameter of the initial scoring model (i.e. determining false positives or negatives of the current round of machine learning model, wherein the false positives/negatives are used back in the machine learning model to adjust the model parameter to the desired level of accuracy) (Sahasi: ¶¶ [0167] [0168] “Classification and/or prediction values may be evaluated at step 1780 to determine whether such values have achieved a desired accuracy level. Performance of the machine learning model 1630 may be evaluated in a number of ways based on a number of true positives, false positives, true negatives, and/or false negatives classifications of the plurality of data points indicated by the machine learning model 1630…For example, the false positives of the machine learning model 1630 may refer to a number of times the machine learning model 1630 incorrectly assigned a high prediction to a media asset associated with a low predetermined prediction. Conversely, the false negatives of the machine learning model 1630 may refer to a number of times the machine learning model assigned a low prediction to a media asset associated with a high predetermined prediction. True negatives and true positives may refer to a number of times the machine learning model 1630 correctly assigned predictions to media assets based on the known, predetermined prediction for each media asset. Related to these measurements are the concepts of recall and precision. Generally, recall refers to a ratio of true positives to a sum of true positives and false negatives, which quantifies a sensitivity of the machine learning model 1630. Similarly, precision refers to a ratio of true positives a sum of true and false positives. When such a desired accuracy level is reached, the training phase ends and the machine learning model 1630 may be output at step 1790; when the desired accuracy level is not reached, however, then a subsequent iteration of the training method 1700 may be performed starting at step 1610 with variations such as, for example, considering a larger collection of media assets.”); and 
determining the initial scoring model as the scoring model in response to confirming that a training end condition is satisfied (i.e. determining the model in response to the desired level of accuracy being satisfied) (Sahasi: ¶ [0168] “Similarly, precision refers to a ratio of true positives a sum of true and false positives. When such a desired accuracy level is reached, the training phase ends and the machine learning model 1630 may be output at step 1790; when the desired accuracy level is not reached, however, then a subsequent iteration of the training method 1700 may be performed starting at step 1610 with variations such as, for example, considering a larger collection of media assets. The machine learning model 1630 may be output at step 1790. The machine learning model 1630 may be configured to determine predicted predictions for media assets that are not within the plurality of media assets used to train the machine learning model.”).
Therefore, it would have been obvious to one of ordinary skill in the art, at the time the invention was made, to add Sahasi’s simulating a virtual advertising platform based on historical data of the advertising platform; determining a training candidate advertisement corresponding to a training exposure request on the virtual advertising platform; determining, by an initial scoring model to be trained, a training competition score of each training candidate advertisement for the training exposure request according to an advertisement state corresponding to the training candidate advertisement and an overall state of the virtual advertising platform, the initial scoring model comprising an initial classification network and multiple initial scoring networks corresponding to different reference advertisement types; determining a training target advertisement exposed by the training exposure request according to the training competition score of each training candidate advertisement for the training exposure request, and simulating a training reward generated by the virtual advertising platform exposing the training target advertisement; determining, by a judgment model, feedback information corresponding to a current round of scoring operation of the initial scoring model according to the overall state of the virtual advertising platform after exposing the training target advertisement and the training reward, and inputting the feedback information into the initial scoring model as reference information in response to that the initial scoring model scores the training candidate advertisement corresponding to the training exposure request in a next round, so as to assist in adjusting a model parameter of the initial scoring model; and determining the initial scoring model as the scoring model in response to confirming that a training end condition is satisfied to Yan’s determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request.  One of ordinary skill in the art would have been motivated to do so in order to allow “for improved model selection and content recommendations” (Sahasi: ¶ [0002]).
With respect to Claim 18:
All limitations as recited have been analyzed and rejected to claim 7. Claim 18 does not teach or define any new limitations beyond claim 7. Therefore it is rejected under the same rationale.

With respect to Claim 8:
Yan does not explicitly disclose the method according to claim 7, wherein determining the training competition score of each training candidate advertisement comprises: determining, by the initial classification network in the initial scoring model, probabilities of each training candidate advertisement belonging to different reference advertisement types; determining a target reference advertisement type to which the training candidate advertisement belongs according to the probabilities of the training candidate advertisement belonging to different reference advertisement types; and determining, by an initial scoring network corresponding to the target reference advertisement type in the initial scoring model, a training competition score of the training candidate advertisement for the training exposure request according to the advertisement state corresponding to the training candidate advertisement, the overall state of the virtual advertising platform, and reference information, the reference information being feedback information provided by the judgment model for a previous round of scoring operation on the initial scoring network, and the scoring operation being performed on the training candidate advertisement corresponding to the training exposure request.
However, Sahasi further discloses:
determining, by the initial classification network in the initial scoring model, probabilities of each training candidate advertisement belonging to different reference advertisement types (i.e. determining/predicting label according to features or different reference advertisement types, wherein the predicted labels reads on probabilities of candidate ad belonging to different reference types because the media asset is predicted to correspond to labeled ad feature) (Sahasi: ¶¶ [0151] [0152] “described herein may be referred to as "at least one machine learning model 1630" or simply the "machine learning model 1630", as shown in FIG. 16 The at least one machine learning model 1630 may be trained by a system 1600 shown in FIG. 16. The system 1600 may be configured to use machine learning techniques to train, based on an analysis of one or more training datasets 1610A-1610B by a training module 1620, the at least one machine learning model 1630. The at least one machine learning model 1630, once trained, may be configured to determine a prediction that a media asset is of interest to a particular user or not of interest to the particular user. A dataset indicative of a plurality of media assets and a labeled ( e.g., predetermined/known) prediction indicating whether the corresponding media assets are of interest to a particular user or not may be used by the training module 1620 to train the at least one machine learning model 1630. Each of the plurality of media assets in the dataset may be associated with a plurality of features that are present within each corresponding media asset. The plurality of features and the labeled predictions may be used to train the at least one machine learning model 1630…The training dataset 1610A may comprise a first portion of the plurality of media assets in the dataset. Each media asset in the first portion may have a labeled (e.g., predetermined) prediction and one or more labeled features. The training dataset 1610B may comprise a second portion of the plurality of media assets in the dataset. Each media asset in the second portion may have a labeled (e.g., predetermined) prediction and one or more labeled features. The plurality of media assets may be randomly assigned to the training dataset 1610A, the training dataset 1610B, and/or to a testing dataset.”); 
determining a target reference advertisement type to which the training candidate advertisement belongs according to the probabilities of the training candidate advertisement belonging to different reference advertisement types (i.e. determining target content features to which the training data belongs to according to the feature occurrence rules corresponding to predicted features belonging to different media assets or content features) (Sahasi: ¶¶ [0155]-[0157] “By way of example, the features described herein may comprise one or more features present within each of the media assets that may be correlative (or not correlative as the case may be) with a particular media asset being of interest to a particular user or not. As another example, the features described herein may comprise an interest attribute, an interest level, a functionality feature, or a content feature as further described and defined herein…A feature selection technique may comprise one or more feature selection rules. The one or more feature selection rules may comprise a feature occurrence rule. The feature occurrence rule may comprise determining which features in the training dataset 1610A occur over a threshold number of times and identifying those features that satisfy the threshold as candidate features. For example, any features that appear greater than or equal to 5 times in the training dataset 1610A may be considered as candidate features…A single feature selection rule may be applied to select features or multiple feature selection rules may be applied to select features. The feature selection rules may be applied in a cascading fashion, with the feature selection rules being applied in a specific order and applied to the results of the previous rule. For example, the feature occurrence rule may be applied to the training dataset 1610A to generate a first list of features. A final list of features may be analyzed according to additional feature selection techniques to determine one or more candidate feature groups (e.g., groups of features that may be used to determine a prediction).”); and 
determining, by an initial scoring network corresponding to the target reference advertisement type in the initial scoring model, a training competition score of the training candidate advertisement for the training exposure request according to the advertisement state corresponding to the training candidate advertisement, the overall state of the virtual advertising platform, and reference information, the reference information being feedback information provided by the judgment model for a previous round of scoring operation on the initial scoring network, and the scoring operation being performed on the training candidate advertisement corresponding to the training exposure request (i.e. determining scores or values based on the learning model, wherein the learning model takes into account advertisement features of trained data, overall state of advertising platform such as revenue or business type, and the feedback/reference information including false positives or negatives of the current round of machine learning model, wherein the false positives/negatives are used back in the machine learning model to adjust the model parameter to the desired level of accuracy and wherein the scoring model is conducted for the current ad request) (Sahasi: ¶¶ [0167] [0168] “Classification and/or prediction values may be evaluated at step 1780 to determine whether such values have achieved a desired accuracy level. Performance of the machine learning model 1630 may be evaluated in a number of ways based on a number of true positives, false positives, true negatives, and/or false negatives classifications of the plurality of data points indicated by the machine learning model 1630…For example, the false positives of the machine learning model 1630 may refer to a number of times the machine learning model 1630 incorrectly assigned a high prediction to a media asset associated with a low predetermined prediction. Conversely, the false negatives of the machine learning model 1630 may refer to a number of times the machine learning model assigned a low prediction to a media asset associated with a high predetermined prediction. True negatives and true positives may refer to a number of times the machine learning model 1630 correctly assigned predictions to media assets based on the known, predetermined prediction for each media asset. Related to these measurements are the concepts of recall and precision. Generally, recall refers to a ratio of true positives to a sum of true positives and false negatives, which quantifies a sensitivity of the machine learning model 1630. Similarly, precision refers to a ratio of true positives a sum of true and false positives. When such a desired accuracy level is reached, the training phase ends and the machine learning model 1630 may be output at step 1790; when the desired accuracy level is not reached, however, then a subsequent iteration of the training method 1700 may be performed starting at step 1610 with variations such as, for example, considering a larger collection of media assets.” Furthermore, as cited in ¶ [0107] “To that end, the content selection unit 1210 can direct an ingestion unit 1220 to obtain a group of directed content assets from directed content storage 1280 retaining a corpus of directed content assets 1284. In some cases, the corpus of directed content assets 1264 can be categorized according to attributes of an end-user. The attributes can include, for example, market type, market segment, geography, business size, business type, revenue, profits, and similar. Accordingly, for a particular user device for which the personalization is being implemented, the content selection unit 1210 can direct the ingestions unit 1220 to obtain directed content assets having a particular set of attributes.” Furthermore, as cited in ¶¶ [0140] [0141] “Each client model 1550 may comprise at least one of: the feature extraction unit 210, the activity monitoring unit 220, the scoring unit 230, the scoring model(s) 248, or the profile generation unit 250 as described herein. Each client model 1550 may be customized based on a number of media assets ( e.g., digital content) the particular client has uploaded/produced for distribution by the present distribution platform…The system shown in FIG. 15 may relate to an "online" or "live" service. For example, the "online" or "live" service may provide a recommendation for content (e.g., a media asset(s)) in real-time as users associated with the particular client are engaging with the client application 106.”).
Therefore, it would have been obvious to one of ordinary skill in the art, at the time the invention was made, to add Sahasi’s determining, by the initial classification network in the initial scoring model, probabilities of each training candidate advertisement belonging to different reference advertisement types; determining a target reference advertisement type to which the training candidate advertisement belongs according to the probabilities of the training candidate advertisement belonging to different reference advertisement types; and determining, by an initial scoring network corresponding to the target reference advertisement type in the initial scoring model, a training competition score of the training candidate advertisement for the training exposure request according to the advertisement state corresponding to the training candidate advertisement, the overall state of the virtual advertising platform, and reference information, the reference information being feedback information provided by the judgment model for a previous round of scoring operation on the initial scoring network, and the scoring operation being performed on the training candidate advertisement corresponding to the training exposure request to Yan’s determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request.  One of ordinary skill in the art would have been motivated to do so in order to allow “for improved model selection and content recommendations” (Sahasi: ¶ [0002]).
With respect to Claim 19:
All limitations as recited have been analyzed and rejected to claim 8. Claim 19 does not teach or define any new limitations beyond claim 9. Therefore it is rejected under the same rationale.

With respect to Claim 9:
Yan does not explicitly disclose the method according to claim 7, wherein determining the training competition score of each training candidate advertisement comprises: determining an input feature of each training candidate advertisement according to the advertisement state corresponding to the training candidate advertisement, the overall state of the virtual advertising platform, and reference information, the reference information being feedback information provided by the judgment model for a previous round of scoring operation on the initial scoring network, and the scoring operation being performed on the training candidate advertisement corresponding to the training exposure request; determining, by the initial classification network in the initial scoring model, probabilities of the training candidate advertisement belonging to different reference advertisement types; performing, based on the probabilities of the training candidate advertisement belonging to different reference advertisement types, weighted processing on the input feature of the training candidate advertisement to obtain an input feature of the training candidate advertisement under each reference advertisement type; and determining, by the initial scoring network in the initial scoring model, a training competition score of the training candidate advertisement for the training exposure request according to the input features of the training candidate advertisement under different reference advertisement types.
However, Sahasi further discloses:
determining an input feature of each training candidate advertisement according to the advertisement state corresponding to the training candidate advertisement, the overall state of the virtual advertising platform, and reference information, the reference information being feedback information provided by the judgment model for a previous round of scoring operation on the initial scoring network, and the scoring operation being performed on the training candidate advertisement corresponding to the training exposure request (i.e. determining advertisement features for a recommended advertisement based on the learning model, wherein the learning model takes into account advertisement features of trained data, overall state of advertising platform such as revenue or business type, and the feedback/reference information including false positives or negatives of the current round of machine learning model, wherein the false positives/negatives are used back in the machine learning model to adjust the model parameter to the desired level of accuracy and wherein the scoring model is conducted for the current ad request) (Sahasi: ¶ [0154] “The training module 1620 may perform feature extraction multiple times, each time using a different feature-extraction technique. In one example, the feature sets generated using the different techniques may each be used to generate different machine learning models 1640. For example, the feature set with the highest quality features (e.g., most indicative of interest or not of interest to a particular user(s)) may be selected for use in training. The training module 1620 may use the feature set( s) to build one or more machine learning models 1640A-1640N that are configured to determine a prediction for a new, unseen media asset.” Furthermore, as cited in ¶¶ [0167] [0168] “Classification and/or prediction values may be evaluated at step 1780 to determine whether such values have achieved a desired accuracy level. Performance of the machine learning model 1630 may be evaluated in a number of ways based on a number of true positives, false positives, true negatives, and/or false negatives classifications of the plurality of data points indicated by the machine learning model 1630…For example, the false positives of the machine learning model 1630 may refer to a number of times the machine learning model 1630 incorrectly assigned a high prediction to a media asset associated with a low predetermined prediction. Conversely, the false negatives of the machine learning model 1630 may refer to a number of times the machine learning model assigned a low prediction to a media asset associated with a high predetermined prediction. True negatives and true positives may refer to a number of times the machine learning model 1630 correctly assigned predictions to media assets based on the known, predetermined prediction for each media asset. Related to these measurements are the concepts of recall and precision. Generally, recall refers to a ratio of true positives to a sum of true positives and false negatives, which quantifies a sensitivity of the machine learning model 1630. Similarly, precision refers to a ratio of true positives a sum of true and false positives. When such a desired accuracy level is reached, the training phase ends and the machine learning model 1630 may be output at step 1790; when the desired accuracy level is not reached, however, then a subsequent iteration of the training method 1700 may be performed starting at step 1610 with variations such as, for example, considering a larger collection of media assets.” Furthermore, as cited in ¶ [0107] “To that end, the content selection unit 1210 can direct an ingestion unit 1220 to obtain a group of directed content assets from directed content storage 1280 retaining a corpus of directed content assets 1284. In some cases, the corpus of directed content assets 1264 can be categorized according to attributes of an end-user. The attributes can include, for example, market type, market segment, geography, business size, business type, revenue, profits, and similar. Accordingly, for a particular user device for which the personalization is being implemented, the content selection unit 1210 can direct the ingestions unit 1220 to obtain directed content assets having a particular set of attributes.” Furthermore, as cited in ¶¶ [0140] [0141] “Each client model 1550 may comprise at least one of: the feature extraction unit 210, the activity monitoring unit 220, the scoring unit 230, the scoring model(s) 248, or the profile generation unit 250 as described herein. Each client model 1550 may be customized based on a number of media assets ( e.g., digital content) the particular client has uploaded/produced for distribution by the present distribution platform…The system shown in FIG. 15 may relate to an "online" or "live" service. For example, the "online" or "live" service may provide a recommendation for content (e.g., a media asset(s)) in real-time as users associated with the particular client are engaging with the client application 106.”); 
determining, by the initial classification network in the initial scoring model, probabilities of the training candidate advertisement belonging to different reference advertisement types (i.e. determining/predicting label according to features or different reference advertisement types, wherein the predicted labels reads on probabilities of candidate ad belonging to different reference types because the media asset is predicted to correspond to labeled ad feature) (Sahasi: ¶¶ [0151] [0152] “described herein may be referred to as "at least one machine learning model 1630" or simply the "machine learning model 1630", as shown in FIG. 16 The at least one machine learning model 1630 may be trained by a system 1600 shown in FIG. 16. The system 1600 may be configured to use machine learning techniques to train, based on an analysis of one or more training datasets 1610A-1610B by a training module 1620, the at least one machine learning model 1630. The at least one machine learning model 1630, once trained, may be configured to determine a prediction that a media asset is of interest to a particular user or not of interest to the particular user. A dataset indicative of a plurality of media assets and a labeled ( e.g., predetermined/known) prediction indicating whether the corresponding media assets are of interest to a particular user or not may be used by the training module 1620 to train the at least one machine learning model 1630. Each of the plurality of media assets in the dataset may be associated with a plurality of features that are present within each corresponding media asset. The plurality of features and the labeled predictions may be used to train the at least one machine learning model 1630…The training dataset 1610A may comprise a first portion of the plurality of media assets in the dataset. Each media asset in the first portion may have a labeled (e.g., predetermined) prediction and one or more labeled features. The training dataset 1610B may comprise a second portion of the plurality of media assets in the dataset. Each media asset in the second portion may have a labeled (e.g., predetermined) prediction and one or more labeled features. The plurality of media assets may be randomly assigned to the training dataset 1610A, the training dataset 1610B, and/or to a testing dataset.”); 
performing, based on the probabilities of the training candidate advertisement belonging to different reference advertisement types, weighted processing on the input feature of the training candidate advertisement to obtain an input feature of the training candidate advertisement under each reference advertisement type (i.e. performing, based on predicted similar content features, weighted processing on advertisement features to determine recommended advertisement features) (Sahasi: ¶ [0059] “A feature vector may be associated with a particular user device(s). A feature vector may comprise a quantification of a level/amount of engagement with each media asset and/or a numerical weight associated with an engagement feature as described herein. The number and arrangement of items in such a feature vector may be the same as those of features vectors used during training of the scoring model 248. The scoring unit 230 can then apply the scoring model 248 to the feature vector to generate an interest attribute representing a level of interest on the media asset. The interest attribute can be a numerical value (e.g., an integer number) or textual label that indicates the level of interest ( e.g., "high", "moderate", and "low").” Furthermore, as cited in ¶ [0184] “At step 2020, the computing device may determine a feature vector associated with each user device of the plurality of user devices. For example, the computing device may determine the feature vector based on the plurality of activity data and the plurality of engagements. Each feature vector associated with each user device of the plurality of user devices may comprise at least one content feature and at least one engagement feature associated with at least one media asset of the plurality media assets. The at least one content feature may comprise at least one: a content type, a content rating, content metadata, a date of creation, a content tag, a content category, a content filter, a language, or one or more words of a content description. The at least one engagement feature of each feature vector may comprise at least one of: a quantification of an engagement with the at least one media asset for a numerical weight associated with an engagement type.”); and 
determining, by the initial scoring network in the initial scoring model, a training competition score of the training candidate advertisement for the training exposure request according to the input features of the training candidate advertisement under different reference advertisement types (i.e. training model determines a score or value according to advertisement features which includes content features and features of the advertiser such as revenue or business type, wherein the different advertisement features correspond to different advertisement reference types) (Sahasi: ¶¶ [0154] [0155] “The training module 1620 may extract a feature set from the training dataset 1610A and/or the training dataset 1610B in a variety of ways. For example, the training module 1620 may extract a feature set from the training dataset 1610A and/or the training dataset 1610B using a classification model (e.g., a machine learning model). The training module 1620 may perform feature extraction multiple times, each time using a different feature-extraction technique. In one example, the feature sets generated using the different techniques may each be used to generate different machine learning models 1640. For example, the feature set with the highest quality features (e.g., most indicative of interest or not of interest to a particular user(s)) may be selected for use in training… The training dataset 1610A and/or the training dataset 1610B may be analyzed to determine any dependencies, associations, and/or correlations between features and the labeled predictions in the training dataset 1610A and/or the training dataset 1610B. The identified correlations may have the form of a list of features that are associated with different labeled predictions (e.g., of interest to a particular user vs. not of interest to a particular user). The term "feature", as used herein, may refer to any characteristic of an item of data that may be used to determine whether the item of data falls within one or more specific categories or within a range. By way of example, the features described herein may comprise one or more features present within each of the media assets that may be correlative (or not correlative as the case may be) with a particular media asset being of interest to a particular user or not.” Furthermore, as cited in ¶ [0107] “To that end, the content selection unit 1210 can direct an ingestion unit 1220 to obtain a group of directed content assets from directed content storage 1280 retaining a corpus of directed content assets 1284. In some cases, the corpus of directed content assets 1264 can be categorized according to attributes of an end-user. The attributes can include, for example, market type, market segment, geography, business size, business type, revenue, profits, and similar. Accordingly, for a particular user device for which the personalization is being implemented, the content selection unit 1210 can direct the ingestions unit 1220 to obtain directed content assets having a particular set of attributes.”).
Therefore, it would have been obvious to one of ordinary skill in the art, at the time the invention was made, to add Sahasi’s determining an input feature of each training candidate advertisement according to the advertisement state corresponding to the training candidate advertisement, the overall state of the virtual advertising platform, and reference information, the reference information being feedback information provided by the judgment model for a previous round of scoring operation on the initial scoring network, and the scoring operation being performed on the training candidate advertisement corresponding to the training exposure request; determining, by the initial classification network in the initial scoring model, probabilities of the training candidate advertisement belonging to different reference advertisement types; performing, based on the probabilities of the training candidate advertisement belonging to different reference advertisement types, weighted processing on the input feature of the training candidate advertisement to obtain an input feature of the training candidate advertisement under each reference advertisement type; and determining, by the initial scoring network in the initial scoring model, a training competition score of the training candidate advertisement for the training exposure request according to the input features of the training candidate advertisement under different reference advertisement types to Yan’s determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request.  One of ordinary skill in the art would have been motivated to do so in order to allow “for improved model selection and content recommendations” (Sahasi: ¶ [0002]).

With respect to Claim 11:
Yan does not explicitly disclose the method according to claim 7, wherein determining the training target advertisement comprises: obtaining an advertisement competition score corresponding to each training candidate advertisement, the advertisement competition score being determined according to an advertisement feature of the training candidate advertisement; and determining the training target advertisement according to a training competition score of each training candidate advertisement for the training exposure request and the advertisement competition score.
However, Sahasi further discloses:
obtaining an advertisement competition score corresponding to each training candidate advertisement, the advertisement competition score being determined according to an advertisement feature of the training candidate advertisement (i.e. utilizing scoring model for each training advertisement according to content features of the advertisement) (Sahasi: ¶ [0150] “Any of the machine learning models or scoring models described herein, such as the scoring models 248 or he client model 1550, may be trained and/or retrained using training datasets comprising user activity data and/or UICs. The training datasets may comprise UICs associated with users who interacted with (e.g., engaged with) a plurality of media assets. The UICs that are used during training and/or retraining may comprise interest attributes, interest levels, functionality features, a content features, a combination thereof, and/or the like. A training module, such as the training module 1620 shown in FIG. 16, may then determine which features in the UICs correlate with the particular features of the plurality of media assets. The machine learning models, once trained ( or retrained as the case may be), may provide a recommendation for a user(s) and a media asset(s) based on the corresponding UIC(s) and the features of that media asset.”); and 
determining the training target advertisement according to a training competition score of each training candidate advertisement for the training exposure request and the advertisement competition score (i.e. determining/selecting candidate advertisement according to training advertisement via scoring model) (Sahasi: ¶ [0164] “At step 1710, the training method 1700 may determine (e.g., access, receive, retrieve, etc.) first media assets and second media assets. The first media assets and the second media assets may each comprise one or more features and a predetermined prediction (e.g., a recommendation). The training method 1700 may generate, at step 1720, a training dataset and a testing dataset. The training dataset and the testing dataset may be generated by randomly assigning media assets from the first media assets and/or the second media assets to either the training dataset or the testing dataset.” Furthermore, as cited in ¶¶ [0140] [0141] “Each client model 1550 may comprise at least one of: the feature extraction unit 210, the activity monitoring unit 220, the scoring unit 230, the scoring model(s) 248, or the profile generation unit 250 as described herein. Each client model 1550 may be customized based on a number of media assets ( e.g., digital content) the particular client has uploaded/produced for distribution by the present distribution platform…The system shown in FIG. 15 may relate to an "online" or "live" service. For example, the "online" or "live" service may provide a recommendation for content (e.g., a media asset(s)) in real-time as users associated with the particular client are engaging with the client application 106.”).
Therefore, it would have been obvious to one of ordinary skill in the art, at the time the invention was made, to add Sahasi’s obtaining an advertisement competition score corresponding to each training candidate advertisement, the advertisement competition score being determined according to an advertisement feature of the training candidate advertisement; and determining the training target advertisement according to a training competition score of each training candidate advertisement for the training exposure request and the advertisement competition score to Yan’s determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request.  One of ordinary skill in the art would have been motivated to do so in order to allow “for improved model selection and content recommendations” (Sahasi: ¶ [0002]).


Claim(s) 10 is rejected under 35 U.S.C. 103 as being unpatentable over Yan and Sahasi in view of U.S. Publication 2016/0037197 to Kitts.

With respect to Claim 10:
Yan teaches: 
The method according to claim 7, wherein simulating the virtual advertising platform comprises: obtaining historical exposure request data, historical exposure log data, […] and playing control parameters of historical placed advertisements of the advertising platform (i.e. how many times the content was accessed, log of interactions with respect to historical advertisements,) (Yan: ¶¶ [0024] [0025] “As an example, content provided by a third party system 130 to users of the online system may be identified along with information identifying an online system user 140 by information received by the online system 140, and the action logger215 logs information identifying the content provided by the third party system 130 in the action log 220 in association with the identified user of the online system 140. For example, the action logger 215 logs information describing a number of times a user of the online system 140 accessed a web page provided by a third party system 130 as well as times the user accessed the content based on information communicated to the online system 140 by a tracking pixel, or other tracking mechanism, included in the content…The action log 220 may be used by the online system 140 to track user actions on the online system 140, as well as actions on third party systems 130 that communicate information to the online system 140. Users may interact with various objects on the online system 140, and information describing these interactions is stored in the action log 220. Examples of interactions with objects include: commenting on posts, sharing links, checking-in to physical locations via a client device 110, accessing content items (including advertisements), and any other suitable interactions. Additional examples of interactions with objects on the online system 140 that are included in the action log 220 include: commenting on a photo album, communicating with a user, establishing a connection with an object, joining an event, joining a group, creating an event, authorizing an application, using an application, expressing a preference for an object ("liking" the object), and engaging in a transaction. Additionally, the action log 220 may record a user's interactions with advertisements on the online system 140 as well as with other applications operating on the online system 140.”).
Yan does not explicitly disclose wherein simulating the virtual advertising platform comprises: obtaining historical exposure request data, historical exposure log data, historical inventory data, and playing control parameters of historical placed advertisements of the advertising platform; constructing the training exposure request based on the historical exposure request data and the historical exposure log data, and determining a training candidate advertisement corresponding to the training exposure request; and determining an advertisement state corresponding to the training candidate advertisement based on the historical inventory data and the playing control parameters of the historical placed advertisements; and determining an overall state of the virtual advertising platform based on the historical inventory data, the historical exposure log data, and the playing control parameters of the historical placed advertisements.
However, Sahasi further discloses constructing the training exposure request based on the historical exposure request data and the historical exposure log data, and determining a training candidate advertisement corresponding to the training exposure request (i.e. constructing the request via the training model based on historical interaction data including exposure request and exposure log data in order to determine a recommended advertisement) (Sahasi: ¶¶ [0053] [0054] “Both the intelligence and knowledge can be generated using historical data identifying one or different types of activities of the user device. The activities can be related to consumption of digital content. In some configurations, the client application 106 can send activity data during consumption of digital content. The activity data can identify an interaction or a combination of interactions of the user device with the digital content. An example of an interaction is trick play (e.g., fast-forward or rewind) of the digital content. Another example of an interaction is reiterated playback of the digital content. Another example of an interaction is aborted playback, e.g., playback that is terminated before the endpoint of the digital content. Yet another example of the interaction is submission ( or "share") of the digital content to a user account in a social media platform. Thus, the activity data can characterize engagement with the digital content…The analytics subsystem 142 can then utilize the activity data to assess a degree of interest of the user device on the digital content (e.g., media assets). To that end, in some embodiments, the analytics subsystem 142 can train a machine learning model to discern a degree of interest on digital content among multiple interest levels. The machine learning model can be trained using unsupervised training, for example, and multiple features determined using digital content and the activity data. By applying the trained machine learning model to new activity data, an interest attribute can be generated.” Furthermore, as cited in ¶ [0142] “The client model 1550 may have been trained using historical user activity data, as further described herein. The client model 1550 may determine the at least one content recommendation 1560 based on activity/interest data 1540 associated with other users.”).
Therefore, it would have been obvious to one of ordinary skill in the art, at the time the invention was made, to add Sahasi’s constructing the training exposure request based on the historical exposure request data and the historical exposure log data, and determining a training candidate advertisement corresponding to the training exposure request to Yan’s determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request.  One of ordinary skill in the art would have been motivated to do so in order to allow “for improved model selection and content recommendations” (Sahasi: ¶ [0002]).
Yan and Sahasi do not explicitly disclose wherein simulating the virtual advertising platform comprises: obtaining historical exposure request data, historical exposure log data, historical inventory data, and playing control parameters of historical placed advertisements of the advertising platform; and determining an advertisement state corresponding to the training candidate advertisement based on the historical inventory data and the playing control parameters of the historical placed advertisements; and determining an overall state of the virtual advertising platform based on the historical inventory data, the historical exposure log data, and the playing control parameters of the historical placed advertisements.
However, Kitts further discloses:
obtaining historical exposure request data, historical exposure log data, historical inventory data, and playing control parameters of historical placed advertisements of the advertising platform (i.e. obtaining historical data representing the ads being request or placed, inventory, and parameters representing how the advertisements are played/controlled) (Kitts: ¶ [0078] “Exemplary embodiments of the present disclosure may include accessing or calculating the current ads that are running with each media asset M; in one or more ways. For example, exemplary embodiments of the present disclosure may include calculating the most frequent ad based on historical placements: as the most frequent historical ad inserted into this media M;.” Furthermore, as cited in ¶ [0107] “Equation 13 may represent the HistCPM30(A1,m,) that the advertiser has logged for patterns of media m, that match the inventory being priced M,, scaled by the typical percentage off historical price that this advertiser historically achieves HistDiscount(A).” Furthermore, as cited in ¶¶ [0110] [0111] “Equation 16 may represent the historical actual clearing price for a pattern of media that matches the inventory being priced…Exemplary embodiments of the present disclosure may train the above model on historical observations of inventory M, advertiser A, SQAD price SQADCPM30, and actual clearing price CPM30.” Furthermore, as cited in ¶ [0078] “Exemplary embodiments of the present disclosure may include accessing or calculating the current ads that are running with each media asset M; in one or more ways. For example, exemplary embodiments of the present disclosure may include calculating the most frequent ad based on historical placements:… as the most frequent historical ad inserted into this media M;. This can be accomplished by counting the most frequent ad counting the number of occurrences of each ad A1 in past airings of media M;….Alternatively, exemplary embodiments of the present disclosure may include setting CurrenAd(M;) to equal the ad which is currently planned to run in media placement M; based on known advertiser upfront and scatter purchases. This ad can be determined by linking the system to sales or inventory tracking systems which have information on which advertisers have bought particular placements.”); 
determining an advertisement state corresponding to the training candidate advertisement based on the historical inventory data and the playing control parameters of the historical placed advertisements (i.e. determining advertisement pricing based on historical inventory data and historical interaction/engagement parameters of historical ads) (Kitts: ¶ [0102] “In order to provide such dynamic pricing, exemplary embodiments of the present disclosure may include providing a yield maximization model that may predict the expected clearing price CPM30(A1,M,) based on advertiser historical prices paid and relevance to the advertiser. An algorithm for providing a yield maximization model is illustrated in FIG. 7.” Furthermore, as cited in ¶ [0107] “Equation 13 may represent the HistCPM30(A1,m,) that the advertiser has logged for patterns of media m, that match the inventory being priced M,, scaled by the typical percentage off historical price that this advertiser historically achieves HistDiscount(A).” Furthermore, as cited in ¶ [0078] “Exemplary embodiments of the present disclosure may include accessing or calculating the current ads that are running with each media asset M; in one or more ways. For example, exemplary embodiments of the present disclosure may include calculating the most frequent ad based on historical placements:… as the most frequent historical ad inserted into this media M;. This can be accomplished by counting the most frequent ad counting the number of occurrences of each ad A1 in past airings of media M;….Alternatively, exemplary embodiments of the present disclosure may include setting CurrenAd(M;) to equal the ad which is currently planned to run in media placement M; based on known advertiser upfront and scatter purchases. This ad can be determined by linking the system to sales or inventory tracking systems which have information on which advertisers have bought particular placements.”); and 
determining an overall state of the virtual advertising platform based on the historical inventory data, the historical exposure log data, and the playing control parameters of the historical placed advertisements (i.e. determining an overall state or tratio of all the advertisers based on historical inventory data, exposure data, and parameters with respect to engagement) (Kitts: ¶¶ [0113] [0114] “Exemplary embodiments of the present disclosure may further include providing a graphical user interface (GUI) to enable a network to view a list of advertisers that may be inserted against their inventory, as shown in FIG. 14…The GUI may be organized into a grid which has TV inventory (1430) running down the page, and candidate advertisers who could be inserted against media inventory running across the page (1450). A series of linked filters may be available on the left and right-hand panes (1420). The GUI may support two-dimensional sorting. Vertical sorting may enable the sorting by, for example, schedule, cost of media, gain in relevance, relevance, or units available, etc. (1410), so that a network may quickly review which inventory to address. Horizontal sorting may enable sorting by advertisers who could be inserted into each position (1450) by, for example, tratio relevance, or other metrics. The GUI may further provide the current or most request ad for each media (1440).” Furthermore, as cited in ¶ [0126] “In FIG. 15, tratio is the match between advertiser's population and the audience of the program (1510). Music players and services are shown as the top advertisers on MTV due to the demographics, while colleges and online education would also be interested in purchasing TV spots (1540). This may be attributed to a predicted 9.09 clearance price for the music companies and a predicted 8.82 clearance price for the technical colleges (1530). In addition, the report shows how much each agency has historically spent (1520). This can be useful for looking for agencies that are likely to buy in the future.”).
Therefore, it would have been obvious to one of ordinary skill in the art, at the time the invention was made, to add Kitts’ obtaining historical exposure request data, historical exposure log data, historical inventory data, and playing control parameters of historical placed advertisements of the advertising platform; constructing the training exposure request based on the historical exposure request data and the historical exposure log data, and determining a training candidate advertisement corresponding to the training exposure request; and determining an advertisement state corresponding to the training candidate advertisement based on the historical inventory data and the playing control parameters of the historical placed advertisements; and determining an overall state of the virtual advertising platform based on the historical inventory data, the historical exposure log data, and the playing control parameters of the historical placed advertisements to Yan’s determining a target advertisement exposed by the current exposure request according to the competition score of each candidate advertisement for the current exposure request.  One of ordinary skill in the art would have been motivated to do so in order to “allow advertisers to understand users' online behaviors through the indirect use of raw data and may maintain privacy of the users and the data.” (Kitts: ¶ [0012]).

Response to Arguments
Applicant’s arguments see pages 17-24 of the Remarks disclosed, filed on 04/30/2025, with respect to the 35 U.S.C. § 101 rejection(s) of claim(s) 1-20 have been considered but are not persuasive:
The Applicant asserts “Specifically, amended claim 1 recites, in addition to the alleged judicial exception: transmitting the target advertisement to a terminal device for playback. The features are not recited at a high level of generality. On the contrary, the recited features impose meaningful limit in terms of how the determined target advertisement is to be utilized. These additional limitations are not merely extra-solution activities… Further, amended claim 1 clearly reflects an improvement to the technology or technical field of online advertisement delivery. In particular, an improved method and process for selecting a target advertisement from multiple candidate advertisements according to a current exposure request and delivering the selected target advertisement for playback is realized. As explained in ¶ [0032] of the specification: “[T]here are generally tens of thousands of advertisements with the targeting conditions meeting the exposure request. Scores are configured for tens of thousands of advertisements, and a finally exposed advertisement is selected therefrom, whereby the scoring model to be trained has a huge action space, and the huge action space will often make the scoring model difficult to converge, thereby causing poor performance of the scoring model obtained by final training and difficulty in configuring scores for various advertisements.” Therefore, in particular, claim 1 represents an improvement to the technology of using machine learning model-based scoring to select target advertisements. Accordingly, amended claim 1 integrates the alleged judicial exception into a practical application, i.e., a machine learning model-based target advertisement selection and delivery method and process.” The Examiner respectfully disagrees. Claims 1, 12, and 20 have recited the following additional elements: by a classification/scoring network in a scoring model; and “the scoring model comprising multiple scoring networks corresponding to different reference advertisement types, each of the multiple scoring networks being configured to score each candidate advertisement based on a corresponding reference advertisement type of the different reference advertisement types”; and terminal device. The additional elements reciting – “by a classification/scoring network in a scoring model” and “terminal device” in claims 1, 12, and 20 are not found to integrate the judicial exception into a practical application. Merely adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea and Generally linking the use of the judicial exception to a particular technological environment or field of use – see MPEP 2106.05(f) is not indicative of integration into a practical application. Accordingly, alone, and in combination, these additional elements are seen as using a computer or tool to perform an abstract idea, adding insignificant-extra-solution activity to the judicial exception. They do no more than link the judicial exception to a particular technological environment or field of use, i.e. networks in a model and device, and therefore do not integrate the abstract idea into a practical application. The courts decided that although the additional elements did limit the use of the abstract idea, the court explained that this type of limitation merely confines the use of the abstract idea to a particular technological environment and this fails to add an inventive concept to the claims (See Affinity Labs of Texas v. DirecTV, LLC,).
The Applicant finally asserts “Specifically, the additional elements recited in amended claim 1, including the claim features relating to utilizing a scoring model comprising multiple scoring networks corresponding to different reference advertisement types in the selection of a target advertisement, are not well-understood, routine, or conventional in the field. Further, the additional elements, in combination, reflect an improvement to the technology or technical field of machine learning model-based scoring for selection of target advertisements, as explained above. Therefore, the additional elements recited in amended claim 1 amount to significantly more than the alleged judicial exception.” The Examiner respectfully disagrees. Claims 1, 12, and 20 recite the following additional limitations including by a classification/scoring network in a scoring model; and “the scoring model comprising multiple scoring networks corresponding to different reference advertisement types, each of the multiple scoring networks being configured to score each candidate advertisement based on a corresponding reference advertisement type of the different reference advertisement types”; and terminal device. The additional elements reciting – “by a classification/scoring network in a scoring model” and “terminal device” do not integrate the judicial exception (abstract idea) into a practical application because of the analysis provided in Step 2A, Prong II. Claims 1, 12, and 20 also recite additional elements – “the scoring model comprising multiple scoring networks corresponding to different reference advertisement types, each of the multiple scoring networks being configured to score each candidate advertisement based on a corresponding reference advertisement type of the different reference advertisement types”. Merely describing a model that has different scoring networks for different advertisement types are not indicative of integration into a practical application because the limitation is adding insignificant extra-solution activity to the judicial exception - see MPEP 2106.05(g). The courts have noted that “performing repetitive calculations” is seen as a well-understood, routine, and conventional computer function (See: Flook, 437 U.S. at 594, 198 USPQ2d at 199 (recomputing or readjusting alarm limit values); Bancorp Services v. Sun Life, 687 F.3d 1266, 1278, 103 USPQ2d 1425, 1433 (Fed. Cir. 2012) ("The computer required by some of Bancorp’s claims is employed only for its most basic function, the performance of repetitive calculations, and as such does not impose meaningful limits on the scope of those claims.")). Claims 1, 12, and 20 do not include additional elements or a combination of elements that result in the claims amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements listed amount to no more than mere instructions to apply an exception using a generic computer component. In addition, the applicant’s specifications describe generic computer-based elements, ¶¶ [0011] [0025], for implementing the “apparatus” or  “processor”, which do not amount to significantly more than the abstract idea of itself, which is not enough to transform an abstract idea into eligible subject matter. Furthermore, there is no improvement in the functioning of the computer or technological field, and there is no transformation of subject matter into a different state. Therefore, the rejection(s) of claim(s) 1-20 under 35 U.S.C. § 101 is maintained above with an updated analysis.
Applicant’s arguments see pages 24-27 of the Remarks disclosed, filed on 04/30/2025, with respect to the 35 U.S.C. § 103 rejection(s) of claim(s) 1-9 and 11-20 over Yan in view of Sahasi with claim 10 being rejected in further view of Kitts have been considered but are not persuasive. The Applicant asserts “Yan is directed to an online system that tracks and stores information identifying content provided by third party systems and accessed by online system users as well as interactions with advertisement performed by online system users. Yan, Abstract. Without acquiescing to the Office’ assertions, Applicant respectfully notes that Yan discloses that an advertisement request includes advertisement content and a bid amount. Id. at paragraph [0031]. Further, Yan discloses various example targeting criteria. Id. at paragraph [0033]. Moreover, Yan discloses that the correlation module 235 determines a correlation between content provided by third party systems accessed by a viewing user and content provided by third party systems accessed by additional users. Id. at paragraph [0034]… Applicant respectfully submits that to the extent Yan discloses advertisement (content), the bidding amount, and targeting criteria, Yan does not disclose or suggest “the plurality of candidate advertisements comprises a contract advertisement and a bid advertisement, the contract advertisement is associated with a contract specifying at least a predetermined playing amount, a selling price, and a targeting condition, the bid contract is associated with an advertisement effect and a pre-offered bid, the contract advertisement and the bid advertisement are mixed in the plurality of candidate advertisements for the determining of the competition score,” as recited in amended claim 1, at least because Yan fails to disclose or suggest two different types of advertisements: the contract advertisement and the bid advertisement, which are mixed in the plurality of candidate advertisements. In fact, Yan is completely silent on the above-quoted claim features recited in claim 1.” The Examiner respectfully disagrees. The Examiner would like to refer the Applicant to ¶¶ [0039] [0040] of the Yan reference; “An expected value associated with an ad request or with a content item represents an expected amount of compensation to the online system 140 for presenting an ad request or a content item. For example, the expected value associated with an ad request is a product of the ad request's bid amount and a likelihood of the user interacting with the ad content from the ad request. The content selection module 240 may rank ad requests based on their associated bid amounts and select ad requests having at least a threshold position in the ranking for presentation to the user. In some embodiments, the content selection module 240 ranks both content items not associated with bid amounts and ad requests in a unified ranking based on bid amounts associated with ad requests and measures of relevance associated with content items and ad requests. Based on the unified ranking, the content selection module 240 selects content for presentation to the user. Selecting ad requests and other content items through a unified ranking is further described in U.S. patent application Ser. No. 13/545, 266, filed on Jul. 10, 2012, which is hereby incorporated by reference in its entirety…For example, the content selection module 240 receives a request to present a feed of content to a user of the online system 140. The feed may include one or more advertisements from ad request as well as content items, such as stories describing actions associated with other online system users connected to the user. The content selection module 240 accesses one or more of the user profile store 205, the content store 210, the action log 220, and the edge store 225 to retrieve information about the user. For example, stories or other data associated with users connected to the identified user are retrieved. Additionally, one or more advertisement requests ("ad requests") may be retrieved from the ad request store 230 The retrieved stories, ad requests, or other content items, are analyzed by the content selection module 240 to identify candidate content that is likely to be relevant to the identified user. For example, stories associated with users not connected to the identified user or stories associated with users for which the identified user has less than a threshold affinity are discarded as candidate content. Based on various criteria, the content selection module 240 selects one or more of the content items or ad requests identified as candidate content for presentation to the identified user. The selected content items or ad requests are included in a feed of content that is presented to the user. For example, the feed of content includes at least a threshold number of content items describing actions associated with users connected to the user via the online system 140.” Furthermore, as cited in ¶¶ [0031] [0032] “The bid amount is associated with an advertisement by an advertiser and is used to determine an expected value, such as monetary compensation, provided by an advertiser to the online system 140 if advertisement content in the ad request is presented to a user, if the advertisement content in the ad request receives a user interaction when presented, or if any suitable condition is satisfied when advertisement content in the ad request is presented to a user. For example, the bid amount specifies a monetary amount that the online system 140 receives from the advertiser if advertisement content in an ad request is displayed…Targeting criteria included in an advertisement request specify one or more characteristics of users eligible to be presented with advertisement content in the advertisement request. For example, targeting criteria are used to identify users having user profile information, edges, or actions satisfying at least one of the targeting criteria. Hence, targeting criteria allow an advertiser to identify users having specific characteristics, simplifying subsequent distribution of content to different users.”) It is clear from the disclosure above that the Yan reference teaches that candidate advertisements include ad requests with bid amounts or contract ads and ad requests associated with content stories or bid advertisements, wherein ad requests associated with bid requests include expected value, targeting criteria and bid amount, the ad requests associated with content stories include an advertisement effect or action and associated bid amount, wherein both ad requests associated with bid amounts and ad requests associated with content stories are included in a unified ranking to select candidate advertisements. Therefore, the rejection(s) of claim(s) 1-20 under 35 U.S.C. § 103 is provided above with updated citations.


Conclusion
The prior art made of record and not relied upon is considered pertinent to Applicant’s disclosure. The following reference are cited to further show the state of the art:
U.S. Publication 2021/0110428 to Lu for disclosing In some examples, a computing device includes at least one processor and at least one module, operable by the at least one processor to receive, from a client device of a user, a request for one or more advertisements to display at the client device with a set of messages. The set of messages is associated with the user in a social network messaging service. The at least one module may be further operable to determine a probabilities that the user will select a candidate advertisement using a machine learning model based on point-wise learning and pair-wise learning. The at least one module may be further operable to determine, based on the probabilities that the user will select the candidate advertisement, a candidate score for the candidate advertisement, determine that the candidate score satisfies a threshold, and send, for display at the client device, the candidate advertisement.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Azam Ansari, whose telephone number is (571) 272-7047. The examiner can normally be reached from Monday to Friday between 8 AM and 4:30 PM.
If any attempt to reach the examiner by telephone is unsuccessful, the examiner's supervisor, Waseem Ashraf, can be reached at (571) 270-3948. 
Another resource that is available to applicants is the Patent Application Information Retrieval (PAIR). Information regarding the status of an application can be obtained from the (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAX. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pairdirect.uspto.gov. Should you have questions on access to the Private PAIR system, please feel free to contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free).
Applicants are invited to contact the Office to schedule either an in-person or a telephonic interview to discuss and resolve the issues set forth in this Office Action. Although an interview is not required, the Office believes that an interview can be of use to resolve any issues related to a patent application in an efficient and prompt manner.

/AZAM A ANSARI/
Primary Examiner, Art Unit 3621                                                                                                                                                                                                        	
May 7, 2025