Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claim 7 and 17 rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the enablement requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to enable one skilled in the art to which it pertains, or with which it is most nearly connected, to make and/or use the invention.
The following limitations in claim 7 are not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor, at the time the application was filed, had possession of the claimed invention: wherein, for each respective variable group comprising multiple input variables, an intra-group correlation of the multiple input variables within the respective variable group is stronger than an intra-group correlation between (i) any one of the multiple input variables within the respective variable group and (ii) any other input variable outside of the respective variable group.
The above recited features of claim 7 are not defined or described in applicant’s specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor, at the time application was filed, had possession of the claimed invention. The term “intra-group correlation” does not appear anywhere in the specification with the only real mention of group correlation between outside variable groups being this line in [0008] “At least one group in the plurality of groups includes two or more input variables, and correlation of input variables within a group is stronger than correlation of input variables between groups”. As such the applicant’s specification fails to define or describe any form of intra group correlation in such a way as to reasonably convey to one skilled in the relevant art that the inventor, at the time the application was filed, had possession of the claimed invention.
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.
Claim’s 7 and 17 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
Regarding Claim 7 and 17, the conditional statement of a stronger relationship between finding the strongest relationship of the variables within the respective variable group would not be a factor in identifying the strongest relationship groups since the comparison of finding the strongest input variable would be a comparison of the variable compared against itself which cannot be used to identify oneself as the stronger variable over the other.
Claim Interpretation
Regarding claim 7 the limitation wherein, for each respective variable group comprising multiple input variables, an intra-group correlation of the multiple input variables within the respective variable group is stronger than an intra-group correlation between (i) any one of the multiple input variables within the respective variable group and (ii) any other input variable outside of the respective variable group has no clear definition on the two conditional statements with “(i) any one of the multiple input variables within the respective variable group and (ii) any other input variable outside of the respective variable group”. At most the specification defines this limitation as “At least one group in the plurality of groups includes two or more input variables, and correlation of input variables within a group is stronger than correlation of input variables between groups.”
Given the indefiniteness of the phrase, and in the interest of simplicity, the examiner has chosen to interpret “wherein, for each respective variable group comprising multiple input variables, an intra-group correlation of the multiple input variables within the respective variable group is stronger than an intra-group correlation between (i) any one of the multiple input variables within the respective variable group and (ii) any other input variable outside of the respective variable group” to mean “wherein, for each respective variable group comprising multiple input variables, an intra-group correlation of the multiple input variables within the respective variable group is stronger than an intra-group correlation between any other input variable outside of the respective variable group.”
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim’s 1-4, 9, 11-15, 19 are rejected under U.S.C. 103 as being unpatentable over Manu Joseph (NPL: Interpretability: Cracking open the black box – Part III) (“Joseph”) in view of Tharrington et al (US10510022B1) (“Tharrington”)
Regarding claim 1, Joseph teaches A computing platform comprising: at least one communication interface; at least one processor; data storage comprising at least one non-transitory computer-readable medium; and program instructions stored in the data storage that, when executed by the at least one processor, cause the computing platform to perform functions comprising:
training a machine learning model by carrying out a machine learning process on a training dataset comprising a set of input vectors and a corresponding set of output values, wherein the trained machine learning model is configured to (i) receive an input vector comprising respective values for a given set of input variables and (ii) based on an evaluation of the received input vector, generate and output a prediction ([Algorithm-Page 13] teaches getting predictions from the model for the coalition vector by converting the coalition vector to the original sample space. The original sample space here is interpreted as a training dataset with the collation vector functioning as a input. The algorithm teaches getting a prediction for the vector teaching the generate and output a prediction part of the limitation.)
based on an evaluation of correlations between input variables in the given set of input variables of the trained machine learning model, dividing the given set of input variables into a plurality of variable groups, wherein each variable group includes at least one input variable from the given set of input variables and each of at least a subset of the variable groups includes multiple input variables from the given set of input variables that are determined to have a threshold level of correlation with one another ([submodular pick and global explanations-Page 6] teaches a program that uses the groupby function on different classes and gets the mean average from each class. The different classes are interpreted as the variable groups with the individual values present in the data frame for each prediction being interpreted as input variable. [interpretation-Page 8] teaches what the correlations with these values with a marital status above 0.5 pointing towards being single and pointing towards you earning less 50k with being married pointing towards you having a higher likelihood of you making above 50k. The threshold level here being the yearly earning with the correlation being the marital status teaching this limitation)
processing the input vector using the trained machine learning model to generate and output a given prediction corresponding to the given input vector ([submodular pick and global explanations-Page 6] the dataframe object is made off of a models prediction on if a person is making above or below 50k [Page 8])
using a model interpretability technique based on game theory to determine, for each respective variable group in the plurality of variable groups, a respective Shapley-based value that quantifies a respective contribution of the respective variable group to the given prediction for the given input vector ([kernel SHAP-Page 13] teaches using the shapely kernel to compute the weight of the sample of the coalition vector. With Shapely values here being a technique similar to game theory [6. Shapely Values-Page 9] teaching the interpretability technique based on game theory. These shapely values here are used to predict based off of different factors whether a person is making above or below 50k [Implementation (local explanation)-Page 20])
and based on the respective Shapley-based values determined for the respective variable groups in the plurality of variable groups, identifying which one or more input variables of the trained machine learning model contributed most to the given prediction ([Interpretation-Page 24] teaches a feature in dependence plots in SHAP where it picks out what feature is having the most interaction with their investigated feature, the feature here being if the person is above or below the 50 mark teaching this limitation.)
Joseph does not teach after training the machine learning model and dividing the given set of input variables into the plurality of variable groups: receiving a given input vector comprising respective values for the given set of input variables;
However, Tharrington does teach after training the machine learning model and dividing the given set of input variables into the plurality of variable groups: receiving a given input vector comprising respective values for the given set of input variables ([Column 10: 48-59] teaches computing a feature contribution value for each variable to a predicted result using the trained model. The feature contribution application is modifying and scores the vectors which is being used along with a model to compute Shapley values. This process is being interpreted as receiving a input vector comprising values for a given set of input variables. [Column 13: Lines 14-20] teaches the relationship between feature contribution application and selection vector which would comprise a value representing a group of variables teaching the plurality of variable groups part of the limitation [Column 14: 60-67 and Column 15: 1-11].)
Joseph and Tharrington are analogous art because they both deal with machine learning and interpretability
It would have been obvious to a person skilled in the art before the effective filling data of the
claimed invention to combine Joseph with the respective value representation of Tharrington. Doing so would be able to provide means of interpreting their prediction results giving users more confidence in how well the model is performing its tasks ([Tharrington Column 1: 22-32])
Regarding claim 2, the combination of Joseph and Tharrington teaches all the limitations of claim 1
Joseph also teaches wherein using the model interpretability technique based on game theory to determine, for each respective variable group in the plurality of variable groups, the respective Shapley-based value that quantifies a respective contribution of the respective variable group to the given prediction for the given input vector comprises: for each respective variable group comprising multiple input variables: using the model interpretability technique based on game theory to determine variable-specific Shapley-based values for the multiple input variables in the respective variable group ([Implementation (Local Explanation): Pages 17-18] teaches a SHAP force plot used on different examples of people with different characteristics ranging from marital status to race. Then plotting those characteristics and the model output values for those different characteristics. Those characteristics are being interpreted as variables with the different examples functioning as variable groups teaching this limitation.
And aggregating the variable-specific Shapley-based values for the multiple input variables into the respective Shapley-based value for the respective variable group ([implementation (Global Explanation): Page 21] teaches aggregating and plotting shapely values across the different variables to see their overall impact and correctly predicting if someone makes above are below 50k.)
and for each respective variable group comprising a single input variable: using the model interpretability technique based on game theory to determine a variable-specific Shapley-based value for the single input variable in the respective variable group ([implementation (local explanation): Page 16-17] teaches calculating shapely values, with SHAP having a relationship to game theory teaching the interpretability technique limitation. The force plot shows different variables and their relationship on accurately predicting if a person is making above or below 50k. Those different factors are being interpreted as input variables with the example they are using that represents a person, that is being interpreted as a variable group, which teaches this limitation)
and utilizing the variable-specific Shapley-based value for the single input variable as the respective Shapley-based value for the respective variable group ([implementation (local explanation): Page 17] the shapely values are used to predict if a person is making above or below 50k teaching this limitation.)
Regarding claim 3, the combination of Joseph and Tharrington teaches all the limitations of claim 1
Joseph also teaches wherein the respective Shapley-based value that is determined for each respective variable group in the plurality of variable groups comprises a respective Shapley Additive Explanation (SHAP) value ([Implementation (Global Explanation): Page 21-22] teaches calculating values and plotting them on a graph for the different examples)
Regarding claim 4, the combination of Joseph and Tharrington teaches all the limitations of claim 1
Joseph also teaches wherein the model interpretability technique based on game theory comprises a TreeSHAP model interpretability technique ([Implementation (Local Explanation): Page 16] teaches for the SHAP values being generated they are using TreeSHAP)
Regarding claim 11, the combination of Joseph and Tharrington teaches all the limitations of claim 1
Tharrington also teaches wherein the trained machine learning model comprises a neural network model or a tree-based model ([Column 9: 22 to 35] teaches a model indicator which includes neural network models)
Regarding claim 12, see the analysis of claim 1
Regarding claim 13, see the analysis of claim 1
Regarding claim 14, see the analysis of claim 2
Regarding claim 15, see the analysis of claim 3
Claims 5-6, 16 are rejected under U.S.C. 103 as being unpatentable over Manu Joseph (NPL: Interpretability: Cracking open the black box – Part III) (“Joseph”) in view of Tharrington et al (US10510022B1) (“Tharrington”) and Yoon et al (US8385662B1) (“Yoon”)
Regarding claim 5, the combination of Joseph and Tharrington teaches all the limitations of claim 1
the combination of Joseph and Tharrington teaches all the limitations of claim 1 does not teach wherein dividing the given set of input variables into the plurality of variable groups based on the evaluation of the correlations between input variables in the given set of input variables comprises: applying a clustering technique for grouping variables based on correlation to a set of input vectors that each comprise respective values for the given set of input variables.
However, Yoon does teach wherein dividing the given set of input variables into the plurality of variable groups based on the evaluation of the correlations between input variables in the given set of input variables comprises: applying a clustering technique for grouping variables based on correlation to a set of input vectors that each comprise respective values for the given set of input variables ([Column 1: Lines 64-67 and Column 2: Lines 1-7] teaches clustering entities with spatial or temporal characteristics. With those entities being represented by feature vectors, teaching the input vector part of the limitation. The vectors would contain values to represent the different aspects of the feature teaching the input variables part of the limitation. The clustering is being done based off of spatial or temporal characteristics teaching the correlation to a set of vectors part of the limitation.)
Joseph, Tharrington, and Yoon are analogous art because they all use machine learning in a way to examine relationships between data
It would have been obvious to a person skilled in the art before the effective filling data of the
claimed invention to combine Joseph with the respective value representation of Tharrington and the clustering technique of Yoon. Doing so would lead to a less computationally intensive process by not having to except the same algorithm multiple times ([Yoon-Column 1: 52-60]).
Regarding claim 6, the combination of Joseph, Tharrington, and Yoon teaches all the limitations of claim 5
Yoon also teaches wherein the clustering technique is based on principal component analysis (PCA) ([Column 2: Lines 2-4] teaches a set of clusters being determined by a PCA algorithm)
Regarding claim 16, see the analysis of claim 5
Claim 7 and 17 are rejected under U.S.C. 103 as being unpatentable over Manu Joseph (NPL: Interpretability: Cracking open the black box – Part III) (“Joseph”) in view of Tharrington et al (US10510022B1) (“Tharrington”) and Nushni et al (US11562180B2) (“Nushni”)
Regarding claim 7, the combination of Joseph and Tharrington teaches all the limitations of claim 1
the combination of Joseph and Tharrington does not teach wherein, for each respective variable group comprising multiple input variables, an intra-group correlation of the multiple input variables within the respective variable group is stronger than an intra-group correlation between (i) any one of the multiple input variables within the respective variable group and (ii) any other input variable outside of the respective variable group
However, Nushni does teach wherein, for each respective variable group comprising multiple input variables, an intra-group correlation of the multiple input variables within the respective variable group is stronger than an intra-group correlation between (i) any one of the multiple input variables within the respective variable group and (ii) any other input variable outside of the respective variable group ([Column 11 Lines 19-34] teaches generating a cluster based on a combination of features having a higher correlation to failed outputs in comparison to other combinations of features. Features here are interpreted as input variables and the combination of features is
interpreted as variable groups. It can be reasonably interpreted that in order to find the combination of features that they would have to be compared against each other teaching the outside grouping of variable groups teaching this limitation)
Joseph, Tharrington, and Nushni are analogous art because they all deal with machine learning and interpretability
It would have been obvious to a person skilled in the art before the effective filling data of the claimed invention to combine Joseph with the respective value representation of Tharrington and the stronger variable group identification of Nushni. Doing so would improve training by understanding what is causing a machine learning model to fail during training ([Nushni-Column 1: 31-38])
Regarding claim 17, see the analysis of claim 7,
Claim’s 8-9 and 18-19 are rejected under U.S.C. 103 as being unpatentable over Manu Joseph (NPL: Interpretability: Cracking open the black box – Part III) (“Joseph”) in view of Tharrington et al (US10510022B1) (“Tharrington”) and Nourian et al (US20210049503A1) (“Nourian”)
Regarding claim 8, the combination of Joseph and Tharrington teaches all the limitations of claim 1
the combination of Joseph and Tharrington does not teach wherein the model interpretability technique based on game theory comprises a model interpretability technique, and wherein the computing platform further comprises program instructions stored in the data storage that, when executed by the at least one processor, cause the computing platform to perform functions comprising: after processing the input vector using the trained machine learning model to generate and output a given prediction corresponding to the given input vector, using a second model interpretability technique based on a partial dependence plot (PDP) framework to determine, for each respective variable group in the plurality of variable groups, a respective PDP-based value that quantifies the respective contribution of the respective variable group to the given prediction for the given input vector.
However, Nourian does teach wherein the model interpretability technique based on game theory comprises a model interpretability technique, and wherein the computing platform further comprises program instructions stored in the data storage that, when executed by the at least one processor, cause the computing platform to perform functions comprising: after processing the input vector using the trained machine learning model to generate and output a given prediction corresponding to the given input vector, using a second model interpretability technique based on a partial dependence plot (PDP) framework to determine, for each respective variable group in the plurality of variable groups, a respective PDP-based value that quantifies the respective contribution of the respective variable group to the given prediction for the given input vector ([0049] teaches using PDP to measure the relationships between features and the predicted outcome of the model. [0048] teaches measuring input vectors and how they contribute to the model’s output. It can be reasonably interpreted that input vectors and its effect on the outcome is also measured when they are using PDP to plot the relationship. [0047] teaches using monotonicity analysis to measure the relationship between features and the predicted outcome teaching another technique used for model interpretability teaching this limitation.)
Joseph, Tharrington, and Nourian are analogous art because they all deal with machine learning and measuring relationships
It would have been obvious to a person skilled in the art before the effective filling data of the claimed invention to combine Joseph with the respective value representation of Tharrington and the PDP quantification of Nourian. Doing so would lead to models less prone to errors or producing undesirable results by providing a proper understanding of the models behavior ([Nourian-0005])
Regarding claim 9, the combination of Joseph, Tharrington, and Nourian teaches all the limitations of claim 8
Joseph also teaches further comprising program instructions stored in the data storage that, when executed by the at least one processor, cause the computing platform to perform functions comprising: for each respective variable group in the plurality of variable groups, aggregating the respective Shapley-based value determined for the respective variable group with the respective PDP-based value determined for the respective variable group and thereby producing an aggregated value that quantifies the respective contribution of the respective variable group to the given prediction for the given input vector ([Implementation (Global Explanations): Page 21] teaches dependence plots for each different variable from their aggregated shapely values. Dependence plots here are said to be analogous to PDP teaching this limitation. These plotted values showcase the contribution of each respective variable group on getting the correct prediction teaching this limitation)
Regarding claim 18, see the analysis of claim 8
Regarding claim 19, see the analysis of claim 9
Claim’s 10 and 20 are rejected under U.S.C. 103 as being unpatentable over Manu Joseph (NPL: Interpretability: Cracking open the black box – Part III) (“Joseph”) in view of Tharrington et al (US10510022B1) (“Tharrington”) and Jiang et al (US20200218987A1) (“Jiang”)
Regarding claim 10, the combination of Joseph and Tharrington teaches all the limitations of claim 1
the combination of Joseph and Tharrington does not teach further comprising program instructions stored in the data storage that, when executed by the at least one processor, cause the computing platform to perform functions comprising: using the identified one or more input variables as a basis for generating one or more adverse action reason codes (AARCs)
However, Jiang does teach further comprising program instructions stored in the data storage that, when executed by the at least one processor, cause the computing platform to perform functions comprising: using the identified one or more input variables as a basis for generating one or more adverse action reason codes (AARCs) ([0049] teaches selecting a certain number of adverse features and the adverse features that make business intuition based on the selection are given adverse action reason codes. The adverse features here are being interpreted as input variables teaching this limitation)
Joseph, Tharrington, and Jiang are analogous art because they all deal with machine learning and interpretability
It would have been obvious to a person skilled in the art before the effective filling data of the claimed invention to combine Joseph with the respective value representation of Tharrington and the AARC generation of Jiang. Doing so would allow for easier identification of contributing factors towards a prediction ([Jiang-0002])
Conclusion
The prior art made of record and relied upon is considered to applicant’s disclosure
Pauws et al US20210118557A1 (2019-04-17) ([Abstract] “The present disclosure pertains to a system for providing model-based predictions of beneficiaries receiving out-of-network care. In some embodiments, the system (i) obtains, from one or more databases, a collection of information related to care utilization and expenditures for a plurality of beneficiaries; (ii) extracts, from the collection of information, information related to healthcare services rendered to the beneficiaries within a predetermined time period; (iii) provides the extracted healthcare services information to a machine learning model to train the machine learning model; (iv) obtains characteristics information related to a current beneficiary and a corresponding healthcare provider; and (v) provides, subsequent to the training of the machine learning model, the current patient and corresponding healthcare provider characteristics information to the machine learning model to predict a likelihood of a future healthcare service provided to the current beneficiary to be rendered out-of-network.”)
Jordan et al US20200387832A1 (2017-10-30) ([Abstract] “Certain aspects involve training tree-based machine-learning models for computing predicted responses and generating explanatory data for the models. For example, independent variables having relationships with a response variable are identified. Each independent variable corresponds to an action or observation for an entity. The response variable has outcome values associated with the entity. Splitting rules are used to generate the tree-based model, which includes decision trees for determining relationships between independent variables and a predicted response associated with the response variable. The tree-based model is iteratively adjusted to enforce monotonicity with respect to representative response values of the terminal nodes. For instance, one or more decision trees are adjusted such that one or more representative response values are modified and a monotonic relationship exists between each independent variable and the response variable. The adjusted model is used to output explanatory data indicating relationships between independent variable changes and response variable changes.”
Any inquiry concerning this communication or earlier communications from the examiner should be directed to URIAH V MOORE whose telephone number is (571)384-8341. The examiner can normally be reached Monday-Friday 8am-5pm.
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, Mariela Reyes can be reached at (571)270-1006. 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.
/URIAH VENDELL MOORE/ Examiner, Art Unit 2142 /Mariela Reyes/Supervisory Patent Examiner, Art Unit 2142