DETAILED ACTION
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 10/15/2025 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 .
Response to Amendment
Applicant has amended the independent claims to include the limitations “that form the basis of an auto regressive model” and “including the basis of the auto regressive model”. Applicant has provided citations to paragraphs in the specification which include support for the additions of such limitations. Examiner has confirmed that the cited sections adequately support the newly added language to the claims.
The newly-added limitations change the scope of the claims so as to limit the training of the decision-tree-based model with training data “including the basis of the autoregressive model”. Accordingly, this change in scope necessitated further search and consideration of the claimed matter.
Response to Arguments
Claim rejections under 35 U.S.C. § 101
Applicant argues that the pending claims allegedly recite techniques for altering the operation of a well which improves the technological field of wells and integrates any alleged abstract idea into a practical application. Applicant provides excerpts from the specification ¶161 in support of this argument.
Examiner disagrees that the excerpt provides adequate support for the improvement of the technological field of wells. The referenced excerpt from the specification states “the models can be used for more than just determining whether, where, and how to drill new wells- they can use actual well production to predict proposed and existing well performance more accurately”. This statement is an admission that the claimed invention (the models) are an improvement of mental process (i.e. determining and predicting) to achieve “more accurate” results. Per MPEP 2106.05(a)(II), “However, it is important to keep in mind that an improvement in the abstract idea itself (e.g. a recited fundamental economic concept) is not an improvement in technology.” The mechanism by which the alteration of well operation (so as to improve wells) is not apparent by the specification.
Applicant further argues that the pending claims particularly recite the alteration of a choke or lift on a well based on the predicted GOR. Applicant argues that Examiner’s previous suggestion that the examined claims recited such alteration in an “unspecified and non-specific way” is incorrect because supposedly one of skill in the art would understand the potentially complex alterations of a choke or lift on a well based on predicted GORs.
Examiner disagrees because altering a choke or life based on a predicted value generated from a model does not provide a specific solution to a particular problem but instead recites the idea of an outcome or a solution without providing details as to how the solution as achieved. Stating that a person having skill in the art would recognize the potentially complex alterations of a choke or lift and merely stating “based on” does not indicate a clear link between the predicted value and how such information influences the alteration of the choke or lift. Per MPEP 2106.04(d)(1), “Conversely, if the specification explicitly sets forth an improvement but in a conclusory manner (i.e., a bare assertion of an improvement without the detail necessary to be apparent to a person of ordinary skill in the art), the examiner should not determine the claim improves technology. An indication that the claimed invention provides an improvement can include a discussion in the specification that identifies a technical problem and explains the details of an unconventional technical solution expressed in the claim, or identifies technical improvements realized by the claim over the prior art.”.
Applicant further argues that Examiner’s analysis is relevant to the memo provided by the Office on August 4, 2025 regarding “Reminders on evaluating subject matter eligibility of claims under 35 U.S.C. § 101.
Regarding applicant’s reliance on the August 2025 memorandum clarifying how examiners should apply § 101 (“Memo”) and the decision of the Appeals Review Panel in Ex parte Desjardins, No. 2024-000567 (P.T.A.B. Sept. 26, 2025), Examiner notes that the Memo explicitly states that it “is not intended to announce any new USPTO guidance or procedure and is meant to be consistent with existing USPTO guidance.”
Applicant argues that the pending claims do not invoke computers or other machinery as “merely a tool to perform an existing process” but instead “the claims purport to improve computer capabilities or to improve an existing technology”. Applicant further asserts that the claims as a whole provide improvement to technology or a technical field.
Examiner disagrees. The recitation and use of an autoregressive model and a decision-tree based model are recitations of generic computing components recited at a high level of generality and functioning in their normal capacity to enable performance of the mental process of prediction. The specificity by which the models are described are not described in the claims in such a way that an inventive concept would provide an improvement to such computing components. Any purported improvement in the claims flows as a direct consequence to the improvement in, or a direct result of the abstract idea(s) itself.
For the reasons stated in this response, in conjunction with the rejection provided in this action, the rejections under 35 U.S.C. § 101 have been maintained.
Claim rejections under 35 U.S.C. § 103
Applicant argues that the cited art does not teach or suggest features incorporated as part of the amendments to the independent claims, particularly with regard to the training data and subsequently training a decision-tree-based model with the training data including the basis of the autoregressive model. Applicant specifically argues that none of the data disclosed by Anderson is used for forming the basis of an autoregressive model included in training data for training the decision-tree-based model. Applicant acknowledges that the data of Anderson is utilized in various models but argues that the ensemble learning method disclosed by Anderson is not the same technique as disclosed in the instant application which does not involve training data including data from autoregressive models for training a decision tree based model. Applicant further acknowledges that Anderson discloses the utilization of several autoregressive models for time series classification tasks but further argues that the disclosed methods are different from the claimed invention.
Applicants arguments have been fully considered but are moot because the new grounds of rejection set forth in this action does not rely on the reference applied in the prior rejection for the subject matter being specifically challenged in the applicant’s argument.
The new grounds set forth relies on Meek for the challenged limitations, wherein Meek discloses and Autoregressive Tree model wherein time-series data is used to construct an autoregressive model via transformation and subsequently the transformed dataset is used to learn/train a decision tree for the target variable ((Meek, Page 229 ¶2) " Roughly speaking, we construct ART models as follows. First, we use a standard \windowing" transformation of a time-series data set into a set of cases suitable for a regression analysis, where the \predictor variables" and \target variable" in the analysis correspond to the preceding values and current value, respectively, in the time series. This transformation is often used when constructing autoregressive models. Then, we use the transformed data set to learn a decision tree for the target variable.")
Claim Objections
Claim 21 is objected to because of the following informalities:
The preamble of the claim recites “The method of claim 1”, however Claim 1 is a system claim which conflicts with such reference. For purposes of this examination, Examiner has assumed that this is a typographical error and should instead read “The system of claim 1”. Alternately the claim could be modified to instead recite “The method of claim 14”.
Appropriate correction is required.
Claim Rejections - 35 USC § 112
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.
Claims 1, 14, and 20 recite “that form the basis of an autoregressive model”. “The basis” lacks antecedent basis. Further, Claims 5, 6 recite “a basis” after the introduction of “the basis” in the respective independent claims which creates ambiguity as to whether the recitations are referring to the same element or not. For purposes of this examination, the initial recitation in the independent claims has been interpreted to recite “that form a basis of an autoregressive model” and the recitations in claims 5 and 6 are interpreted to recite “the basis” such that the claims all refer to such basis as a singular element.
The dependent claims 2-6, 8-13, 15-17, 19, and 21 incorporate the deficiency of the independent claims and are therefore rejected under the same rationale.
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 1-6, 8-17, and 19-21 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. Particularly, recitation of Mental Processes (MPEP 2106.04(a)(2)(III)) are noted. The following section follows the 2019 Patent Eligibility Guidance (PEG) for analyzing subject matter eligibility:
Step 1 - Statutory Category:
Step 1 of the PEG analysis entails considering whether the claimed subject matter falls within the four statutory categories of patentable subject matter identified by 35 USC 101 (process, machine, manufacture, or composition of matter). (MPEP 2106.03)
Step 2A Prong 1 - Judicial exception:
In Step 2A Prong 1, examiners evaluate whether the claim recites a judicial exception (an abstract idea, law of nature, or a natural phenomenon). (MPEP 2106.04)(a-b)
Step 2a Prong 2 - Integration into a practical application:
If claims recite a judicial exception, the claim requires further analysis in Step 2A Prong 2. In Step 2A Prong 2, examiners evaluate whether the claim as a whole integrates the exception into a practical application. (MPEP 2106.04)(d)
Step 2B Significantly More:
If the additional elements identified in Step 2A Prong 2 do not integrate the exception into a practical application, then the claim is directed to the recited judicial exception and requires further analysis under Step 2B- Significantly More.
As noted in the MPEP 2106.05(II): The identification of the additional element(s) in the claim from Step 2A Prong 2, as well as the conclusions from Step 2A Prong 2 on the considerations discussed in MPEP 2106.05(a) -(c), (e), (f), and (h) are to be carried over. Claim limitations identified as Insignificant Extra-Solution Activities are further evaluated to determine if the elements are beyond what is well -understood, routine, and conventional (WURC) activity, as dictated by MPEP 2106.05(II). (MPEP 2106.05)
Independent Claims:
Claim 1:
Step 1: Claim 1 and its dependent claims 2-6, 8-13, and 21 are directed to a system which falls within one of the four statutory categories of a machine.
Step 2A Prong 1: Claim limitations that have been identified as judicial exceptions are indicated in bold. Claim 1 recites wherein the one or more predicted production values are generated by the decision-tree-based model based on an internal structure of the decision-tree-based model, the new time-independent input feature values, and the new time-dependent input feature values, wherein the one or more predicted production values of the well include a predicted gas oil ratio (GOR); The claim limitation can be reasonably read to entail evaluating the internal structure of the model, the new time-independent input feature values, and the new time-dependent input feature values to derive predicted production values. This task can be done within the human mind or using a pen and paper as an assistive physical aid. The limitation includes the usage of a decision-tree-based model, which is a computing component recited at a high level of generality. This limitation, as drafted, is a process that, under broadest reasonable interpretation, covers performance of the mind using generic computing components as a tool to perform the concept (MPEP 2106.04(a)(2)(III)(C)) and is thus the recitation of the judicial exception of abstract ideas as a mental process. The claim also recites the limitation wherein each further predicted production value of the further predicted production values is recursively generated by the decision-tree-based model based on the internal structure of the decision tree- based model using (i) ground-truth initial production values of the well, (ii) a set of the one or more predicted production values of the well, and (iii) a set of prior further predicted production values, wherein the further predicted production values of the well include further predicted GORs; which can reasonably be read to entail utilizing ground truth initial production values, a set of the one or more predicted production values, and a set of prior further predicted production values to further predict production values by leveraging a recursive decision-tree-based model. This task can be performed within the human mind or using a pen and paper as an assistive physical aid because the task is executing algorithmic steps on known data sources. The limitation includes the usage of a decision-tree-based model, which is a computing component recited at a high level of generality. This limitation, as drafted, is a process that, under broadest reasonable interpretation, covers performance of the mind using generic computing components as a tool to perform the concept (MPEP 2106.04(a)(2)(III)(C)) and is thus the recitation of the judicial exception of abstract ideas as a mental process. Therefore, the claim recites a judicial exception.
Step 2a Prong 2: Additional elements were identified and are noted in italics. With regard to the limitations: “persistent storage containing training data related to well production, wherein entries in the training data respectively include time-independent input feature values and time-dependent input feature values both mapped to ground-truth production values of corresponding wells at particular points in time, wherein the time-dependent input feature values include ground-truth production values of the corresponding wells at respectively earlier points in time that form the basis of an autoregressive model; and one or more processors configured to” and “train a decision-tree-based model with the training data including the basis of the autoregressive model”. These limitations have been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)) because they limit the use of the judicial exception to a particular technological environment. The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). These limitations are further characterized as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) because the claim limitations invoke the use of computing components (persistent storage, one or more processors, training a decision tree based model) as tools to perform an existing process. The courts have also found that invoking the use of generic computing components does not integrate the judicial exception into a practical application. The limitations: “obtain, from the persistent storage, the training data”; “provide, to the decision-tree-based model, new time-independent input feature values and new time-dependent input feature values for a well”; “receive, from the decision-tree-based model, one or more predicted production values of the well”; “receive, from the decision-tree-based model, further predicted production values of the well”; and “write, to the persistent storage, the one or more predicted production values and the further predicted production values; and” have been identified as the Insignificant Extra Solution activity of necessary data gathering and outputting MPEP 2106.05(g). The courts have ruled that adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). The limitation “alter a choke or a lift on the well based on the predicted GOR and the further predicted GORs generated using the decision-tree-based model” has been identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) because the limitation amounts to merely utilizing the value obtained in the mental process in an unspecified and generic way to alter a choke or lift of the well which amounts to the recitation of “apply it” with regard to the judicial exception. The courts have found that reciting the words “apply it” or an equivalent with the judicial exception does not integrate the judicial exception into a practical application.
Step 2B: As discussed in Step 2A Prong 2, additional elements were identified as linking the judicial exception to a Field of Use and Technological Environment (MPEP 2106.05(h)) and merely reciting the words “apply it” as in Mere Instructions to Apply an Exception (MPEP 2106.05(f)) which the courts have found does not amount to significantly more than the abstract idea. Additional elements were identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) which must be further evaluated to determine if they are beyond WURC activities: “obtain, from the persistent storage, the training data”; “provide, to the decision-tree-based model, new time-independent input feature values and new time-dependent input feature values for a well”; “receive, from the decision-tree-based model, one or more predicted production values of the well”; “receive, from the decision-tree-based model, further predicted production values of the well”; and “write, to the persistent storage, the one or more predicted production values and the further predicted production values; and.” When read in light of the specification, these activities are recited at a high level of generality to be data gathering and outputting. The claim limitation “obtain, from the persistent storage, the training data” encompasses retrieving information in memory. The claim limitation “provide, to the decision-tree-based model, new time-independent input feature values and new time-dependent input feature values for a well”, when read in light of the specification ¶47-58 encompasses transmitting data over a network. The limitation “receive, from the decision-tree-based model, one or more predicted production values of the well,” and “receive, from the decision-tree-based model, further predicted production values of the well” when read in light of the specification ¶47-58 encompasses receiving data over a network. The limitation and “write, to the persistent storage, the one or more predicted production values and the further predicted production values” encompasses storing information in memory. The Insignificant Extra Solution Activities of necessary data gathering at outputting, as receiving or transmitting data over a network and storing and retrieving information in memory, are recognized by the court as Well-Understood, Routine, and Conventional computer functions when they are claimed in a merely generic manner (MPEP 2106.05(d)(II)(i)). Therefore, the claim does not include additional elements, alone or in combination that are sufficient to amount to significantly more than the recited judicial exceptions.
Conclusion: Based on this rationale, the claim has been deemed to be ineligible subject matter under 35 USC 101.
Claim 14:
Step 1: Claim 14 and its dependent claims 15-17 and 19 are directed to a method which falls within one of the four statutory categories of a process.
Step 2A Prong 1: Claim limitations that have been identified as judicial exceptions are indicated in bold. Claim 14 recites wherein the one or more predicted production values are generated by the decision-tree-based model based on an internal structure of the decision-tree-based model, the new time-independent input feature values, and the new time-dependent input feature values, wherein the one or more predicted production values of the well include a predicted gas oil ratio (GOR); and The claim limitation can be reasonably read to entail evaluating the internal structure of the model, the new time-independent input feature values, and the new time-dependent input feature values to derive a predicted production value. This task can be done within the human mind or using a pen and paper as an assistive physical aid. The limitation includes the usage of a decision-tree-based model, which is a computing component recited at a high level of generality. This limitation, as drafted, is a process that, under broadest reasonable interpretation, covers performance of the mind using generic computing components as a tool to perform the concept (MPEP 2106.04(a)(2)(III)(C)) and is thus the recitation of the judicial exception of abstract ideas as a mental process. The claim also recites the limitation wherein each further predicted production value of the further predicted production values is recursively generated by the decision-tree-based model based on the internal structure of the decision-tree-based model using (i) ground-truth initial production values of the well, (ii) a set of the one or more predicted production values of the well, and (iii) a set of prior further predicted production values, wherein the further predicted production values of the well include further predicted GORs; which can reasonably be read to entail utilizing ground truth initial production values, a set of the one or more predicted production values, and a set of prior further predicted production values to further predict production values by leveraging a recursive decision-tree-based model. This task can be performed within the human mind or using a pen and paper as an assistive physical aid because the task is executing algorithmic steps on known data sources. The limitation includes the usage of a decision-tree-based model, which is a computing component recited at a high level of generality. This limitation, as drafted, is a process that, under broadest reasonable interpretation, covers performance of the mind using generic computing components as a tool to perform the concept (MPEP 2106.04(a)(2)(III)(C)) and is thus the recitation of the judicial exception of abstract ideas as a mental process. Therefore, the claim recites a judicial exception.
Step 2a Prong 2: Additional elements were identified and are noted in italics. With regard to the limitations: “wherein entries in the training data respectively include time-independent input feature values and time dependent input feature values both mapped to ground-truth production values of corresponding wells at particular points in time, wherein the time-dependent input feature values include ground-truth production values of the corresponding wells at respectively earlier points in time that form the basis of an autoregressive model”; and “training a decision-tree-based model with the training data including the basis of the autoregressive model”; these limitations have been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)). The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). These limitations are further characterized as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) because the claim limitations invoke the use of computing components (persistent storage, one or more processors, training a decision tree based model) as tools to perform an existing process. The courts have also found that invoking the use of generic computing components does not integrate the judicial exception into a practical application. The limitations: “obtaining, from persistent storage, training data related to well production”; “providing, to the decision-tree-based model, new time-independent input feature values and new time-dependent input feature values for a well”; “receiving, from the decision-tree-based model, one or more predicted production values of the well”; “receiving, from the decision-tree-based model, further predicted production values of the well,”; and “writing, to the persistent storage, the one or more predicted production values and the further predicted production values; and” have been identified as the Insignificant Extra Solution activities of necessary data gathering and outputting MPEP 2106.05(g). The courts have ruled that adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). The limitation “altering a choke or a lift on the well based on the predicted GOR and the further predicted GORs generated using the decision-tree-based model.” has been identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) because the limitation amounts to merely utilizing the value obtained in the mental process in an unspecified and generic way to alter a choke or lift of the well which amounts to the recitation of “apply it” with regard to the judicial exception. The courts have found that reciting the words “apply it” or an equivalent with the judicial exception does not integrate the judicial exception into a practical application.
Step 2B: As discussed in Step 2A Prong 2, additional elements were identified as linking the judicial exception to a Field of Use and Technological Environment (MPEP 2106.05(h)) and merely reciting the words “apply it” as in Mere Instructions to Apply an Exception (MPEP 2106.05(f)) which the courts have found does not amount to significantly more than the abstract idea. Additional elements were identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) which must be further evaluated to determine if they are beyond WURC activities: “obtaining, from persistent storage, training data related to well production”; “providing, to the decision-tree-based model, new time-independent input feature values and new time-dependent input feature values for a well”; “receiving, from the decision-tree-based model, one or more predicted production values of the well”; “receiving, from the decision-tree-based model, further predicted production values of the well”; and “writing, to the persistent storage the one or more predicted production values and the further predicted production values; and”. When read in light of the specification, these activities are recited at a high level of generality to be data gathering and outputting. The claim limitation “obtaining, from persistent storage, training data related to well production” encompasses retrieving information in memory. The claim limitation “providing, to the decision-tree-based model, new time-independent input feature values and new time-dependent input feature values for a well”, when read in light of the specification ¶47-58 encompasses transmitting data over a network. The limitation “receiving, from the decision-tree-based model, one or more predicted production values of the well and writing, to the persistent storage, the one or more predicted production values” and “receiving, from the decision-tree-based model, further predicted production values of the well,” when read in light of the specification ¶47-58 encompasses receiving data over a network. The “limitation and “writing, to the persistent storage the one or more predicted production values and the further predicted production values; and” encompasses storing information in memory. The Insignificant Extra Solution Activities of necessary data gathering at outputting, as receiving or transmitting data over a network and storing and retrieving information in memory, are recognized by the court as Well-Understood, Routine, and Conventional computer functions when they are claimed in a merely generic manner. (MPEP 2106.05(d)(II)(i)). Therefore, the claim does not include additional elements, alone or in combination that are sufficient to amount to significantly more than the recited judicial exceptions.
Conclusion: Based on this rationale, the claim has been deemed to be ineligible subject matter under 35 USC 101.
Claim 20:
Step 1: Claim 20 is directed to an article of manufacture which falls within one of the four statutory categories of a manufacture.
Step 2A Prong 1: Claim limitations that have been identified as judicial exceptions are indicated in bold. Claim 14 recites wherein the one or more predicted production values are generated by the decision-tree-based model based on an internal structure of the decision-tree-based model, the new time-independent input feature values, and the new time-dependent input feature values, wherein the one or more predicted production values of the well include a predicted gas oil ratio (GOR); . The claim limitation can be reasonably read to entail evaluating the internal structure of the model, the new time-independent input feature values, and the new time-dependent input feature values to derive a predicted production value. This task can be done within the human mind or using a pen and paper as an assistive physical aid. The limitation includes the usage of a decision-tree-based model, which is a computing component recited at a high level of generality. This limitation, as drafted, is a process that, under broadest reasonable interpretation, covers performance of the mind using generic computing components as a tool to perform the concept (MPEP 2106.04(a)(2)(III)(C)) and is thus the recitation of the judicial exception of abstract ideas as a mental process. The claim additionally recites the limitation wherein each further predicted production value of the further predicted production values is recursively generated by the decision-tree-based model based on the internal structure of the decision-tree-based model using (i) ground-truth initial production values of the well, a set of the one or more (ii) predicted production values of the well, and (iii) a set of prior further predicted production values, wherein the further predicted production values of the well include further predicted GORs; which can reasonably be read to entail utilizing ground truth initial production values, a set of the one or more predicted production values, and a set of prior further predicted production values to further predict production values by leveraging a recursive decision-tree-based model. This task can be performed within the human mind or using a pen and paper as an assistive physical aid because the task is executing algorithmic steps on known data sources. The limitation includes the usage of a decision-tree-based model, which is a computing component recited at a high level of generality. This limitation, as drafted, is a process that, under broadest reasonable interpretation, covers performance of the mind using generic computing components as a tool to perform the concept (MPEP 2106.04(a)(2)(III)(C)) and is thus the recitation of the judicial exception of abstract ideas as a mental process. Therefore, the claim recites a judicial exception.
Step 2a Prong 2: Additional elements were identified and are noted in italics. With regard to the limitations: “wherein entries in the training data respectively include time-independent input feature values and time-dependent input feature values both mapped to ground-truth production values of corresponding wells at particular points in time, wherein the time-dependent input feature values include ground-truth production values of the corresponding wells at respectively earlier points in time that form the basis of an autoregressive model”; and “training a decision-tree-based model with the training data including the basis of the autoregressive model”; these limitations have been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)). The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). These limitations are further characterized as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) because the claim limitations invoke the use of computing components (persistent storage, one or more processors, training a decision tree based model) as tools to perform an existing process. The courts have also found that invoking the use of generic computing components does not integrate the judicial exception into a practical application. The limitations: “obtaining, from persistent storage, training data related to well production”; “providing, to the decision-tree-based model, new time-independent input feature values and new time-dependent input feature values for a well”; “receiving, from the decision-tree-based model, one or more predicted production values of the well”; “receiving, from the decision-tree-based model, further predicted production values of the well”; and “writing, to the persistent storage, the one or more predicted production values and the further predicted production values; and” have been identified as the Insignificant Extra Solution activity of necessary data gathering and outputting MPEP 2106.05(g). The courts have ruled that adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). The limitation “altering a choke or a lift on the well based on the predicted GOR and the further predicted GORs generated using the decision-tree-based model.” has been identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) because the limitation amounts to merely utilizing the value obtained in the mental process in an unspecified and generic way to alter a choke or lift of the well which amounts to the recitation of “apply it” with regard to the judicial exception. The courts have found that reciting the words “apply it” or an equivalent with the judicial exception does not integrate the judicial exception into a practical application.
Step 2B: As discussed in Step 2A Prong 2, additional elements were identified as linking the judicial exception to a Field of Use and Technological Environment (MPEP 2106.05(h)) and merely reciting the words “apply it” as in Mere Instructions to Apply an Exception (MPEP 2106.05(f)) which the courts have found does not amount to significantly more than the abstract idea. Additional elements were identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) which must be further evaluated to determine if they are beyond WURC activities: “obtaining, from persistent storage, training data related to well production”; “providing, to the decision-tree-based model, new time-independent input feature values and new time-dependent input feature values for a well”; “receiving, from the decision-tree-based model, one or more predicted production values of the well”; “receiving, from the decision-tree-based model, further predicted production values of the well,”; and “writing, to the persistent storage, the one or more predicted production values and the further predicted production values; and”. When read in light of the specification, these activities are recited at a high level of generality to be data gathering and outputting. The claim limitation “obtaining, from persistent storage, training data related to well production” encompasses retrieving information in memory. The claim limitation “providing, to the decision-tree-based model, new time-independent input feature values and new time-dependent input feature values for a well”, when read in light of the specification ¶47-58 encompasses transmitting data over a network. The limitation “receiving, from the decision-tree-based model, one or more predicted production values of the well” and “receiving, from the decision-tree-based model, further predicted production values of the well,” when read in light of the specification ¶47-58 encompasses receiving data over a network. The “limitation and “writing, to the persistent storage the one or more predicted production values and the further predicted production values; and” encompasses storing information in memory. The Insignificant Extra Solution Activities of necessary data gathering at outputting, as receiving or transmitting data over a network and storing and retrieving information in memory, are recognized by the court as Well-Understood, Routine, and Conventional computer functions when they are claimed in a merely generic manner. (MPEP 2106.05(d)(II)(i)). Therefore, the claim does not include additional elements, alone or in combination that are sufficient to amount to significantly more than the recited judicial exceptions.
Conclusion: Based on this rationale, the claim has been deemed to be ineligible subject matter under 35 USC 101.
Ordered Combination:
With regard to the dependent claims, limitations are analyzed both individually and as part of the ordered combination. Unless explicitly stated otherwise, dependent claims do not appear to substantially change the functionality of the device as a whole beyond what is noted previously in their respective independent claims analysis. For the sake of brevity, the discussion of the ordered combination in view of the dependent limitations has not been restated.
Dependent Claims:
Regarding claim 2, the rejection of independent claim 1 is further incorporated. Claim 2 recites additional element “wherein the time-independent input feature values relate to lateral lengths of the corresponding wells, proppant pumped into the corresponding wells, fluid pumped into the corresponding wells, or porosity of rock in which the corresponding wells are respectively disposed”. This limitation has been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)). The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional elements viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The additional element amounts to no more than field of use and technological environment which does not amount to significantly more than the abstract idea (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 3, the rejection of independent claim 1 is further incorporated. Claim 3 recites additional element “wherein the time-dependent input feature values include inter-well spacing values, artificial lift, choke, pressure, or whether the corresponding wells are parent wells”. This limitation has been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)). The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional element viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The additional element amounts to no more than field of use and technological environment which does not amount to significantly more than the abstract idea (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 4, the rejection of independent claim 1 is further incorporated. Claim 4 recites additional element “wherein the ground-truth production values represent volumes of hydrocarbons or water extracted from the corresponding wells”. This limitation has been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)). The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional elements viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The additional element amounts to no more than field of use and technological environment which does not amount to significantly more than the abstract idea (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 5, the rejection of independent claim 1 is further incorporated. Claim 5 additionally recites the limitation wherein the ground-truth production values of the corresponding wells at respectively earlier points in time are data that form a basis of an order 1 autoregressive model, which can reasonably be read to entail utilizing historical data points in an order 1 autoregressive model. When read in light of the specification, an autoregressive model can be expressed as a mathematical formula (Instant Specification, ¶163-167). Thus, this limitation includes recitation of the mathematical concept of Mathematical Formulas or Equations (2106.04(a)(2)(1)(B)). Therefore, the claim recites a judicial exception (Step 2A Prong 1). The claim does not include any additional elements that integrate the judicial exception into a practical application (Step 2A Prong 2), nor amount to significantly more (Step 2B) than the judicial exception. This claim is not eligible subject matter under 35 USC 101.
Regarding claim 6, the rejection of independent claim 1 is further incorporated. Claim 6 additionally recites limitations wherein the ground-truth production values of the corresponding wells at respectively earlier points in time are data that form a basis of an order 2 autoregressive model. which can reasonably be read to entail utilizing historical data points in an order 2 autoregressive model. When read in light of the specification, an autoregressive model can be expressed as a mathematical formula (Instant Specification, ¶163-167). Thus, this limitation includes recitation of the mathematical concept of Mathematical Formulas or Equations (2106.04(a)(2)(1)(B)). Therefore, the claim recites a judicial exception (Step 2A Prong 1). The claim does not include any additional elements that integrate the judicial exception into a practical application (Step 2A Prong 2), nor amount to significantly more (Step 2B) than the judicial exception. This claim is not eligible subject matter under 35 USC 101.
Regarding claim 8, the rejection of independent claim 1 is further incorporated. Claim 8 recites additional element “wherein the particular points in time are regular intervals”. This limitation has been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)). The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional element viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The additional element amounts to no more than field of use and technological environment which does not amount to significantly more than the abstract idea (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 9, the rejection of independent claim 1 and dependent claim 8 is further incorporated. Claim 9 recites additional element the regular intervals are 5 days, 10 days, 15 days, 60 days, or 90 days. This limitation has been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)). The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional element viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The additional element amounts to no more than field of use and technological environment which does not amount to significantly more than the abstract idea (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 10, the rejection of independent claim 1 is further incorporated. Claim 10 recites additional element “wherein the time-dependent input feature values include anomalous events”. This limitation has been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)). The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional element viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The additional element amounts to no more than field of use and technological environment which does not amount to significantly more than the abstract idea (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 11, the rejection of independent claim 1 and dependent claim 10 is further incorporated. Claim 11 additionally recites limitations generate a variation of the time-dependent input feature values with the one or more of the anomalous events removed which can reasonably be read to entail evaluating the indications of the anomalous events to make a judgement of which time-dependent input feature values are to be removed. This task can be done within the human mind or using a pen and paper as an assistive physical aid. The limitation includes the usage of one or more processors, which is a computing component recited at a high level of generality. This limitation, as drafted, is a process that, under broadest reasonable interpretation, covers performance of the mind using generic computing components as a tool to perform the concept (MPEP 2106.04(a)(2)(III)(C)) and is thus the additional recitation of the judicial exception of abstract ideas as a mental process. Claim 11 further recites wherein the one or more counterfactual predicted production values are generated by the decision-tree-based model based on the internal structure of the decision-tree-based model, the new time-independent input feature values, and the variation of the time-dependent input feature values, which can reasonably be read to entail evaluating the structure of the decision-tree-based model along with the new time-independent input feature values and a variation of the new time-dependent input feature values to predict a hypothetical production value. This task can be done within the human mind or using a pen and paper as an assistive physical aid. The limitation includes the usage of one or more processors, which is a computing component recited at a high level of generality. This limitation, as drafted, is a process that, under broadest reasonable interpretation, covers performance of the mind using generic computing components as a tool to perform the concept (MPEP 2106.04(a)(2)(III)(C)) and is thus an additional recitation of the judicial exception of abstract ideas as a mental process. Therefore, the claim recites judicial exceptions (Step 2A Prong 1). Claim 11 further recites additional elements “receive indications of one or more of the anomalous events to remove from the time-dependent input feature values”, “provide, to the decision-tree-based model, the new time-independent input feature values and the variation of the time-dependent input feature values for the well; and” and “receive, from the decision-tree-based model, one or more counterfactual predicted production values of the well,”. These limitations have been identified as the Insignificant Extra Solution activity of necessary data gathering and outputting MPEP 2106.05(g). The courts have ruled that adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional elements viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The Insignificant Extra Solution Activities of necessary data gathering at outputting, as receiving or transmitting data over a network, are recognized by the court as Well-Understood, Routine, and Conventional computer functions when they are claimed in a merely generic manner. (MPEP 2106.05(d)(II)(i)). Therefore, the claim does not include additional elements, alone or in combination that are sufficient to amount to significantly more than the recited judicial exception. (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 12, the rejection of independent claim 1 and dependent claim 10 is further incorporated. Claim 12 additionally recites limitations generate a variation of the time-dependent input feature values with the one or more of the hypothetical anomalous events which can reasonably be read to entail evaluating the indications of the anomalous events to make a judgement of which time-dependent input feature values are to be added. This task can be done within the human mind or using a pen and paper as an assistive physical aid. The limitation includes the usage of one or more processors, which is a computing component recited at a high level of generality. This limitation, as drafted, is a process that, under broadest reasonable interpretation, covers performance of the mind using generic computing components as a tool to perform the concept (MPEP 2106.04(a)(2)(III)(C)) and is thus the additional recitation of the judicial exception of abstract ideas as a mental process. Claim 12 further recites wherein the one or more counterfactual predicted production values are generated by the decision-tree-based model based on the internal structure of the decision-tree-based model, the new time-independent input feature values, and the variation of the time-dependent input feature values, which can reasonably be read to entail evaluating the structure of the decision-tree-based model along with the new time-independent input feature values and a variation of the new time-dependent input feature values to predict a hypothetical production value. This task can be done within the human mind or using a pen and paper as an assistive physical aid. The limitation includes the usage of one or more processors, which is a computing component recited at a high level of generality. This limitation, as drafted, is a process that, under broadest reasonable interpretation, covers performance of the mind using generic computing components as a tool to perform the concept (MPEP 2106.04(a)(2)(III)(C)) and is thus an additional recitation of the judicial exception of abstract ideas as a mental process. Therefore, the claim recites judicial exceptions (Step 2A Prong 1). Claim 12 further recites additional elements “receive indications of one or more hypothetical anomalous events to add to the time dependent input feature values;” , “provide, to the decision-tree-based model, the new time-independent input feature values and the variation of the new time-dependent input feature values for the well; and” and “receive, from the decision-tree-based model, one or more counterfactual predicted production values of the well”. These limitations have been identified as the Insignificant Extra Solution activity of necessary data gathering and outputting MPEP 2106.05(g). The courts have ruled that adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional elements viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The Insignificant Extra Solution Activities of necessary data gathering at outputting, as receiving or transmitting data over a network, are recognized by the court as Well-Understood, Routine, and Conventional computer functions when they are claimed in a merely generic manner. (MPEP 2106.05(d)(II)(i)). Therefore, the claim does not include additional elements, alone or in combination that are sufficient to amount to significantly more than the recited judicial exception. (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 13, the rejection of independent claim 1 is further incorporated. Claim 13 additionally recites limitations determine Shapley data for the time-independent input feature values and the time dependent input feature values, wherein the Shapley data indicate correlations between: (i) each of the time-independent input feature values and the time-dependent input feature values, and (ii) the ground-truth production values of the corresponding wells, which can reasonably be read to entail evaluating the time-independent and time-dependent input feature values along with the ground-truth production values to determine correlations and Shapley values. This task can be done within the human mind or using a pen and paper as an assistive physical aid. The limitation includes the usage of one or more processors, which is a computing component recited at a high level of generality. This limitation, as drafted, is a process that, under broadest reasonable interpretation, covers performance of the mind using generic computing components as a tool to perform the concept (MPEP 2106.04(a)(2)(III)(C)) and is thus the recitation of the judicial exception of abstract ideas as a mental process. Furthermore, when read in light of the specification in ¶203-204, the Shapley data quantifies and assigns a contribution to each of the input features with respect to the output values of a model, which is the recitation of mathematical relationships 2106.04(a)(2)(I)(A) and thus the claim additionally recites the judicial exception of abstract ideas as mathematical concepts. Therefore, the claim recites judicial exceptions (Step 2A Prong 1). Claim 13 further recites additional elements “generate a representation of a graphical user interface for the Shapley data, wherein each of the time-independent input feature values and the time-dependent input feature values is associated with a scatterplot of the respective Shapley data for the corresponding wells; and” and “transmit, to a client device, the representation of the graphical user interface.”. These limitations have been identified as the Insignificant Extra Solution activity of necessary data outputting MPEP 2106.05(g). The courts have ruled that adding insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional elements viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The limitation “generate a representation of a graphical user interface for the Shapley data, wherein each of the time-independent input feature values and the time-dependent input feature values is associated with a scatterplot of the respective Shapley data for the corresponding wells; and” is interpreted to mean outputting a scatterplot of Shapley for display. One having ordinary skill in the art would recognize this activity as well-understood, routine, and conventional. The factual evidence to support this determination is that commercially off-the-shelf software products such as the well-known Python SHAP library have functionality that enable visualization of Shapley values via scatterplots. For example, a tutorial for visualizing Shapley data to include the use of scatterplots (using the Python SHAP library) was published on Kaggle, prior to the effective filing date of this application, whereby Kaggle is a widely-recognized data-science website full of training and resources for the field. (Kaggle, “Tutorial: Machine Learning Interpretability”, April 22, 2019, https://www.kaggle.com/code/datacog314/tutorial-machine-learning-interpretability). A screenshot of the reference is provided. Activities that are recognized as well-understood, routine, and conventional do not amount to significantly more than the abstract idea.
PNG
media_image1.png
186
840
media_image1.png
Greyscale
PNG
media_image2.png
930
1902
media_image2.png
Greyscale
The limitation “transmit, to a client device, the representation of the graphical user interface” is recited at a high level of generality and when read in light of the specification ¶47-58 encompasses receiving data over a network. The Insignificant Extra Solution Activities of necessary data outputting, as transmitting data over a network, is recognized by the court as Well-Understood, Routine, and Conventional computer functions when they are claimed in a merely generic manner (MPEP 2106.05(d)(II)(i)). Therefore, the claim does not include additional elements, alone or in combination that are sufficient to amount to significantly more than the recited judicial exception. (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 15, the rejection of independent claim 14 is further incorporated. Claim 15 recites additional element “wherein the time-independent input feature values relate to lateral lengths of the corresponding wells, proppant pumped into the corresponding wells, fluid pumped into the corresponding wells, or porosity of rock in which the corresponding wells are respectively disposed”. This limitation has been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)). The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional element viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The additional element amounts to no more than field of use and technological environment which does not amount to significantly more than the abstract idea (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 16, the rejection of independent claim 14 is further incorporated. Claim 16 recites additional element “wherein the time-dependent input feature values include inter-well spacing values, artificial lift, choke, pressure, or whether the corresponding wells are parent wells”. This limitation has been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)). The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional element viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The additional element amounts to no more than field of use and technological environment which does not amount to significantly more than the abstract idea (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 17, the rejection of independent claim 14 is further incorporated. Claim 17 recites additional element “wherein the ground-truth production values represent volumes of hydrocarbons or water extracted from the corresponding wells.” This limitation has been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)). The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional element viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The additional element amounts to no more than field of use and technological environment which does not amount to significantly more than the abstract idea (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 19, the rejection of independent claim 14 is further incorporated. The claim language for claim 19 is substantially similar to that of claim 11 and therefore the mirrored limitations are rejected under the same rationale noted previously. Claim 19 additionally recites limitation identified as an additional element that is not within claim 11 language: “wherein the time-dependent input feature values include anomalous events, the method further comprising:”. This limitation has been interpreted as Field of Use and Technological Environment (MPEP 2106.05(h)). The courts have ruled that generally linking the use of a judicial exception to a particular technological environment or field of use, as discussed in MPEP 2016.05(h) does not integrate the judicial exception into a practical application (MPEP 2106.04(d)). With the additional element viewed as part of the ordered combination, the claim as a whole does not appear to integrate the judicial exception into a practical application (Step 2A Prong 2). The additional element amounts to no more than field of use and technological environment which does not amount to significantly more than the abstract idea (Step 2B). This claim is not eligible subject matter under 35 USC 101.
Regarding claim 21, the rejection of independent claim 1 is further incorporated. Claim 21 recites the additional element “wherein the choke or the lift on the well is altered to achieve a target production on the well”. This limitation has been identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) for amounting to a generic recitation of the words “apply it”. The limitation, as written recites the outcome of an idea or solution without providing how the solution to the problem is accomplished. That is to say that the claim does not specify what particularly is altered on the choke or lift so as to achieve the desired outcome of achieving targe production. The courts have found that reciting the words “apply it” or a generic equivalent in addition to a recited judicial exception does not integrate the judicial exception into a practical application nor amount to significantly more (Step 2A Prong 2 and Step 2B). This claim is not eligible subject matter under 35 USC 101.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 1-6, 8, 10-11, 14-17, and 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over Anderson et al. (U.S. Patent Pub No. 2017/0364795 A1), hereinafter referred to as Anderson, in view of Meek et al (Meek, C., Chickering, D., and Heckerman, D., “Autoregressive Tree Models for Time-Series Analysis”, 2002, Proceedings of the 2002 SIAM International Conference on Data Mining, Society for Industrial and Applied Mathematics, pp. 229-244), hereinafter referred to as Meek, further in view of Yajun et al (CN109941950A), hereinafter referred to as Yajun, and further in view of Kittilsen et al (Kittilsen, P., Fjalestad, K., and, Aashein, R., “Stabilized and Increased Well Production Using Automatic Choke Control”, November 2014, Paper presented at the Abu Dhabi International Petroleum Exhibition and Conference), hereinafter referred to as Kittilsen.
Regarding Claim 1, Anderson teaches the claim limitations indicated in bold (except those surrounded by brackets ([[…]]): A system comprising: persistent storage containing training data related to well production, wherein entries in the training data respectively include time-independent input feature values and time-dependent input feature values both mapped to ground-truth production values of corresponding wells at particular points in time, wherein the time-dependent input feature values include ground-truth production values of the corresponding wells at respectively earlier points in time [[that form the basis of an autoregressive model]]; and one or more processors configured to: Anderson teaches the PALM system. ((Anderson, ¶3) "The Petroleum Analytics Learning Machine (PALM) is a machine learning based, "brutally empirical" analysis system for use in all upstream and midstream oil and gas operations."). Anderson teaches a system that utilizes the Hadoop framework for data storage and analytics, whereby Hadoop has built-in capabilities for persistent storage of data ((Anderson, ¶60) "A distributed file system facilitates the storage and maintenance of the data, and provides efficient data computations and analytics. For example, Hadoop is a framework that allows for the distributed storage of data and distributed processing of large data sets across clusters of computing resources."). Anderson teaches well production data that includes time-independent feature values and time-dependent feature values which are associated with measured in the field production values (ground-truth production values). ((Anderson, ¶4) “The PALM also uses Support Vector Regression, logistic regression, Bayesian models, nearest neighbors, neural networks and deep learning networks uniquely combined as ensemble learning to weigh the importance of hundreds to thousands of geological, geophysical, and engineering attributes, both measured in the field and computed from theoretical analyses such as reservoir simulation models and 4D seismic and gravity gradiometry monitoring of production changes over time."; (Anderson, ¶11) "The system integration database retrieves, compares and combines geology and geophysics, reservoir modeling, rock properties, drilling, completion, hydraulic fracturing, production and pipeline gathering data into a uniform data repository by linking heterogeneous data sources with normalization based on common unique identifiers."; See also Fig. 4 for data stored within the system integration database (SID)).). The data points for all collected data are associated with a particular point in time. ((Anderson, ¶5) “The time and depth for each data point of the collected data are recorded."). The data stored within the System Integration Database includes historical data and the data is used for subsequent machine learning components such that the machine learning components have “learned” rankings based on the data and therefore the data is considered training data. ((Anderson, ¶5) "Incoming data over a communications network are received and stored into a system integration database by a processor based server or cloud-based distribution of servers to provide collected data for analyses. The incoming data comprises digital exogenous data, real-time and historical endogenous data, historical data from surrounding production wells, hydraulic fracture completion data, and progress, status and maintenance data from new vertical and horizontal wells, including kickoffs, sidetracks, step-outs, pipeline gathering systems, compressor stations and other kinds of oil and gas sensor data including from public and private data sources now existent and of future design. The time and depth for each data point of the collected data are recorded. … The normalized data are processed to determine clusters of correlation in multi-dimensional space to identify a machine learned ranking of Importance Weights for each attribute.".
obtain, from the persistent storage, the training data; See Fig 1 where training data stored within the SID (1300) is obtained by the MAP sub-systems (1200). ((Anderson, ¶4) “These subsystems use the PALM System Integration Database (SID) to retrieve integrated data, then perform machine learning and other statistical analyses of that data, and return to the SID results of computation and predictive and prescriptive actions that can be forwarded by the TOTALVU user inter face (UI) to controllers, human and/or automated, so that real-time optimization of production and minimization of costs can be realized for new wells.")
train a decision-tree-based model with the training data [[including the basis of the autoregressive model]]; Anderson teaches Petroleum Analytics Learning Machine (PALM) as a model which contains a decision tree; the decision tree may be exemplified by CART. ((Anderson, ¶13) “The PALM utilizes at one of the following classification: logistic regression, support vector machine and support vector regression, nearest neighbors, decision trees and random forest, neural networks and deep learning networks."); ((Anderson, ¶59) " Turning now to FIG. 3, in accordance with an exemplary embodiment, there are listed machine analytics algorithms and tools commonly accessible to all MAP 1200 subsystems within the PALM 1000[[…]]. A Regression Predictor 1411 computes linear and support vector regression, and classification and regression trees (CART)[[…]]. A Classification Prescriptor 1421 computes logistic regression, support vector machines, K-Nearest neighbor, Decision tree modeling, and Neural networks and Deep learning."). The data from the SID is used on components of PALM to produce “learned rankings” and “learned seeds”, indicating that training data is used to train components of PALM to produce the learned model values. ((Anderson, ¶5) "The normalized data are processed to determine clusters of correlation in multi-dimensional space to identify a machine learned ranking of Importance Weights for each attribute."; (Anderson, ¶6) “In accordance with an exemplary embodiment of the claimed invention, unstructured textual data are classified to correlate with optimal production by utilizing progressive clustering using region growing from learned seeds, information extraction and retrieval, image recognition, textual mining, keyword and key phrase extraction, semantic and sentiment analysis, entity and pattern recognition and knowledge discovery processing to capture the dynamics of said at least one or all wells of oil and natural gas fields.")
provide, to the decision-tree-based model, new time-independent input feature values and new time-dependent input feature values for a well; Anderson discusses using labeled data for categorization and classification of in PALM (input data). ((Anderson, ¶7) "Categorization and classification results from labeled data sets to identify patterns are provided."). PALM contains functionality for decision-tree models, as stated previously ((Anderson, ¶59) " Turning now to FIG. 3, in accordance with an exemplary embodiment, there are listed machine analytics algorithms and tools commonly accessible to all MAP 1200 subsystems within the PALM 1000[[…]]. A Regression Predictor 1411 computes linear and support vector regression, and classification and regression trees (CART)[[…]]. A Classification Prescriptor 1421 computes logistic regression, support vector machines, K-Nearest neighbor, Decision tree modeling, and Neural networks and Deep learning.").The PALM system uses historical data (time-dependent feature values) and attribute data (time independent feature values) for predicting future production of a well ((Anderson, ¶10) "In accordance with an exemplary embodiment of the claimed invention, a real-time processor of the Petroleum Analytics Learning Machine system convolves importance weight values of attributes received by the Petroleum Analytics Learning Machine system from historical data and attribute data from each new well as it progresses in real time to predict future production of the new well before oil and gas are delivered to the surface."). Anderson also indicates that the received data is modified to create “new” data by combining the received data with real-time exogeneous data. ((Anderson, ¶11) "The received data are combined with real time exogenous data comprising weather forecasts.").
receive, from the decision-tree-based model, one or more predicted production values of the well, wherein the one or more predicted production values are generated by the decision-tree-based model based on an internal structure of the decision-tree-based model, the new time-independent input feature values, and the new time-dependent input feature values, [[wherein the one or more predicted production values of the well include a predicted gas oil ratio (GOR);]] Forecasted (predicted) production values for wells are produced by the MAPPROD subsystem of PALM, where time-independent and time-dependent feature values stored in the system integrated database are used for the prediction. ((Anderson, ¶16) “In accordance with an exemplary embodiment of the claimed invention, the aforesaid system and method queries one or more system integration databases of multiple surrounding wells in an area or querying one integrated master system integration database comprising regionally relevant geologic and geophysical data, reservoir models, drilling data, hydraulic fracturing data, the historical exogenous data, the real-time exogenous data, and the real-time endogenous data to forecast production of said each well."; (Anderson, ¶56) “In accordance with an exemplary embodiment of the claimed invention, the MAPPROD 1240 is a production forecaster that convolves the actual attribute values of hundreds to thousands of attributes coming into the system from historical wells, as well as each new well as it progresses, to maximize production for all wells in a play”). The MAPPROD subsystem (1240) produces the predicted production results and is capable of leveraging all machine learning algorithms and tools included in PALM, which include decision tree modeling amongst others in the classification prescriptor (1421). Because outputs of models are dependent upon the structure of the models, it would be understood by a person having ordinary skill in the art that the generated production value is based on the internal structure of the model used. ((Anderson, ¶59) See Fig. 3; "Turning now to FIG. 3, in accordance with an exemplary embodiment, there are listed machine analytics algorithms and tools commonly accessible to all MAP 1200 subsystems within the PALM 1000."). The PALM system uses historical data (time-dependent feature values) and attribute data (time independent feature values) for predicting future production of a well ((Anderson, ¶10) "Analytics Leaming Machine system from historical data and attribute data from each new well as it progresses in real time to predict future production of the new well before oil and gas are delivered to the surface.").
receive, from the decision-tree-based model, further predicted production values of the well, wherein each further predicted production value of the further predicted production values is recursively generated by the decision-tree-based model based on the internal structure of the decision-tree- based model using (i) ground-truth initial production values of the well, (ii) a set of the one or more predicted production values of the well, and (iii) a set of prior further predicted production values, [[wherein the further predicted production values of the well include further predicted GORs;]] The PALM system takes as input historical data (ground-truth production values) to predict (receive) production values (set of one or more predicted production values) ((Anderson, ¶10) "In accordance with an exemplary embodiment of the claimed invention, a real-time processor of the Petroleum Analytics Learning Machine system convolves importance weight values of attributes received by the Petroleum Analytics Learning Machine system from historical data and attribute data from each new well as it progresses in real time to predict future production of the new well before oil and gas are delivered to the surface."). An initial forecast for production (set of one or more predicted production values) is then used to predict future production (a set of prior further predicted production values) ((Anderson, ¶84) "In accordance with an exemplary embodiment of the claimed invention, the MAPPROD optimizer 1240 convolves the Importance Weights for all wells in each study area f with g which is each attribute value specific to the well for which future production of oil, gas and water is being calculated, wherein f * g is an integral transform of the product of the two functions as attributes specific to that well under study. The integral transform then predicts the future production of the well under study before the oil and gas are delivered to the surface and uses future production to calculate an accuracy of that initial forecast."). Performing predictive optimization is proposed by utilizing unique combinations of machine learning and statistical algorithms to include regressions and decision trees as part of an ensemble learning approach, thereby indicating that a decision tree model may be combined with an autoregressive model to recursively predict based on the decision tree based structure ((Anderson, ¶6) "In accordance with an exemplary embodiment of the claimed invention, predictive and prescriptive optimization are performed on the normalized data utilizing unique combinations of machine learning and statistical algorithm ensembles. The ensembles include at least two of the following: linear and non-linear support vector machines and regressions, naive Bayes, logistic regression, decision trees, hidden Markov models, random forests, gradient boosting machines, neural networks, deep learning networks, among other machine learning models"). The MAPPROD optimizer has access to decision-tree modeling functionality, as stated previously ((Anderson, ¶59) " Turning now to FIG. 3, in accordance with an exemplary embodiment, there are listed machine analytics algorithms and tools commonly accessible to all MAP 1200 subsystems within the PALM 1000[[…]]. A Regression Predictor 1411 computes linear and support vector regression, and classification and regression trees (CART)[[…]]. A Classification Prescriptor 1421 computes logistic regression, support vector machines, K-Nearest neighbor, Decision tree modeling, and Neural networks and Deep learning."). The MAPPROD optimizer has access to autoregression functionality which would be understood by a person having skill in the art to utilize a recursive process to predict time series data ((Anderson, ¶13) " The PALM system utilizes at least one of the following largescale or big data analyses: autoregressive integrated moving average (ARIMA), multivariate time series analysis, hidden Markov models, nonparametric Bayesian models, autoregressive conditional heteroskedasticity (ARCH), exponentially weighted moving average, and generalized autoregressive conditional heteroskedasticity (GARCH).").
write, to the persistent storage, the one or more predicted production values and the further predicted production values; and See Fig. 1 where the SID which has persistent storage capability is connected with no indicated direction to the MAP subsystems (which include the MAPPROD subsystem that produces the predicted production values), indicating that data can be both read from and written to the storage. Furthermore, Anderson teaches that recommended actions are produced based on production values and changes in operations based on the actions are stored within the database, which would reasonably include the predicted production value. ((Anderson, ¶20) "The aforesaid system and method generates at least one predicted condition by the Petroleum Analytics Learning Machine system, and stores resulting changes in operations in the system integration database from field operations in response to a recommended action."). Furthermore, Anderson states that the results are returned to the system integration database. ((Anderson, ¶4) “These subsystems use the PALM System Integration Database (SID) to retrieve integrated data, then perform machine learning and other statistical analyses of that data, and return to the SID results of computation and predictive and prescriptive actions that can be forwarded by the TOTALVU user inter face (UI) to controllers, human and/or automated, so that real-time optimization of production and minimization of costs can be realized for new wells."). The frac control center stores data from action undertaken based on future production values, indicating future production values are also stored in the SID ((Anderson, ¶31) "The Frac control center generates at least one recommendation to increase production or cut costs of a well in progress by controlling a mix of the hydraulic fracturing class outcome using decision trees of the Petroleum Analytics Learning Machine system to maximize an overall ell production. The Frac control center stores data from actions undertaken based on at least one recommendation in the system integration database to provide a feedback to the Petroleum Analytics Learning Machine system about its recommendations based on the future production.")
[[alter a choke or a lift on the well based on the predicted GOR and the further predicted GORs generated using the decision-tree-based model.]]
Anderson does not disclose; however Anderson in view of Meek discloses that form the basis of an autoregressive model Time series data is described as being used in a transformation to construct an autoregressive model ((Meek, Page 229, ¶2) "Roughly speaking, we construct ART models as follows. First, we use a standard windowing" transformation of a time-series data set into a set of cases suitable for a regression analysis, where the \predictor variables" and \target variable" in the analysis correspond to the preceding values and current value, respectively, in the time series. This transformation is often used when constructing autoregressive models.")
including the basis of the autoregressive model A decision tree is learned (trained) using the transformed data which forms the autoregressive model as stated above (( Meek, Page 229, ¶2) "This decision tree has linear regressions at its leaves, thus, producing a piecewise-linear auto-regression model. We use a Bayesian technique to learn the structure and parameters of the decision tree."); ((Meek, Page 238, ¶1) "We evaluate the quality of a learned model by computing the sequential pre-dictive score for the holdout data set corresponding to the training data from which the model was learned.")
Meek is analogous to the claimed invention because it is reasonably pertinent to the problem faced by the inventor- that is making predictions using time-series data. It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have implemented the autoregressive tree model as disclosed by Meek as the predictive model leveraged by Anderson because some teaching, suggestion, or motivation in the prior art would have led one having ordinary skill in the art to combine the prior art references in order to arrive at the claimed invention. Anderson discloses a petroleum analytics learning system that leverages autoregressive models and decisions tree models as part of an ensemble learning method. Meek discloses a combined autoregressive tree model for time series forecasting and particularly touts the benefits of the model to be useful for data mining purposes because such models can be learned efficiently, support accurate predictions, and have easy interpretability ((Meek, Page 229, ¶1) "In this paper, our goal is to identify models for continuous-valued time-series data that are useful for data mining in that they (1) can be learned efficiently from data, (2) support accurate predictions, and (3) are easy to interpret."). As such and so to achieve the touted benefits of simultaneously being able to produce accurate predictions with interpretable results, it would have accordingly been obvious to use particularly the ART model disclosed by Meeks for forecasting time series data over the ensemble methods disclosed by Anderson.
The proposed combination does not teach; however the combination in view of Yajun teaches wherein the one or more predicted production values of the well include a predicted gas oil ratio (GOR);
wherein the further predicted production values of the well include further predicted GORs; Gas-oil ratio is predicted using a classification and regression tree approach ((Yajun, ¶52) "Step 8: Determine the control mode of the vacuum-assisted oil and gas recovery system: according to the ambient temperature, oil temperature and flow signal of the fuel dispenser as input data, predict the oil-gas ratio of the oil and gas recovery by inputting the data input test and adjusted CART model,").
Anderson and Yajun are analogous arts because they are both related to the same field of endeavor as using machine learning methods for optimizing oil well production. It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have combined the teachings of Yajun into the methodology taught by Anderson because some teaching, suggestion, or motivation in the prior art would have led one of ordinary skill to combine the prior are reference teachings to arrive at the claimed invention. Anderson teaches predicting oil production values using a machine learning approach to include gas volume and oil volume but does not explicitly teach predicting the gas oil ratio (GOR). Anderson suggests that a classification and regression tree (CART) machine learning model can be used for the predictions. Yajun teaches utilizing a CART machine learning model to predict the gas-oil ratio for adaptively controlling the oil and gas recovery process. Meek teaches the utilization of an autoregressive tree model for predicting time series data wherein the decision tree has linear regressions at its leaves. In applying the teachings of Yajun to predict the GOR into the method of Anderson (which predicts an initial value and subsequent values of production metrics ((Anderson, ¶84) "The integral transform then predicts the future production of the well under study before the oil and gas are delivered to the surface and uses future production to calculate an accuracy of that initial forecast.")) and further modified by Meek (which can be used for forecasting future time series data ((Meek, Page 236 ¶5- page 237 ¶1) "We now consider the problem of using ART models to forecast. Given a sequence of observations, the task of forecasting is to calculate the distributions for future observations in the sequence. We distinguish between two important types of forecasting: (1) one-step forecasting and (2) multi-step forecasting.")), the proposed combination of the methods would yield the prediction of a predicted GOR and further predicted GOR values. One having skill in the art would be motivated to modify the production prediction methodology taught by Anderson in view of Meek to include predicting the gas-oil ratio as taught by Yajun because the gas-oil ratio is an important metric to monitor in the oil-gas recovery process in order to ensure environmental protection and safety as well as to reduce energy usage and cost. ((Yajun, ¶73) "The oil and gas recovery method can realize the adaptive adjustment of the gas-liquid ratio, so that the oil-gas recovery ratio is in the range of 1.00-1.20, so as to achieve environmental protection, safety, energy saving and low cost.")
The proposed combination does not explicitly teach; however, the proposed combination in view of Kittilsen teaches alter a choke or a lift on the well based on the predicted GOR and the further predicted GORs generated using the decision-tree-based model. The GOR is described as being linked to pressure within the well. The pressure within the well is described as being controlled by a choke in response to the GOR. ((Kittilsen, Page 2, ¶6) "An increasing well stream GOR gives a lower liquid hold-up in the well, which in turn means less static pressure drop across the well. The wellhead pressure eventually increases. Normally, the pressure downstream the well head choke is controlled, and thus the pressure drop across the choke increases. This results in a higher gas rate that makes the well stream GOR increase. In total there is an unstable well reservoir system. Production trends from manual operation of a gas coning well are shown in the left part of Fig. 7. The plot illustrates how the gas and oil rates drift over time when not controlled"); ((Kittilsen, Page 7, ¶2) " The gas rate response to a step in choke position shows two distinct parts, as illustrated with typical responses in Fig. 4. The first part of the response is an almost immediate change in gas rate. This is followed by a second phase with a slowly changing gas rate due to the change in GOR and corresponding change in differential pressure across the choke, as discussed in the introduction. ")
Kittilsen is analogous art to the claimed invention because it is related to the optimization of oil well performance. It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have combined the teachings of Kittilsen into the methodology taught by the proposed combination of Anderson, Meek and Yajun because a simple substitution of one known element for another to obtain predictable results would yield the claimed invention. Anderson teaches a machine learning method for predicting oil well production values. Meek discloses the utilization of an autoregressive tree model for predicting and forecasting time series data. Yajun teaches a machine learning method for predicting gas oil ratio values for adaptive control of the oil well, wherein a vacuum pump of the system is controlled in response to the GOR to optimize gas and oil recovery . Kittilsen teaches automated control of a choke in response to GOR to improve process operation. In combining the predictive capabilities for oil applications of Anderson with the particular ART model disclosed by Meek with the GOR prediction and subsequent system control of Yajun with the choke adjustment in response to GOR taught by Kittilsen, one having ordinary skill would arrive at the claimed invention. It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have substituted the choke control of the system in response to GOR described by Kittilsen for the control of the pump for the vacuum pump in response to predicted GOR taught by Yajun to adaptively realize oil and gas recovery because Kittilsen sufficiently describes the link between choke adjustment and GOR such that results of the substitution would be predictable.
Regarding claim 2, the proposed combination in further view of Anderson teaches wherein the time-independent input feature values relate to lateral lengths of the corresponding wells, proppant pumped into the corresponding wells, fluid pumped into the corresponding wells, or porosity of rock in which the corresponding wells are respectively disposed. Attributes (features) of the data stored in the database are listed in Appendix 3 and include features related to Lateral lengths (Perforated Lateral Length), proppant pumped (Volume 100 Mesh Sand Proppant, Large Proppant Design), fluid pumped (Fluid Resistivity, Total Pump Output, Fluid Design, Fluid and Gas Composition), and porosity of rock (Porosity, Effective Porosity). (Anderson, Appendix 3)
Regarding claim 3, the proposed combination in further view of Anderson teaches wherein the time-dependent input feature values include inter-well spacing values, artificial lift, choke, pressure, or whether the corresponding wells are parent wells. Attributes (features) of the data stored in the database are listed in Appendix 3 and include features including inter-well spacing (Fracture Spacing, Cluster Spacing), artificial lift (total pump output, GPM measure for water), choke (choke position), pressure (standpipe pressure, avg pressure, breakdown pressure, breakdown pressure to avg pressure, initial pressure, max pressure, sample pressure, daily casing pressure, daily tubing pressure, differential pressure, pressure, separator pressures, pressure), and parent well determination (planned stages). (Anderson, Appendix 3)
Regarding claim 4, the proposed combination in further view of Anderson teaches wherein the ground-truth production values represent volumes of hydrocarbons or water extracted from the corresponding wells. ((Anderson, ¶64 and see Fig 4, item1340) "Within the SID 1300, in accordance with an exemplary embodiment of the claimed invention, production data 1340 include gas analysis, such as BTU calculation, depletion (Z) factor, sample pressure, sample temperature, molar component percent, GPM (gallons per minute) measure, production estimates, such as daily gas, daily condensate, daily water, daily casing pressure, daily tub pressure, daily pad volume, condensate haul tickets, water haul tickets, tank gauges-top, tank gauges-bottom, and SCADA (supervisory control and data acquisition), such as gas rate, differential pressure, tubing pressure, casing pressure, ESD ( emergency shutdown) alarms, separator pressures, choke position, LEL (lower explosive limit) readings, condensate density, water density, tank gauges-top, tank gauges bottom, EBU Data, flash separation data, VRU (vapor recovery unit) data, battery voltage data."). Anderson includes production values measured in volumes of gallons per minute of condensate (hydrocarbons) and water in Appendix 3. The data in Appendix 3 is a glossary of the attributes obtained in the well data which includes data measured in the field (ground truth). ((Anderson, ¶4) "The PALM also uses Support Vector Regression, logistic regression, Bayesian models, nearest neighbors, neural networks and deep learning networks uniquely combined as ensemble learning to weigh the importance of hundreds to thousands of geological, geophysical, and engineering attributes, both measured in the field and computed from theoretical analyses such as reservoir simulation models and 4D seismic and gravity gradiometry monitoring of production changes over time.")
Regarding claim 5, the proposed combination in further view of Anderson teaches wherein the ground-truth production values of the corresponding wells at respectively earlier points in time are data that form a basis of an order 1 autoregressive model. Anderson describes utilizing production data of historical wells as the time series data that is fed into a time-series classification step that utilizes a variant (ARCH/GARCH) of an autoregressive model. ((Anderson, ¶26) “The aforesaid real-time processor correlates stages of each class to an average highest production of historical wells."; (Anderson, ¶27) "In accordance with an exemplary embodiment of the claimed invention, the aforesaid real-time processor performs the automated time series classification by discovering sequential patterns and interactions among time series variables utilizing at least one of the following: an autoregressive integrated moving average (ARIMA) model, a multivariate time series analysis, a hidden Markov model, an autoregressive conditional heteroskedasticity (ARCH) model, an exponentially weighted moving average and a generalized autoregressive conditional heteroskedasticity (GARCH) model."). Anderson also discusses PALM subsystem MAPGATHER’s ability to achieve day-ahead forecasting and all subsystems have access to the same algorithms and tools (Anderson, ¶59) so this functionality is widespread for the PALM utility. With day-ahead forecasting, there is at least 1 day of data by which to base a model and therefore, 1 day of data could be the basis for the model of order 1. ((Anderson, ¶87) "Forecasting of day-ahead and week-ahead pipe line gathering system capacity by the MAPGATHER subsystem 1250 leads to the identification of maintenance that will prevent the need to shut-in wells because of excessive gathering system capacity.")
Regarding claim 6, the proposed combination in further view of Anderson teaches wherein the ground-truth production values of the corresponding wells at respectively earlier points in time are data that form a basis of an order 2 autoregressive model. Anderson describes utilizing production data of historical wells as the time series data that is fed into a time-series classification step that utilizes a variant (ARCH/GARCH) of an autoregressive model. ((Anderson, ¶26) “The aforesaid real-time processor correlates stages of each class to an average highest production of historical wells."; (Anderson, ¶27) "In accordance with an exemplary embodiment of the claimed invention, the aforesaid real-time processor performs the automated time series classification by discovering sequential patterns and interactions among time series variables utilizing at least one of the following: an autoregressive integrated moving average (ARIMA) model, a multivariate time series analysis, a hidden Markov model, an autoregressive conditional heteroskedasticity (ARCH) model, an exponentially weighted moving average and a generalized autoregressive conditional heteroskedasticity (GARCH) model."). Anderson also discusses PALM subsystem MAPGATHER’s ability to achieve day-ahead and week-ahead forecasting and all subsystems have access to the same algorithms and tools (Anderson, ¶59) so this functionality is widespread for the PALM utility. With day-ahead and week-ahead forecasting, there is at least 1 day of data by which to base a model and up to 7 days (week) by which to base a model. Therefore, it is obvious that 2 days’ worth of data could be the basis for the model of order 2. ((Anderson, ¶87) “Forecasting of day-ahead and week-ahead pipe line gathering system capacity by the MAPGATHER subsystem 1250 leads to the identification of maintenance that will prevent the need to shut-in wells because of excessive gathering system capacity.")
Regarding claim 8¸ the proposed combination in further view of Anderson teaches wherein the particular points in time are regular intervals. Data is generated every day, indicating regular intervals of time for collecting data points, where each data point is associated with a time. ((Anderson, ¶60) "In the petroleum industry, terabytes of data are generated every day, such as time series hydraulic fracture data, well log and measurement-while-drilling data, and sensor data that monitors production and delivery to processing plants.")
Regarding claim 10, the proposed combination in further view of Anderson teaches wherein the time-dependent input feature values include anomalous events. When read in light of the instant specification, anomalous events cover any sort of random, one-time, or unexplained event that impacts well production. ((Instant Specification, ¶193) "These anomalous events may be called "frac hits" for purposes of convenience, but cover any sort of random, one-time, or unexplained event that impacts well production. Thus, anomalous events may include pumping or stimulation of a nearby well, adverse weather, human error, and so on.") Collected data, which includes time-dependent feature values contain extraneous and noisy data. ((Anderson, ¶5) "The collected data are 'cleaned' to eliminate extraneous and noisy data. The cleaned data are normalized and stored."). Therefore, a person having ordinary skill in the art would understand extraneous and noisy data to be included per the definition of the specification for anomalous events.
Regarding claim 11, the proposed combination in further view of Anderson teaches receive indications of one or more of the anomalous events to remove from the time- dependent input feature values; Extraneous and noisy data of data that includes time-dependent feature values are anomalous events indicated by the outlier nature of the data. Such indications motivate cleaning of the data to remove the outliers. ((Anderson, ¶5) “The collected data are 'cleaned' to eliminate extraneous and noisy data.")
generate a variation of the time-dependent input feature values with the one or more of the anomalous events removed; The data is cleaned, where the anomalous events are removed, and stored as a modified version of the original data for further use. ((Anderson, ¶5) “The cleaned data are normalized and stored.")
provide, to the decision-tree-based model, the new time-independent input feature values and the variation of the time-dependent input feature values for the well; and A cleaned and normalized variation of the collected data is generated. ((Anderson, ¶5) "The collected data (new data from the database) are 'cleaned' to eliminate extraneous and noisy data. The cleaned data are normalized and stored."). The data is further processed by the PALM tool, which is a modeling tool which includes the use of decision trees. ((Anderson, ¶5) “The normalized data are processed to determine clusters of correlation in multi-dimensional space to identify a machine learned ranking of Importance Weights for each attribute."; (Anderson, ¶13) “The PALM utilizes at one of the following classification: logistic regression, support vector machine and support vector regression, nearest neighbors, decision trees and random forest, neural networks and deep learning networks.")
receive, from the decision-tree-based model, one or more counterfactual predicted production values of the well, wherein the one or more counterfactual predicted production values are generated by the decision-tree-based model based on the internal structure of the decision-tree-based model, the new time-independent input feature values, and the variation of the time-dependent input feature values. Forecasted (predicted) production values for wells are produced by the MAPPROD subsystem of PALM, where cleaned (Anderson, ¶5) time-independent and time-dependent feature values stored in the system integrated database are used for the prediction. ((Anderson, ¶16) “In accordance with an exemplary embodiment of the claimed invention, the aforesaid system and method queries one or more system integration databases of multiple surrounding wells in an area or querying one integrated master system integration database comprising regionally relevant geologic and geophysical data, reservoir models, drilling data, hydraulic fracturing data, the historical exogenous data, the real-time exogenous data, and the real-time endogenous data to forecast production of said each well."; (Anderson, ¶56) “In accordance with an exemplary embodiment of the claimed invention, the MAPPROD 1240 is a production forecaster that convolves the actual attribute values of hundreds to thousands of attributes coming into the system from historical wells, as well as each new well as it progresses, to maximize production for all wells in a play”) The MAPPROD subsystem (1240) produces the predicted production results and is capable of leveraging all machine learning algorithms and tools included in PALM, which include decision tree modeling amongst others in the classification prescriptor (1421). Because outputs of models are dependent upon the structure of the models, it would be understood by a person having ordinary skill in the art that the output generated production value is based on the internal structure of the model used. ((Anderson, ¶59) See Fig. 3; "Turning now to FIG. 3, in accordance with an exemplary embodiment, there are listed machine analytics algorithms and tools commonly accessible to all MAP 1200 subsystems within the PALM 1000.")
Regarding claim 14, the claim limitations A computer-implemented method comprising:
obtaining, from persistent storage, training data related to well production, wherein entries in the training data respectively include time-independent input feature values and time- dependent input feature values both mapped to ground-truth production values of corresponding wells at particular points in time, wherein the time-dependent input feature values include ground-truth production values of the corresponding wells at respectively earlier points in time that form the basis of an autoregressive model;
training a decision-tree-based model with the training data including the basis of the autoregressive model;
providing, to the decision-tree-based model, new time-independent input feature values and new time-dependent input feature values for a well;
receiving, from the decision-tree-based model, one or more predicted production values of the well, wherein the one or more predicted production values are generated by the decision- tree-based model based on an internal structure of the decision-tree-based model, the new time-independent input feature values, and the new time-dependent input feature values, wherein the one or more predicted production values of the well include a predicted gas oil ratio (GOR);
receiving, from the decision-tree-based model, further predicted production values of the well, wherein each further predicted production value of the further predicted production values is recursively generated by the decision-tree-based model based on the internal structure of the decision-tree-based model using (i) ground-truth initial production values of the well, (ii) a set of the one or more predicted production values of the well, and (iii) a set of prior further predicted production values, wherein the further predicted production values of the well include further predicted GORs;
writing, to the persistent storage, the one or more predicted production values and the further predicted production values; and
altering a choke or a lift on the well based on the predicted GOR and the further predicted GORs generated using the decision-tree-based model. are substantially similar to that of claim 1 and so claim 14 is therefore rejected under the same rationale as provided for claim 1. Claim 1 is a system claim and claim 14 is a method claim. Anderson discloses a method and the other relied-on references disclose the remaining limitations as provided in the rejection of claim 1. ((Anderson, Page 14, Claim 1) "A method for optimizing exploration, production and gathering from at least one well of oil and natural gas fields using a petroleum analytics learning machine system to maximize production while minimizing costs, comprising the steps of:")
Regarding claim 15, the limitations wherein the time-independent input feature values relate to lateral lengths of the corresponding wells, proppant pumped into the corresponding wells, fluid pumped into the corresponding wells, or porosity of rock in which the corresponding wells are respectively disposed are substantially similar to that of claim 2 and so claim 15 is therefore rejected under the same rationale as provided for claim 2.
Regarding claim 16, the limitations wherein the time-dependent input feature values include inter-well spacing values, artificial lift, choke, pressure, or whether the corresponding wells are parent wells are substantially similar to that of claim 3 and so claim 16 is therefore rejected under the same rationale as provided for claim 3.
Regarding claim 17, the limitations wherein the ground-truth production values represent volumes of hydrocarbons or water extracted from the corresponding wells are substantially similar to that of claim 4 and so claim 17 is therefore rejected under the same rationale as provided for claim 4.
Regarding claim 19, the limitations wherein the time-dependent input feature values include anomalous events, the method further comprising:
receiving indications of one or more of the anomalous events to remove from the time- dependent input feature values;
generating a variation of the time-dependent input feature values with the one or more of the anomalous events removed;
providing, to the decision-tree-based model, the new time-independent input feature values and the variation of the time-dependent input feature values for the well; and
receiving, from the decision-tree-based model, one or more counterfactual predicted production values of the well, wherein the one or more counterfactual predicted production values are generated by the decision-tree-based model based on the internal structure of the decision-tree-based model, the new time-independent input feature values, and the variation of the time-dependent input feature values are substantially similar to the combination of claims 10 and 11 and so claim 19 is therefore rejected under the same rationale as provided for claims 10 and 11.
Regarding claim 20, the claim limitations An article of manufacture including a non-transitory computer-readable medium, having stored thereon program instructions that, upon execution by a computing system, cause the computing system to perform operations comprising:
obtaining, from persistent storage, training data related to well production, wherein entries in the training data respectively include time-independent input feature values and time- dependent input feature values both mapped to ground-truth production values of corresponding wells at particular points in time, wherein the time-dependent input feature values include ground-truth production values of the corresponding wells at respectively earlier points in time that form the basis of an autoregressive model;
training a decision-tree-based model with the training data including the basis of the autoregressive model;
providing, to the decision-tree-based model, new time-independent input feature values and new time-dependent input feature values for a well;
receiving, from the decision-tree-based model, one or more predicted production values of the well, wherein the one or more predicted production values are generated by the decision- tree-based model based on an internal structure of the decision-tree-based model, the new time-independent input feature values, and the new time-dependent input feature values, wherein the one or more predicted production values of the well include a predicted gas oil ratio (GOR);
receiving, from the decision-tree-based model, further predicted production values of the well, wherein each further predicted production value of the further predicted production values is recursively generated by the decision-tree-based model based on the internal structure of the decision-tree-based model using (i) ground-truth initial production values of the well, (ii) a set of the one or more predicted production values of the well, and (iii) a set of prior further predicted production values, wherein the further predicted production values of the well include further predicted GORs;
writing, to the persistent storage, the one or more predicted production values and the further predicted production values; and
altering a choke or a lift on the well based on the predicted GOR and the
further predicted GORs generated using the decision-tree-based model. are substantially similar to that of claim 1 and so claim 20 is therefore rejected under the same rationale as provided for claim 1. Claim 1 is a system claim and claim 20 is a manufacture claim. Anderson discloses computer executable instructions stored on memory and executed by a real-time processor and the other relied-on references disclose the remaining limitations as provided in the rejection of claim 1. ((Anderson, ¶31) " In accordance with an exemplary embodiment of the claimed invention, the aforesaid real-time processor comprises a memory to store computer-executable instructions. The aforesaid real-time processor is coupled to at least one transmitter to communicate with the hydraulic fracturing control center via a bi-directional messaging interface. The aforesaid real-time processor executes the computer-executable instructions to cause the hydraulic fracturing control center (or Frac control center) to perform multiple actions.")
Regarding claim 21, the proposed combination discloses The method of claim 1, as disclosed previously.
The proposed combination in further view of Kittilsen discloses wherein the choke or the lift on the well is altered to achieve a target production on the well. Automatic choke control is described as being beneficial for production and injection wells, so as to keep key parameters at the optimal values. Process fluctuations ((Kittilsen, Page 1, ¶1-2) " Better operation is achieved by reduced process fluctuations so that the average production may be closer to the constraints. This is achieved by automatic process control which also reduces operator load. Wells which produce oil and free gas (coning) from the reservoir are sensitive to changes in operating conditions, and the gas and oil rates drift over time. This is natural because variations in coned free gas generate variations in wellhead pressure. This will in turn affect the production rate and the well drawdown. In manual operation, a common strategy is to operate with an excess of free gas or use artificial gas lift and manipulate the wellhead choke manually. An automatic wellhead choke control solution is developed to reduce fluctuations and increase production rates."); ((Kittilsen, Page 2, ¶2) "Processing constraints usually imply production loss due to fluctuations and the need for operational margins to the constraints. Increased production is achieved by stabilizing the process by automatic control and moving the average closer to constraints as illustrated in Fig. 1."); See also Figure 1 which depicts a fluctuating production value with a target, wherein automatic control is used to reduce process fluctuations so as to achieve the target.
Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over the proposed combination as applied to claims 8 and 1 above, and further in view of Rebeschini et Al. (Building Neural-Network-Based Models Using Nodal and Time-Series Analysis for Short-Term Production Forecasting, October 28, 2013, Paper presented at the SPE Middle East Intelligent Energy Conference and Exhibition), hereinafter referred to as Rebeschini.
Regarding claim 9, Anderson is further relied upon to teach (except the limitations surrounded by brackets) [[wherein the]] regular intervals [[are 5 days, 10 days, 15 days, 60 days, or 90 days]] Data is generated every day, indicating regular intervals of time for collecting data points, where each data point is associated with a time. ((Anderson, ¶60) "In the petroleum industry, terabytes of data are generated every day, such as time series hydraulic fracture data, well log and measurement-while-drilling data, and sensor data that monitors production and delivery to processing plants.")
The proposed combination in further view of Anderson does not teach, however, Rebeschini teaches wherein the … are 5 days, 10 days, 15 days, 60 days, or 90 days. A 90-day historical data is used as the basis for a neural network model for production forecasting. ((Rebeschini, Page 2, ¶1) "The first part used historical data from the previous 90 days to generate a NN model comparable to actual results.")
Anderson and Rebeschini are related to the same field of endeavor of predictive modeling for oil well production. It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have implemented the 90-day interval of time taught by Rebeschini as the period of time for the regular intervals taught by the proposed combination including the teachings of Anderson because 90 days of well production history has low variability and is a suitable time frame for predicting short-term production predictions. ((Rebeschini, Page 3, ¶2) "Considering that the 90-day well production history likely has low variability and does not incorporate many operational events or changes in well completions, the nodal analysis technique was implemented for individual wells."). Achieving short term predictions has the predictable results of providing insight for day-to-day changes in production flow, thereby being effective for predictive maintenance applications. ((Rebeschini, Page 1, ¶5) “As indicated in Al-Jasmi et al. (2013b), short-term predictions are useful for accommodating day-to-day changes in production flow.").
Claim 12 is rejected under 35 U.S.C. 103 as being unpatentable over the proposed combination as applied to claims 10 and 1 above, and further in view of Vargas et Al. (A realistic and public dataset with rare undesirable real events in oil wells, October 2019, Journal of Petroleum Science and Engineering), hereinafter referred to as Vargas.
Regarding claim 12, Anderson is further relied upon to teach (except the limitations surrounded by brackets) [[receive indications of one or more hypothetical anomalous events to add to the time- dependent input feature values;
generate a variation of the time-dependent input feature values with the one or more of the hypothetical anomalous events;]]
provide, to the decision-tree-based model, the new time-independent input feature values [[and the variation of the new time-dependent input feature values for the well;]] The labeled data from within the System Integrated Database (SID) is processed by the PALM tool, which is a modeling tool which includes the use of decision trees. ((Anderson, ¶5) “The normalized data are processed to determine clusters of correlation in multi-dimensional space to identify a machine learned ranking of Importance Weights for each attribute."; (Anderson, ¶13) "The PALM utilizes at one of the following classification: logistic regression, support vector machine and support vector regression, nearest neighbors, decision trees and random forest, neural networks and deep learning networks.").
and receive, from the decision-tree-based model, one or more [[counterfactual]] predicted production values of the well, wherein the one or more [[counterfactual]] predicted production values are generated by the decision-tree-based model based on its internal structure, the new time-independent input feature values, [[and the variation of the new time-dependent input feature values.]] Forecasted (predicted) production values for wells are produced by the MAPPROD subsystem of PALM, where time-independent and time-dependent feature values stored in the system integrated database are used for the prediction. ((Anderson, ¶16) "In accordance with an exemplary embodiment of the claimed invention, the aforesaid system and method queries one or more system integration databases of multiple surrounding wells in an area or querying one integrated master system integration database comprising regionally relevant geologic and geophysical data, reservoir models, drilling data, hydraulic fracturing data, the historical exogenous data, the real-time exogenous data, and the real-time endogenous data to forecast production of said each well."; (Anderson, ¶56) “In accordance with an exemplary embodiment of the claimed invention, the MAPPROD 1240 is a production forecaster that convolves the actual attribute values of hundreds to thousands of attributes coming into the system from historical wells, as well as each new well as it progresses, to maximize production for all wells in a play”) The MAPPROD subsystem (1240) produces the predicted production results and is capable of leveraging all machine learning algorithms and tools included in PALM, which include decision tree modeling amongst others in the classification prescriptor (1421). Because outputs of models are dependent upon the structure of the models, it would be understood by a person having ordinary skill in the art that the output generated production value is based on the internal structure of the model used. ((Anderson, ¶59) See Fig. 3; "Turning now to FIG. 3, in accordance with an exemplary embodiment, there are listed machine analytics algorithms and tools commonly accessible to all MAP 1200 subsystems within the PALM 1000.")
The proposed combination in further view of Anderson does not teach; however, Vargas teaches receive indications of one or more hypothetical anomalous events to add to the time- dependent input feature values; The man-made dataset taught by Vargas includes indications of hypothetical anomalous events that are distinguished from normal conditions. ((Vargas, page 7, section 4.2, ¶1) " This benchmark intends to encourage the development, evaluation, and comparison of anomaly detecting algorithms. In this task, undesirable events (anomalies) must be distinguished from the normal condition.") The dataset contains the indication of the anomalous event as a “type of event”. ((Vargas, Table 2)). Vargas proposes the use of this dataset in conjunction with (as an addition to) existing datasets which may be sparse. ((Vargas, page 8, section 5 ¶1) " The events are characterized by eight process variables, and the resulting dataset can be readily used as a benchmark for the development of machine learning techniques related to inherent difficulties of actual data. This resource can also be explored in tasks associated with detecting and diagnosing undesirable events in such wells. "). Time-dependent data is used (See Vargas page 5, section 2.5 discussing Multivariate Time Series data) and the data includes feature values, as is indicated by the suggestion that feature engineering is a proposed application of the dataset ((Vargas, page 2, ¶10) "This work gave rise to two relevant contributions. The first one is the 3W dataset, which has been made available in the supporting repository (Vargas et al., 2019) for this paper and can be readily used as a benchmark dataset for development of machine learning techniques related to inherent difficulties of actual data. An enormous number of possibilities in terms of, for example, preprocessing (normalization, NaN values, missing values, frozen variables, outliers, etc.), filters (smoothing, resampling, etc.), transformations (multiscale, wavelet, etc.), family of classifiers (based on trees, artificial neural networks, distances, ensembles, etc.), hyperparameter optimization, feature engineering, and performance metrics can be investigated using this dataset. ")
generate a variation of the time-dependent input feature values with the one or more of the hypothetical anomalous events; Vargas teaches using the man-made dataset with other published datasets, thus implying the application of generating an expanded dataset. ((Vargas, page 8, section 5 ¶3) " Specific benchmarks that practitioners and researchers can use together with the published dataset are defined. Along with the proposed dataset, these challenges are expected to be a significant motivation to the community of engineers and scientist who develop machine learning and data analytics methods for the oil and gas field. ")
…and the variation of the new time-dependent input feature values for the well; As stated previously, Vargas proposes the use of this dataset in conjunction with (as an addition to) existing datasets which may be sparse (Vargas, page 8, section 5 ¶1) and teaches using the man-made dataset with other published datasets, thus implying the application of generating an expanded dataset (Vargas, page 8, section 5 ¶3).
…counterfactual…Because the dataset taught by Vargas is man-made to include real and simulated data, predictions based on this dataset would be counterfactual in nature. ((Vargas, page 6 ¶2) "Each instance, whether real, simulated, or hand-drawn, was saved in a standardized and dedicated Comma-Separated Values (CSV) file.")
…and the variation of the new time-dependent input feature values…. (Vargas, page 8, section 5 ¶1 and ¶3)
Vargas is reasonably pertinent to the problem faced by the inventor. Vargas specifically targets naturally flowing wells but suggests that that the dataset can similarly be applied to wells that leverage artificial lift techniques such as that described in Anderson. ((Vargas, page 3, section 2.1 ¶3) “It is important to note that naturally flowing wells can also be equipped to be operated with artificial lift methods under certain circumstances. That is, it is not uncommon for a well to be operated in an intercalated fashion between an artificial lift technique and the natural method"). Therefore, the teachings within the arts indicate analogous subject matter. It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have incorporated the synthetic dataset that includes anomalous behavior taught by Vargas into the modeling and prediction system taught by the proposed combination that includes the teachings of Anderson because Anderson already indicates the usage of reservoir modeling simulation data as part of the comprehensive System Integrated Dataset that the PALM modeling utility uses. ((Anderson, ¶5) "Structured digital data and unstructured textual data from geological, geophysical, reservoir modeling simulation, drilling, hydraulic fracturing and completion, and production of crude oil, natural gas, ethane, butane, propane and condensates are collected."). Furthermore, the proposed combination particularly in view of Meek suggests that forecasting using the ART as in Meek provides a mechanism by which to interpret results (so as to be able to visualize the causal components of the model for the results). Including a simulated dataset that includes anomalous events for modeling purposes yields the improvement of having a comprehensive view of oil well behavior in order to provide more accurate and hypothetical predictions of undesirable events, especially if historical data does not already contain such events. Additionally, accurate prediction of undesirable events enables potential to achieve higher production by mitigating or addressing anomalies before occurrence, thereby supporting the objective of optimizing production ((Vargas, page 2 ¶4) "In oil and gas wells, AEM can help prevent production losses, environmental accidents, and human casualties and reduce maintenance costs."). Further explicit motivation to combine the teachings of both arts is the Vargas suggests utilizing the publicly-available created dataset in applications that target machine learning and data analytics methods for the oil and gas field, such as taught by Anderson ((Vargas, page 8, section 5 ¶3) "Specific benchmarks that practitioners and researchers can use together with the published dataset are defined. Along with the proposed dataset, these challenges are expected to be a significant motivation to the community of engineers and scientist who develop machine learning and data analytics methods for the oil and gas field")
Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over the proposed combination as applied to claim 1 above, and further in view of Kaggle (Kaggle, “Tutorial: Machine Learning Interpretability”, April 22, 2019, www.kaggle.com/code/datacog314/tutorial-machine-learning-interpretability), hereinafter referred to as Kaggle.
Regarding claim 13, the proposed combination in further view of Anderson teaches determine Shapley data for the time-independent input feature values and the time- dependent input feature values, wherein the Shapley data indicate correlations between: (i) each of the time-independent input feature values and the time-dependent input feature values, and (ii) the ground-truth production values of the corresponding wells; When read in light of the specification, Shapley data assigns a contribution to each of the input features to quantify how much the input feature contributes to the output. (Instant Specification, ¶203-204). Anderson describes computing Importance Weights via the MAPGEORES subsystem of PALM to indicate correlations between the attributes of the dataset (that include time-dependent and time-independent variables) and the production value. ((Anderson, ¶68) "The MAPGEORES 1210 utilizes machine learning of the historical structured data to compute Importance Weights for the attributes that represent all the data available before spud. The machine learning algorithms of the MAPGEORES 1210 uniquely combine the parameters of support vector and linear regression, allowing the construction of the Tornado diagrams, as exemplary shown in FIG. 5, to represent the Importance Weights of each attribute that correlates with a positive production prediction result (the bars to the right) and the importance of negative weights of each attribute that correlates with a positive production prediction result (the bars to the left). The predicted production is then compared to the actual production to derive an accuracy score."). Because Shapley data and the Importance Weights achieve the same end goal of identifying correlations for attributes/features to production output, one having ordinary skill in the art would understand this to be equivalent or interchangeable ways to quantify this relationship.
generate a representation of a graphical user interface for the Shapley data, wherein each of the time-independent input feature values and the time-dependent input feature values is associated with [[a scatterplot of]] the respective Shapley data for the corresponding wells; Importance weights of attributes of the time-independent and time-dependent feature values enable generation of a tornado diagram for presentation to the user. ((Anderson, ¶18) "The attributes comprise: relevant geological and geophysical data; reservoir modeling results and calculations, including correction factors and assumptions; rock property measurements including poisons ratio, young's module, gamma ray radioactivity, organic and British Thermal Unit (BTU) content; and combining parameters of the support vector regression and linear regression to enable construction of tornado diagrams representing visually the importance weights of each attribute that correlates with a positive production prediction result and the importance weights of each attribute that correlates with a negative production prediction result for all wells in the area or play.")
and transmit, to a client device, the representation of the graphical user interface. Data can be forwarded (transmitted) to the TOTALVU user interface. ((Anderson, ¶4) “These subsystems use the PALM System Integration Database (SID) to retrieve integrated data, then perform machine learning and other statistical analyses of that data, and return to the SID results of computation and predictive and prescriptive actions that can be forwarded by the TOTALVU user interface (UI) to controllers, human and/or automated, so that real-time optimization of production and minimization of costs can be realized for new wells "). Figure 5 shows a tornado diagram that depicts the importance weight values on the TOTALVU dashboard. ((Anderson, ¶67 & ¶42) “The tornado diagram of importance weights calculated by MAPGEORES 1210 as exemplary displayed by the TotalVU 1500 is shown in FIG. 5.")
The proposed combination in further view of Anderson does not explicitly disclose; however the proposed combination in view of Kaggle discloses a scatterplot of Shapley data.
PNG
media_image2.png
930
1902
media_image2.png
Greyscale
Kaggle is analogous to the claimed invention because it is related to the same field of endeavor of using predictive models for understanding data. It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have implemented particularly the scatterplot representation of Shapley data into the proposed combination because some teaching, suggestion, or motivation in the prior art would have led one having ordinary skill in the art to do so in order to arrive at the claimed invention. The proposed combination particularly in view of Anderson discloses the utilization of a tornado diagram to visualize the importance weights. Kaggle provides a tornado diagram representation of Shapley data which leverages scatter plotting to visualize the data. Because Anderson suggests a tornado diagram and Kaggle provides a particular means by which to generate such a diagram which includes scattered points, it would have accordingly been obvious to one having skill to make the combination to arrive at the claimed invention.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to EMILY GORMAN LEATHERS whose telephone number is (571)272-1880. The examiner can normally be reached Monday-Friday, 9:00 am-5:00 pm ET.
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, EMERSON PUENTE can be reached at (571) 272-3652. 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.
/E.G.L./Examiner, Art Unit 2187
/EMERSON C PUENTE/Supervisory Patent Examiner, Art Unit 2187
Prosecution Insights