DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application is being examined under the pre-AIA first to invent provisions.
The amendment filed on 5/8/2026 has been acknowledged but not being entered due to the non-compliant issue as stated in the Notice of Non-Compliant Amendment mailed on 5/21/2026.
The action is responsive to amendment filed on 6/1/2026.
Claims 1-24 are cancelled.
Claims 25-44 are pending in this application.
Response to Arguments
Applicant’s arguments with respect to the rejections previously made and the amended claims filed on 5/8/2026 and 6/1/2026 have been fully considered. In view of the claim amendments, the rejections are being updated accordingly.
Double Patenting
In view of the approved terminal disclaimers, the rejections as set forth in the previous office action are hereby withdrawn.
35 USC 101 Rejections
Applicant’s arguments have been fully considered.
In response to the arguments, it is submitted that the intended inventive concept may be directed to "a specific machine-learning implementation using gradient boosted decision trees with feature space partitioning that improves the technical process of ranking matches in a dating service” as stated by the Applicant.
However, such concept is not being recited in claim 25 as nowhere in claim 25 recite any usage of gradient boosted decision trees with feature space partitioning.
The newly added limitations of “the fourth entity being an entity for which behavioral features are absent, wherein the ranking model comprises a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages” merely recite (i) the forth is an entity that does not have behavioral futures and (ii) the ranking model includes a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages.
Applying the ranking model to rank a potential match and having the boosted regression tree that is an aggregate of regression trees computed in a sequence of stages being included in the ranking model as claimed do not necessary require any usage of gradient boosted decision trees with feature space partitioning as intended because the ranking model application could involve a subset of elements other than the boosted regression tree.
Also, the newly added limitations further emphasize the claimed process is directed to mathematic relationships with additional recitations of boosted regression tree that is an aggregate of regression trees computed in a sequence of stages, which involves mathematical calculations with mathematical algorithms and formulas known by one skilled in the art.
Plus, the amended claim recite a process comprising detecting behavior features indicate degrees of interest of matches between entities, determining a probability of relevance, dividing features values into regions with value partitioning, assigning each region of the partitioned regions with a value to define a function with mathematical symbols (T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri), training a ranking model using features defined by profile data and probability of relevant as well as minimizing a total loss, and ranking a potential match by applying the ranking model with additional mathematical calculations of boosted regression tree. The process additionally comprising the applying is based on vectors and that the ranking model minimizes a total loss based on a gradient descent method as recited in the dependent claims
Thus, the claimed process is similar to a method of mathematic relationships directed to a series of mathematic relationships involving degrees of interest, probability of relevance, feature value division, usage of mathematical symbols to represent a partitioned region, total loss calculation and ranking data element.
E.g. the degree of interest is direct to mathematical relationship that indicates a quantity of interest by an entity.
Similarly, the probability of probability of relevance, feature value division, usage of mathematical symbols to represent a partitioned region, total loss calculation and ranking are all directed to using mathematics involving numbers and numeric operations.
Moreover, the claims explicitly recite mathematic relationships based on “value <Di to define a function T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri”, which explicitly supported by supported by para [0052] of the specification that the defined function by partitioning clearly involves a mathematical formula, which further emphasize that the claimed process is directed to a series of mathematic relationships.
Likewise, aggregate of regression trees computed as claimed by the added limitation is directed to addition of calculated trees via computing, which involves mathematic relationships.
Besides, the claimed steps with mathematic relationships do not render the claims to a practical application because the claimed process does not show how performing a series of mathematic relationships would direct the claimed process to a particular useful application.
In addition, the element of “to rank a potential match” is directed to non-functional descriptive material that describes that intended outcome of in the field usage in a dating service when the machine-learning model is applied, which not impose a meaningful limit on the judicial exception since the ranking step is not functionally involved to any practical application or outcome. Nowhere in the claim show how the claimed elements would improve the dating service or rendering the dating service practical application, and merely reciting the intended purpose as claimed would not integrate the abstraction idea into practical application.
Additionally, the claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. Additional elements (e.g. behavioral features, profiles, click feedback) are directed to types of information, which do not impose a meaningful limit on the judicial exception, such that the claims are more than a drafting effort design to monopolize exception, because the claimed steps could be performed in a same manner to achieve the same outcome with other types of information other than the ones being used in the claims.
Hence, the claims do not include additional elements or the combination of the elements are sufficient to amount to significantly more than the judicial exception and fail to integrate the judicial exception into practical application according to Prong Two in Step 2A of the 2019 Patent Subject Matter Eligibility Guidance because the claimed elements or their combination do not impose any meaningful limits on practicing the abstract idea.
Further, in view of Step 2B of the 2019 Patent Subject Matter Eligibility Guidance, it is determined that the computing elements (such a computer readable storage medium, a memory, processor) in the claim amount to no more than usage of a generic computing system having a generic computing components, which fails to provide an inventive concept or significantly more than abstract idea because the elements do not necessary improve the functional of a computing system or an improvement to a technical field since network computing is well known.
Therefore, for at least the reasoning above, the pending claims are not patent eligible.
35 USC 101 Rejections
Applicant’s arguments have been fully considered.
In response to the arguments, it is submitted that that as stated above, “dating services or match-making” is directed to the intended of the claimed process, as further emphasized by “a method for a dating device”, and intended use element does not carry any patentable weight.
Plus, cited reference Hueter explicitly discloses a method for dating service application with candidate profiles and potential match determination ([0028]).
Moreover, regarding the newly added limitations of “the fourth entity being an entity for which behavioral features are absent, wherein the ranking model comprises a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages”, it is submitted that the limitation are merely directed to non-functional descriptive materials for the following reasons:
(i) the forth is an entity that does not have behavioral futures--which does not impact any of the functional nor outcome of any claimed steps--as all the claimed steps would performed the same regardless of whether the forth entity is an entity that having behavioral futures or not, and
(ii) the ranking model includes a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages-- which also does not impact any of the functional nor outcome of any claimed steps--as all the claimed steps would performed the same regardless of whether a boosted regression tree is included in the ranking model or not, and regardless of whether the boosted regression tree that is an aggregate of regression trees computed in a sequence of stages or not.
Likewise, as previously stated, regarding the steps of “partitioning a space of feature values into regions…assigning each of the regions a value...”, these steps are merely directed to splitting a space of values into partitions (i.e. regions) and assign a data value to each partition (i.e. region).
Nowhere in these limitations define any specific mathematical structure or operation of the regression tree used in the ranking model as presented in the claim. And the partitioning of feature space into regions and the assignment of values are not or required to involve in determine how the function T(u,v) computes relevance scores for entity pairs as presented in the claim.
In addition, according to the claim language, the machine ranking model is trained (i) using a subset of features defined by profiles and probability of relevance, and (ii) by
minimizing a total loss, at least in part based on the first entity and the second entity.
Thus, the training is not depending partitioning step nor the assigning step. And the assigned value is not being used in the training or nor the applying ranking model as set forth by the claim language, which rendering the partitioning and assigning steps being interpreted independently from the training and apply steps.
Likewise, nowhere in the claim cited any limitation on changing the partitioning or the assigned values would change the output of the function and thus the ranking results, and hence such concept is not given any patentable weight.
Further, it is submitted that the newly added limitations are properly addressed by the
new ground of rejection; see rejection below for detail.
Furthermore, it is submitted that all limitations in pending claims--including those not specifically argued--are properly addressed. The reason is set forth in the rejections. See claim analysis below for detail.
Claim Objections
Claims 25, 32 and 38 are objected to because of the following informalities: fail to mark the amended elements in the claims to indicate the changes made in accordance to MPEP 714.
It is noted that the claim amendments filed on 5/8/2026 are not being entered due to non-compliant issue. Hence the amendment filed on 6/1/2026 need to be based on the claim amendment filed on 1//2026.
Appropriate correction is required.
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 25-44 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
The amended claims 25-44 recite a process comprising detecting behavior features indicate degrees of interest of matches between entities, determining a probability of relevance, dividing features values into regions with value partitioning, assigning each region of the partitioned regions with a value to define a function with mathematical symbols (T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri), training a ranking model using features defined by profile data and probability of relevant as well as minimizing a total loss, and ranking a potential match by applying the ranking model. The process also comprising the applying is based on vectors and that the ranking model minimizes a total loss based on a gradient descent method as recited in the dependent claims.
The claimed process is similar to a method of mathematic relationships, which is one of the groupings of abstract ideas according to Prong One in Step 2A of the 2019 Patent Subject Matter Eligibility Guidance since the claimed steps are directed a series of mathematic relationships involving degrees of interest, probability of relevance, feature value division, usage of mathematical symbols to represent a partitioned region, total loss calculation and ranking data element.
E.g. the degree of interest is direct to mathematical relationship that indicates a quantity of interest by an entity. Similarly, the probability of probability of relevance, feature value division, usage of mathematical symbols to represent a partitioned region, total loss calculation and ranking are all directed to using mathematics involving numbers and numeric operations.
Plus, the claims explicitly recite mathematic relationships based on “value <Di to define a function T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri”, which explicitly supported by supported by para [0052] of the specification that the defined function by partitioning clearly involves a mathematical formula, which further emphasize that the claimed process is directed to a series of mathematic relationships.
Besides, the newly added limitation “wherein the ranking model comprises a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages” further emphasize that the claimed process is directed to mathematic relationships with additional recitations of boosted regression tree that is an aggregate of regression trees computed in a sequence of stages, which involves mathematical calculations.
In addition, the claimed steps with mathematic relationships do not render the claims to a practical application because the claimed process does not show how performing a series of mathematic relationships would direct the claimed process to a particular useful application.
Additionally, the element of “to rank a potential match” is directed to non-functional descriptive material that describes that intended outcome when the machine-learning model is applied, which not impose a meaningful limit on the judicial exception since the ranking step is not functionally involved to any practical application or outcome. Merely reciting the intended purpose as claimed would not integrate the abstraction idea into practical application.
Also, the claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. Additional elements (e.g. behavioral features, profiles, click feedback) are directed to types of information, and the newly amended limitation of “the fourth entity being an entity for which behavioral features are absent” is directed to a type of descriptive data information. None of the data information impose a meaningful limit on the judicial exception, such that the claims are more than a drafting effort design to monopolize exception, because the claimed steps could be performed in a same manner to achieve the same outcome with other types of information other than the ones being used in the claims.
Hence, the claims do not include additional elements or the combination of the elements are sufficient to amount to significantly more than the judicial exception and fail to integrate the judicial exception into practical application according to Prong Two in Step 2A of the 2019 Patent Subject Matter Eligibility Guidance because the claimed elements or their combination do not impose any meaningful limits on practicing the abstract idea.
Further, in view of Step 2B of the 2019 Patent Subject Matter Eligibility Guidance, it is determined that the computing elements (such a computer readable storage medium, a memory, processor) in the claim amount to no more than usage of a generic computing system having a generic computing components, which fails to provide an inventive concept or significantly more than abstract idea because the elements do not necessary improve the functional of a computing system or an improvement to a technical field since network computing is well known.
Thus, for at least the reasoning above, the pending claims are not patent eligible.
Examiner Comments
The term “for” in the claims (e.g. claims 25, 32, 38) indicates intended use; Minton v. Nat ’l Ass ’n of Securities Dealers, Inc., 336 F.3d 1373, 1381, 67 USPQ2d 1614, 1620 (Fed. Cir. 2003) “whereby clause in a method claim is not given weight when it simply expresses the intended result of a process step positively recited.” Examples of claim language, although not exhaustive, that may raise a question as to the limiting effect of the language in a claim are: (A) “adapted to” or “adapted for” clauses; (B) “wherein” clauses; and (C) “whereby” clauses. Therefore intended use limitations are not required to be taught, see MPEP 2103, 2106 Section II(C), MPEP 2111.04 [R-3]).
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of pre-AIA 35 U.S.C. 103(a) which forms the basis for all obviousness rejections set forth in this Office action:
(a) A patent may not be obtained though the invention is not identically disclosed or described as set forth in section 102, if the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art to which said subject matter pertains. Patentability shall not be negated by the manner in which the invention was made.
This application currently names joint inventors. In considering patentability of the claims under pre-AIA 35 U.S.C. 103(a), the examiner presumes that the subject matter of the various claims was commonly owned at the time any inventions covered therein were made absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and invention dates of each claim that was not commonly owned at the time a later invention was made in order for the examiner to consider the applicability of pre-AIA 35 U.S.C. 103(c) and potential pre-AIA 35 U.S.C. 102(e), (f) or (g) prior art under pre-AIA 35 U.S.C. 103(a).
Claims 25-44 are rejected under pre-AIA 35 U.S.C. 103(a) as being unpatentable over Hueter et al (Pub No US 2009/0248682, hereinafter Hueter) in view of Liu et al (Pub No. US 2009/0106222, hereinafter Liu), and futher in view of Sarlos (Pub No. US 2009/0204602)
Heuter and Lin are cited in the previous office action.
With respect to claim 25, Hueter discloses a method for a dating service including a plurality of candidate profiles, the plurality of candidate profiles including a profile of a first entity and a profile of a second entity (abstract, [0027-0028], Fig 5: a method for a dating service application with multiple profiles), the method comprising:
detecting a first behavioral feature for a first potential match for the first entity and a second behavioral feature for a second potential match for the second entity, wherein the first behavioral feature indicates a degree of at least one-way interest in the first entity by a third entity, and the second behavioral feature indicates a degree of at least one-way interest in the second entity by the third entity (elements on indicating degree of one way interest appears to be directed to non-functional descriptive material for not functionally impacting the structure of the claim because the steps in the claim would be performed the same regardless of the indicated degree of one-way interest ; [0012-0014], [0023-0024], [0043],Fig 2-5: detect behavioral features based on a subject of an active entity’s behaviors toward other subjects representing other entities for respective potential match. The detected behavioral features including and not limited to explicit and inferred/derived behavior features of the active entity that represents the 3rd entity. The explicit behavioral features include explicit input or feedbacks of relevancy indicating a degree of at least one-way interest by the 3rd entity toward search results being presented to the 3rd entity. Inferred/derived behavior features include implicit feedback—including and not limited to non-selection of search results--which also indicate a degree of a degree of at least one-way interest by the 3rd entity toward search results being presented to the 3rd entity. Each search result represents an entity. Hence a 1st and 2nd entities are being represented by two search results among the plurality of search results being presented. Since the plurality of search results being presented and the active entity are being tracked, hence a first behavior feature for a first potential match and a 2nd behavior features for a second potential match are being detected upon explicit and inferred/derived behaviors of the active entity);
determining a probability of relevance of the first and second potential matches based at least in part upon the first and second behavioral features ([0012-0015], [0020-0024], [0031], [0043], Fig 2-5: determine probability of relevance based behavioral features when training a learning method and/or when determine relevance of one or more matches, which are based on the behavior features in view of input or lack of input of the 3rd entity. Also the scores such as the relevance/behavioral score may be corresponding to the probability of relevance as well since it is also based on the behavioral features as well);
partitioning a space of feature values into regions ([0042-0044], [0046-0047], Fig 4: the partitioning a space of presented search results representing the feature values into at least two 2 regions according to feedback, e.g. a region of selected result, and a region of unselected region, and/or a region of relevant and a region of not relevant);
assigning each of the regions a value ([0042-0044], [0046-0047], Fig 4: assigning each region a value, e.g. a value of selected or unselected, and/or a value of relevant or not relevant);
training a machine-learned ranking model (i) using a subset of features defined by the plurality of candidate profiles and the probability of relevance of each of the first and second potential matches as inputs ([0030-0031],[0042-0046], Fig 4-6: training a ranking model at least in part is based on profile features and the probability of relevance as the profiles features are being used matching processing to generate return results representing the entities including the 1st and 2nd entities as described in [0030] & [0042], and then used to update entity profiles based on feedback for the subsequent matches that includes a potential match for a fourth entity indicated by subsequent search results as described in [0046]));
applying the ranking model to rank a potential match for a fourth entity [0030-0031],[0042-0046], Fig 4-6: apply ranking model to rank potential matches-- such as match in the subsequent search for the 4th entity of the potential matches which is the intended use entity of the ranking model application and does not necessary carry any patentable weight), the fourth entity being an entity for which behavioral features are absent (the limitation is directed to non-functional descriptive material which does not impact functionality of any claimed steps; [0043], Fig 3 & 6: the forth entity represented by a given result return to the user is being entity without behavioral features with no feedback from the user yet).
Hueter does not explicitly discloses R9, j=1, 2, ... , J for the regions; a value <Di to define a function T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri for each of the regions;
training the machine-learned ranking model by minimizing a total loss, at least in part based on the first entity and the second entity, and wherein the ranking model comprises a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages as claimed.
However, Liu discloses training a machine-learned ranking model by minimizing a total loss, at least in part based on a first entity and a second entity ([0021-0023]: training a machine learning ranking model represented by a ranking function by minimizing a total loss a first and second entities with a usage of a loss function); and
partitioning a space of feature values into regions and assigning each region a value that involve usage of different symbol values such as j=1, 2, ...n(i) ([0020-0022], [0041], [0055-0061]: partition a space represented by a list of features vectors each with an assigned vales, such that the math equation is defined using different mathematical symbol values to rank potential matches)
Since (i) Hueter further discloses the training of ranking model adopts function to minimize squared error, and (ii) Hueter and Liu are from the same field of because both are directed to determine matches between entities based on associated feature indicators, it would have been obvious to one skilled in the art at the time of the invention to modify the error computation & ranking model application of Hueter and incorporate the total loss computation & ranking model application of Liu in ranking model training of teachings of Hueter for minimizing loss and to rank potential match as claimed. The motivation to combine is to provide intelligently rank matches resulting highly relevant information provision for users (Hueter, [0001]; Liu, [0003]).
Liu does not explicitly disclose R9, j=1, 2, ... , J for the regions; a value <Di to define a function T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri for each of the regions, wherein the ranking model comprises a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages as claimed.
However, this difference is only found in the nonfunctional descriptive data materials that describe the following:
data labels of R9, j=1, 2, ... , J for the partitioned regions,
a data value assigned to each of the regions,
an intended use of region value assignment that is directed to define a function T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri for each of the regions, and
the ranking includes boosted regression tree that is an aggregate of regression trees computed in a sequence of stages.
Neither of data label (or data value) or the intended use, nor the boosted regression tree that is an aggregate of regression trees computed in a sequence of stage is functionally involved in the steps recited.
All the steps in the claims (e.g. detecting, determining, partitioning, assigning, training, and apply) would be performed the same regardless of what the data labels for the regions and region values are. Similarly, all steps in the claims would be performed the same regardless of intended use of the region value assignment. Likewise, all steps in the claims would be performed the same regardless of whether the ranking includes boosted regression tree that is an aggregate of regression trees computed in a sequence of stages or not. None of them hve any of the functionalities of the claimed steps as all the steps would be performed the same to achieve the same outcome.
Therefore, it would have been obvious to one skilled in the art before the effective filing date of the claimed invention to (i) use any data label for the partitions, (ii) assign any value to each region for any usage and (iii) include any data information in the ranking model because such data material does not functionally relate to the steps in the method claimed. Also, it is because the subjective interpretation of the data does not patentably distinguish the claimed invention.
Further, Sarlos discloses wherein the ranking model comprises a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages ([0032]: ranking model with a boosted regression trees which computes in stages as boosted regression trees are configured).
Since Hueter. Liu and Sarlos are from the same field of because all are directed to query processing to provide relevant data information, it would have been obvious to one skilled in the art at the time of the invention to modify and combined their teachings by incorporate the boosted regression trees technique of Sarlos into the combined teachings of Hueter and Liu to rank potential match as claimed. The motivation to combine is to provide intelligently rank matches resulting highly relevant information provision for users with improved search results (Hueter, [0001]; Liu, [0003]; Sarlos, [0005])
With respect to claim 32, Hueter discloses a non-transitory, computer-readable medium storing thereon program instructions for performing operations for a dating service including a plurality of candidate profiles, the plurality of candidate profiles including a profile of a first entity and a profile of a second entity (Abstract, [0001], [0027-0028], Fig 5: operations of a method for a dating service application with multiple profiles), the operations comprising:
detecting a first behavioral feature for a first potential match for the first entity and a second behavioral feature for a second potential match for the second entity, wherein the first behavioral feature indicates a degree of at least one-way interest in the first entity by a third entity, and the second behavioral feature indicates a degree of at least one-way interest in the second entity by the third entity (elements on indicating degree of one way interest appears to be directed to non-functional descriptive material for not functionally impacting the structure of the claim because the steps in the claim would be performed the same regardless of the indicated degree of one-way interest ; [0012-0014], [0023-0024], [0043],Fig 2-5: detect behavioral features based on a subject of an active entity’s behaviors toward other subjects representing other entities for respective potential match. The detected behavioral features including and not limited to explicit and inferred/derived behavior features of the active entity that represents the 3rd entity. The explicit behavioral features include explicit input or feedbacks of relevancy indicating a degree of at least one-way interest by the 3rd entity toward search results being presented to the 3rd entity. Inferred/derived behavior features include implicit feedback—including and not limited to non-selection of search results--which also indicate a degree of a degree of at least one-way interest by the 3rd entity toward search results being presented to the 3rd entity. Each search result represents an entity. Hence a 1st and 2nd entities are being represented by two search results among the plurality of search results being presented. Since the plurality of search results being presented and the active entity are being tracked, hence a first behavior feature for a first potential match and a 2nd behavior features for a second potential match are being detected upon explicit and inferred/derived behaviors of the active entity);
determining a probability of relevance of the first and second potential matches based at least in part upon the first and second behavioral features ([0012-0015], [0020-0024], [0031], [0043], Fig 2-5: determine probability of relevance based behavioral features when training a learning method and/or when determine relevance of one or more matches, which are based on the behavior features in view of input or lack of input of the 3rd entity. Also the scores such as the relevance/behavioral score may be corresponding to the probability of relevance as well since it is also based on the behavioral features as well);
partitioning a space of feature values into regions ([0042-0044], [0046-0047], Fig 4: the partitioning a space of presented search results representing the feature values into at least two 2 regions according to feedback, e.g. a region of selected result, and a region of unselected region, and/or a region of relevant and a region of not relevant);
assigning each of the regions a value ([0042-0044], [0046-0047], Fig 4: assigning each region a value, e.g. a value of selected or unselected, and/or a value of relevant or not relevant);
training a machine-learned ranking model (i) using a subset of features defined by the plurality of candidate profiles and the probability of relevance of each of the first and second potential matches as inputs ([0030-0031],[0042-0046], Fig 4-6: training a ranking model at least in part is based on profile features and the probability of relevance as the profiles features are being used matching processing to generate return results representing the entities including the 1st and 2nd entities as described in [0030] & [0042], and then used to update entity profiles based on feedback for the subsequent matches that includes a potential match for a fourth entity indicated by subsequent search results as described in [0046]));
applying the ranking model to rank a potential match for a fourth entity [0030-0031],[0042-0046], Fig 4-6: apply ranking model to rank potential matches-- such as match in the subsequent search for the 4th entity of the potential matches which is the intended use entity of the ranking model application and does not necessary carry any patentable weight), the fourth entity being an entity for which behavioral features are absent (the limitation is directed to non-functional descriptive material which does not impact functionality of any claimed steps; [0043], Fig 3 & 6: the forth entity represented by a given result return to the user is being entity without behavioral features with no feedback from the user yet).
Hueter does not explicitly discloses R9, j=1, 2, ... , J for the regions; a value <Di to define a function T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri for each of the regions;
training the machine-learned ranking model by minimizing a total loss, at least in part based on the first entity and the second entity, and wherein the ranking model comprises a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages as claimed.
However, Liu discloses training a machine-learned ranking model by minimizing a total loss, at least in part based on a first entity and a second entity ([0021-0023]: training a machine learning ranking model represented by a ranking function by minimizing a total loss a first and second entities with a usage of a loss function); and
partitioning a space of feature values into regions and assigning each region a value that involve usage of different symbol values such as j=1, 2, ...n(i) ([0020-0022], [0041], [0055-0061]: partition a space represented by a list of features vectors each with an assigned vales, such that the math equation is defined using different mathematical symbol values to rank potential matches)
Since (i) Hueter further discloses the training of ranking model adopts function to minimize squared error, and (ii) Hueter and Liu are from the same field of because both are directed to determine matches between entities based on associated feature indicators, it would have been obvious to one skilled in the art at the time of the invention to modify the error computation & ranking model application of Hueter and incorporate the total loss computation & ranking model application of Liu in ranking model training of teachings of Hueter for minimizing loss and to rank potential match as claimed. The motivation to combine is to provide intelligently rank matches resulting highly relevant information provision for users (Hueter, [0001]; Liu, [0003]).
Liu does not explicitly disclose R9, j=1, 2, ... , J for the regions; a value <Di to define a function T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri for each of the regions, wherein the ranking model comprises a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages as claimed.
However, this difference is only found in the nonfunctional descriptive data materials that describe the following:
data labels of R9, j=1, 2, ... , J for the partitioned regions,
a data value assigned to each of the regions,
an intended use of region value assignment that is directed to define a function T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri for each of the regions, and
the ranking includes boosted regression tree that is an aggregate of regression trees computed in a sequence of stages.
Neither of data label (or data value) or the intended use, nor the boosted regression tree that is an aggregate of regression trees computed in a sequence of stage is functionally involved in the steps recited.
All the steps in the claims (e.g. detecting, determining, partitioning, assigning, training, and apply) would be performed the same regardless of what the data labels for the regions and region values are. Similarly, all steps in the claims would be performed the same regardless of intended use of the region value assignment. Likewise, all steps in the claims would be performed the same regardless of whether the ranking includes boosted regression tree that is an aggregate of regression trees computed in a sequence of stages or not. None of them hve any of the functionalities of the claimed steps as all the steps would be performed the same to achieve the same outcome.
Therefore, it would have been obvious to one skilled in the art before the effective filing date of the claimed invention to (i) use any data label for the partitions, (ii) assign any value to each region for any usage and (iii) include any data information in the ranking model because such data material does not functionally relate to the steps in the method claimed. Also, it is because the subjective interpretation of the data does not patentably distinguish the claimed invention.
Further, Sarlos discloses wherein the ranking model comprises a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages ([0032]: ranking model with a boosted regression trees which computes in stages as boosted regression trees are configured).
Since Hueter. Liu and Sarlos are from the same field of because all are directed to query processing to provide relevant data information, it would have been obvious to one skilled in the art at the time of the invention to modify and combined their teachings by incorporate the boosted regression trees technique of Sarlos into the combined teachings of Hueter and Liu to rank potential match as claimed. The motivation to combine is to provide intelligently rank matches resulting highly relevant information provision for users with improved search results (Hueter, [0001]; Liu, [0003]; Sarlos, [0005])
With respect to claim 38, Hueter discloses an apparatus (Abstract, [0001]), comprising:
a processor; and a memory including instructions, the processor, upon executing the instructions ([0001]), to
detect a first behavioral feature for a first potential match for the first entity and a second behavioral feature for a second potential match for the second entity, wherein the first behavioral feature indicates a degree of at least one-way interest in the first entity by a third entity, and the second behavioral feature indicates a degree of at least one-way interest in the second entity by the third entity (elements on indicating degree of one way interest appears to be directed to non-functional descriptive material for not functionally impacting the structure of the claim because the steps in the claim would be performed the same regardless of the indicated degree of one-way interest ; [0012-0014], [0023-0024], [0043],Fig 2-5: detect behavioral features based on a subject of an active entity’s behaviors toward other subjects representing other entities for respective potential match. The detected behavioral features including and not limited to explicit and inferred/derived behavior features of the active entity that represents the 3rd entity. The explicit behavioral features include explicit input or feedbacks of relevancy indicating a degree of at least one-way interest by the 3rd entity toward search results being presented to the 3rd entity. Inferred/derived behavior features include implicit feedback—including and not limited to non-selection of search results--which also indicate a degree of a degree of at least one-way interest by the 3rd entity toward search results being presented to the 3rd entity. Each search result represents an entity. Hence a 1st and 2nd entities are being represented by two search results among the plurality of search results being presented. Since the plurality of search results being presented and the active entity are being tracked, hence a first behavior feature for a first potential match and a 2nd behavior features for a second potential match are being detected upon explicit and inferred/derived behaviors of the active entity);
determine a probability of relevance of the first and second potential matches based at least in part upon the first and second behavioral features ([0012-0015], [0020-0024], [0031], [0043], Fig 2-5: determine probability of relevance based behavioral features when training a learning method and/or when determine relevance of one or more matches, which are based on the behavior features in view of input or lack of input of the 3rd entity. Also the scores such as the relevance/behavioral score may be corresponding to the probability of relevance as well since it is also based on the behavioral features as well);
partition a space of feature values into regions ([0042-0044], [0046-0047], Fig 4: the partitioning a space of presented search results representing the feature values into at least two 2 regions according to feedback, e.g. a region of selected result, and a region of unselected region, and/or a region of relevant and a region of not relevant);
assign each of the regions a value ([0042-0044], [0046-0047], Fig 4: assigning each region a value, e.g. a value of selected or unselected, and/or a value of relevant or not relevant);
train a machine-learned ranking model (i) using a subset of features defined by the plurality of candidate profiles and the probability of relevance of each of the first and second potential matches as inputs ([0030-0031],[0042-0046], Fig 4-6: training a ranking model at least in part is based on profile features and the probability of relevance as the profiles features are being used matching processing to generate return results representing the entities including the 1st and 2nd entities as described in [0030] & [0042], and then used to update entity profiles based on feedback for the subsequent matches that includes a potential match for a fourth entity indicated by subsequent search results as described in [0046]));
perform an application of the ranking model to rank a potential match for a fourth entity [0030-0031],[0042-0046], Fig 4-6: apply ranking model to rank potential matches-- such as match in the subsequent search for the 4th entity of the potential matches which is the intended use entity of the ranking model application and does not necessary carry any patentable weight),
the fourth entity being an entity for which behavioral features are absent (the limitation is directed to non-functional descriptive material which does not impact functionality of any claimed steps; [0043], Fig 3 & 6: the forth entity represented by a given result return to the user is being entity without behavioral features with no feedback from the user yet).
Hueter does not explicitly discloses R9, j=1, 2, ... , J for the regions; a value <Di to define a function T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri for each of the regions;
train the machine-learned ranking model by minimizing a total loss, at least in part based on the first entity and the second entity;
wherein the ranking model comprises a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages as claimed.
However, Liu discloses train a machine-learned ranking model by minimizing a total loss, at least in part based on a first entity and a second entity ([0021-0023]: training a machine learning ranking model represented by a ranking function by minimizing a total loss a first and second entities with a usage of a loss function); and
partitioning a space of feature values into regions and assigning each region a value that involve usage of different symbol values such as j=1, 2, ...n(i) ([0020-0022], [0041], [0055-0061]: partition a space represented by a list of features vectors each with an assigned vales, such that the math equation is defined using different mathematical symbol values to rank potential matches)
Since (i) Hueter further discloses the training of ranking model adopts function to minimize squared error, and (ii) Hueter and Liu are from the same field of because both are directed to determine matches between entities based on associated feature indicators, it would have been obvious to one skilled in the art at the time of the invention to modify the error computation & ranking model application of Hueter and incorporate the total loss computation & ranking model application of Liu in ranking model training of teachings of Hueter for minimizing loss and to rank potential match as claimed. The motivation to combine is to provide intelligently rank matches resulting highly relevant information provision for users (Hueter, [0001]; Liu, [0003]).
Liu does not explicitly disclose R9, j=1, 2, ... , J for the regions; a value <Di to define a function T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri for each of the regions; wherein the ranking model comprises a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages as claimed.
However, this difference is only found in the nonfunctional descriptive data materials that describe the following:
data labels of R9, j=1, 2, ... , J for the partitioned regions,
a data value assigned to each of the regions,
an intended use of region value assignment that is directed to define a function T(u, v), where T(u, v) = (Di if feature vector fu,v E Ri for each of the regions, and
the ranking includes boosted regression tree that is an aggregate of regression trees computed in a sequence of stages.
Neither of data label (or data value) or the intended use, nor the boosted regression tree that is an aggregate of regression trees computed in a sequence of stage is functionally involved in the steps recited.
All the steps in the claims (e.g. detecting, determining, partitioning, assigning, training, and apply) would be performed the same regardless of what the data labels for the regions and region values are. Similarly, all steps in the claims would be performed the same regardless of intended use of the region value assignment. Likewise, all steps in the claims would be performed the same regardless of whether the ranking includes boosted regression tree that is an aggregate of regression trees computed in a sequence of stages or not. None of them hve any of the functionalities of the claimed steps as all the steps would be performed the same to achieve the same outcome.
Therefore, it would have been obvious to one skilled in the art before the effective filing date of the claimed invention to (i) use any data label for the partitions, (ii) assign any value to each region for any usage and (iii) include any data information in the ranking model because such data material does not functionally relate to the steps in the method claimed. Also, it is because the subjective interpretation of the data does not patentably distinguish the claimed invention.
Further, Sarlos discloses wherein the ranking model comprises a boosted regression tree that is an aggregate of regression trees computed in a sequence of stages ([0032]: ranking model with a boosted regression trees which computes in stages as boosted regression trees are configured).
Since Hueter. Liu and Sarlos are from the same field of because all are directed to query processing to provide relevant data information, it would have been obvious to one skilled in the art at the time of the invention to modify and combined their teachings by incorporate the boosted regression trees technique of Sarlos into the combined teachings of Hueter and Liu to rank potential match as claimed. The motivation to combine is to provide intelligently rank matches resulting highly relevant information provision for users with improved search results (Hueter, [0001]; Liu, [0003]; Sarlos, [0005]).
With respect to claim 26, 33 and 39, the combined teachings of Hueter and Liu further discloses wherein the applying is based at least in part on a first feature vector indicating features of the profile of the first entity and a second feature vector indicating features of the profile of the second entity (the limitations are directed to non-functional descriptive material on describe what the applying is, and not necessary what the applying does, which may not carry patentable weight; Hueter, [0042-0044], [0046-0047], Fig 4: the applying of ranking is performed in part by partitioning a space of presented search results representing the feature values into 2 regions according to feedback. One region being selected and the region being unselected that being used to derive relevance scores and weights (E.g. the selected are consider high relevance and higher weight when the non-selected are consider low-relevance and low weight) in user/entity profiling, which is being used in the ranking model application for subsequent matches to rank subsequent search results. Also, object vector cluster techniques are being used in matching and ranking of entities: region of interest and a region of non-interest; Liu; [0020-0022], [0041]:the applying is based on feature vectors that indicate the respective features,).
With respect to claim 27, 34, and 40, the combined teachings of Hueter and Liu further discloses wherein the ranking model ranks the potential match for the fourth entity, based at least in part on a vector of the fourth entity and at least one of the regions (the limitation is directed to non-functional descriptive material on describe what the ranking model is intended for, which may not carry patentable weight; Hueter, [0042], [0046-0047]: retrieve source and target vectors--which include a vector the 4th entity as the 4th entity is one of the targets –in the matching and ranking of entities/objects based on at least one of the regions, such as the relevance and weight of the selected search results in the profile, or region of interest as set forth by the vector clustering techniques are being used; Liu; [0020], [0041]: ranking model ranks based on vectors).
With respect to claims 28, 35 and 41, the combined teachings of Hueter and Liu further disclose wherein the first behavioral feature pertains to a view of the profile of the first entity, and the second behavioral feature pertains to a view of the profile of the second entity (elements appears to be directed to non-functional descriptive material for not functionally impacting the structure of the claims and the profile is merely a type of data, the behavior features are pertain or related of the profile which are not necessary views of the profiles, hence the steps in the claim would be performed the same regardless of the type of data being used and which may not carry patentable weight; Hueter [0012-0014], [0023-0024], [0043], Fig 2-5: the behavioral features, such as and not limited to explicit feedbacks of relevancy, selection, and non-selection of results representing the entities in the potential matches, are pertain to a view of a profile, as the result presented may be directed to profile of the respective entity of a dating service since the apparatus is applicable to dating service application as described in [0028]; Liu, [0020-0025]: the features pertain to different type of attribute, which is merely a type of data of the items correspond to the profile).
With respect to claims 29, 36 and 42, the combined teachings of Hueter and Liu further discloses wherein the first and second behavioral features do not indicate click feedback from the first entity, the second entity, or the third entity ( the term “or” indicates that only one of the listed--the first entity, the second entity, or the third entity-- is needed to read on the limitation; Hueter, [0012-0014], [0023-0024], [0043], Fig 2-5: when the 1st and 2nd behavioral features are being inferred/derived behavior features such as non-selection of search results by the 3rd entity when search results are being presented to the 3rd entity. Each search result represents an entity. The non-selections of two results representing the first and second behavioral features. Also the first behavioral feature indicates a degree of at least one-way interest in the first entity by a third entity and the second behavioral feature indicates a degree of at least one-way interest in the second entity by the third entity as presented in the independent claims 45 & 52 & 59 have indicated that the first and second behavioral features are from the 3rd entity, hence the first and second behavioral features do not provide click feedback from the first entity and the second entity since neither behavioral features are from the first and second entities. Even if they the first and second entities do provide behavioral features, those behavioral features are not limited to click feedback because behavioral features include inferred/derived behavior features as stated above).
With respect to claims 30 and 43, the combined teachings of Hueter and Liu further discloses wherein the ranking model is trained to predict features that correlate with relevance (the limitation is directed to non-functional descriptive material on describe what the ranking model is, which may not carry patentable weight; Hueter, [0031], [0043], [0046], Fig 3-5; Liu, [0020-0023], [0041]: training ranking model based on the probability of relevance to predict features/matches).
With respect to claims 31, 37 and 44, the combined teachings of Hueter and Liu further discloses wherein the ranking model minimizes the total loss, based on a gradient descent method (Liu, [0021], [0041]: minimize the total loss using neural network and gradient decent as optimization algorithm).
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Michelle Owyang whose telephone number is (571)270-1254. The examiner can normally be reached Monday-Friday, 8am-6pm EST.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Charles Rones can be reached at (571)272-4085. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/MICHELLE N OWYANG/Primary Examiner, Art Unit 2168
Prosecution Insights