DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
Applicant’s arguments received 23 December 2025, have been fully considered. Claims 1, 7-11, and 15-20 are pending. Claims 1, 8, 15, 21, and 22 are amended.
Applicant’s efforts to amend the claims to address the claim objections are satisfactory, therefore all objections to the claim language are dropped.
Applicant’s arguments concerning the prior art rejections under 35 USC 103 have been considered and will be addressed below.
Applicant argues that Ibemere and Chen would not have motivated a person of ordinary skill in the art to input geochemical data that comprises abundances of the hydrocarbon molecules presented on multi-dimensional plots into the trained machine learning network. The examiner agrees that neither Ibemere nor Chen teach inputting a multi-dimensional plot into an ML model, however the examiner disputes that the claim language must be interpreted that way or that the specification has support for such an interpretation. See 112(b) rejection below.
Applicant argues that, because Chen uses PCA to reduce the dimensionality of raw datasets, one of ordinary skill would not have been motivated by Chen to perform the limitations of the independent claims. The examiner disagrees; one is free to define the PCA and ML processes of Chen’s workflow as a single ML algorithm. In this case, raw data would be fed into the ML algorithm. Furthermore, note that Applicant’s specification discusses using PCA as a form of regularization while training a CNN (¶70), and it would be reasonable from this to consider that PCA may be performed on inputs to the trained model as well.
Applicant argues that Chen inputs uncorrelated features into the ML in contrast to the claimed invention, and that Chen teaches away from inputting raw datasets into the ML algorithms. This argument seems to depend on the independent claims requiring one to input a visual graphic such as the “correlation star diagram” (see Figs. 5A-5B and ¶45 of specification) into the trained ML (the word “correlation” is used four times: once in the title and thrice as “correlation star diagram”). As previously argued, such an interpretation is unnecessary. Even assuming that interpretation, Chen teaches that rather than inputting the principal components themselves into ML models, the highest correlated original features are used (pg. 3, column 2, paragraph above Section 2.3: “The loading matrix in PCA provided the correlations between the original features and new principal components (PC). For better interpretations, feature selection was proceeded by choosing the highest correlated original features.”).
Regarding claim 22, Applicant argues that the prior art of record does not disclose or render obvious the amended feature of “separating the training geochemical data and the training geological data from a database using an artificial intelligence algorithm.” This amendment finds support from ¶73 of Applicant’s specification (¶73: “An AI algorithm (904) may access the database (902) to categorize the geochemical data (906) and geological data (908) that is stored in the database (902) among other data. In some embodiments, the AI algorithm (904) may separate the geochemical data (906) and geological data (908) from other data in the database (902). In some embodiments, the AI algorithm (904) may further separate the geological data (908) by origin data, depositional environment data, organofacies data, oil family data, and/or alteration mechanism data.”).
Considering the application of references to those claims which claim 22 depends from, the above language requires only that the input and output data be separated from a database using AI. The examiner considers that such a teaching would have been obvious in light of Chen. In Chen, the dataset drawn from includes both input and output data (pg. 7, column 1, under Section 3.1: “The experimental data used in this study came from the work published by Song et al. … Briefly, aliquot chemically dispersed oil, and non-dispersed oil samples were generated and weathered from day 1 to day 60 in a simulated marine environment. Five types of biomarkers (terpanes, steranes, triaromatic steranes or TA-steranes, monoaromatic steranes or MA-steranes, and diamantanes) were detected and analyzed to differentiate dispersed oil and non-dispersed oil. The samples were analyzed for each biomarker in 10 days intervals, through day 1 to day 60 weathering process.”). It would have been obvious for Chen to have, in the step of drawing input and output data from the database for training, have performed some separation step to distinguish data meant to be input into a ML model and data meant to be compared to the ML model’s output. Doing so would be useful to ensure output data is not treated as input data and vice versa.
See 103 rejections below.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1, 7-11, and 15-24 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 and 15 recite that the “geochemical data comprises abundances of the hydrocarbon molecules presented on multi-dimensional plots” then continues to recite that “the geochemical data [is input] into a trained machine learning network.” It is unclear from the claim language whether the claim requires the trained machine learning network to accept as input the plots themselves, or only the hydrocarbon molecule abundance data, where the data is capable of being presented on multi-dimensional plots. The former interpretation does not seem to have support in the specification and would therefore warrant a rejection under 35 U.S.C. 112(a). For examination purposes, the examiner will assume that the latter interpretation.
Claims 7-11 and 16-24 depend from claims 2 and 15, respectively, therefore they inherit the same issues and are rejected for the same reasons.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
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.
Claims 1, 7-11, and 15-24 are rejected under 35 U.S.C. 103 as being unpatentable over Ibemere ("Crude Oil Properties Elucidation Using Fingerprinting Technique") in view of Chen ( “A data-driven binary-classification framework for oil fingerprinting analysis”) and Knight (US 20170139078 A1) and Walters (“The Origin of Petroleum”).
Regarding claim 1, Ibemere discloses a method, comprising:
obtaining an oil sample from a region of interest (Abstract, crude oil samples are taken from the Niger Delta), wherein the oil sample comprises hydrocarbon molecules (crude oil comprises hydrocarbon molecules);
for each oil sample:
determining geochemical data for the oil sample using gas chromatography (Abstract, ¶2: “analysis was done with Gas Chromatographic instrument equipped with flame ionization detector (GC-FID) and HP-PONA capillary column”), and
wherein the geochemical data comprises abundances of the hydrocarbon molecules presented on multi-dimensional plots (see ex. Fig. 2, showing a chromatogram of an oil sample which comprises abundances of hydrocarbon molecules; see also Fig. 4 showing abundance ratios on a star plot); and
predicting geological data for the oil sample (pg. 5, Ph/Phy Ratio: “This ratio evaluates the depositional environment (Palo-environment), at which the initial hydrocarbon was deposited during formation and source origin information. This depositional environment is usually described as oxic (acidic) or anoxic (basic). When Pr/Phy values is above unitary (1), it indicates oxic environment, while values below 1 indicate anoxic.”),
wherein the geological data comprises at least one of origin data, depositional environment data, or organofacies data (paleoenvironment is depositional environment).
Ibemere does not explicitly disclose obtaining an oil sample from each of two or more wells within a subterranean region of interest. However, Ibemere says that the samples are of crude oil, and crude oil typically is taken from wells which bring up oil from underground. Furthermore, Ibemere goes on to state differences in the oil samples such as varying paleoenvironment, therefore it would be reasonable to assume that the oil samples are not all taken from the same well. Therefore it would be reasonable to assume that the oil samples were taken from each of two or more wells within a subterranean region of interest (i.e. from underground locations in the Niger Delta region).
Ibemere does not explicitly disclose predicting geological data for the oil sample by inputting the geochemical data into a trained machine learning network.
Chen discloses a method of training machine learning (ML) to aid fingerprinting using geochemical data of oil samples (pg. 3, below Table 2: “This study introduces ML as a new analysis tool by proposing a new binary classification framework to aid source identification in oil fingerprinting by distinguishing weathered crude oil (WCO) and CDO…The total 862 diagnostic ratios based on five types of biomarkers (terpanes, steranes, triaromatic steranes or TA-steranes, monoaromatic steranes or MA-steranes, and diamantanes) are chosen…as the features for ML.” pg. 3, column 2, Section 2.2: PCA is applied to the raw datasets and features were selected to be input into ML algorithms; the PCA and ML process can be considered a single ML algorithm.). Chen teaches that ML is a practical tool for determining the source of oil samples (pg. 11, Conclusion, ¶2: “This research…proved the practical value of adopting ML to facilitate oil fingerprinting”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Chen with the invention of Ibemere by predicting geological data for the oil sample by inputting the geochemical data into a trained machine learning network. Doing so would enable one to determine depositional environment of a new oil sample using ML, which would provide speed, accuracy, and automation to investigative techniques for oil exploration.
Ibemere in view of Chen does not explicitly disclose:
generating a geological map of the subterranean region of interest using the geological data for the oil sample from each of the two or more wells,
wherein the geological map separates the at least one of the origin data, the depositional environment data, or the organofacies data among the subterranean region of interest;
determining an oil field management plan using the geological map comprising at least one of determining a casing or determining if hydraulic fractures should be induced; and
completing at least one of the two or more wells based on the oil field management plan comprising at least one of installing the casing within the at least one of the two or more wells or inducing the hydraulic fractures in rock surrounding the at least one of the two or more wells.
Knight teaches that a geological map is a useful tool for hydrocarbon production (¶467: “A high-resolution subsurface geologic map of a region is typically a useful tool when determining the value and method of production of a well.”).
Furthermore, Walters teaches that anoxic environments are generally better than oxic environments for preserving oil-prone organic matter (pg. 85, Fig. 4 and paragraph above Fig. 4).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Knight and Walters with the invention of Ibemere in view of Chen by generating a geological map of the subterranean region of interest using the geological data for the oil sample from each of the two or more wells,
wherein the geological map separates the at least one of the origin data, the depositional environment data, or the organofacies data among the subterranean region of interest;
determining an oil field management plan using the geological map comprising at least one of determining a casing or determining if hydraulic fractures should be induced; and
completing at least one of the two or more wells based on the oil field management plan comprising at least one of installing the casing within the at least one of the two or more wells or inducing the hydraulic fractures in rock surrounding the at least one of the two or more wells.
Generating a geological map of the subterranean region of interest displaying oxic vs. anoxic paleoenvironment would be a useful aid in determining which locations have higher potential for oil production. Creating an oil field management plan comprising at least one of determining a casing or if hydraulic fractures should be induced would be useful to develop a definite plan about how oil should be extracted from a region of interest. Finally, the step of completing a well or inducing fractures would be useful because carrying out the management plan would enable one to actually extract oil.
Regarding claim 15, many of the limitations of claim 15 are found in claim 1 and are rejected for the same reasons. Claim 15 also recites a gas chromatography system which determines geochemical data and a computer processor configured to perform the method of claim 1; the gas chromatography system is disclosed by Ibemere as seen in the rejection of claim 1, and it would have been obvious to include a computer processor in order to perform the method of claim 1 autonomously (furthermore the use of ML implies that at least part of the method is performed with a computer processor).
Regarding claim 7, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 1. Ibemere further discloses that the geochemical data comprise ratios of the abundances of the hydrocarbon molecules (Fig. 4 depicts ratios of the abundances of hydrocarbon molecules, such as N-C29/N-C30).
Regarding claim 8, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 1. Ibemere further discloses that the multi-dimensional plots comprise star diagrams (Fig. 4 is a star diagram).
Regarding claim 9, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 1. Ibemere further discloses that the hydrocarbon molecules range from light to heavy hydrocarbons (see Figs. 1-3, depicting peaks for a wide range of hydrocarbons). Considering the unprocessed nature of Ibemere’s chromatographic results in Figs. 1-3, and noting that crude oil comprises at least some light hydrocarbons (such as methane, which would be gaseous at STP), one may reasonably conclude that Ibemere discloses that the hydrocarbon molecular range includes gas.
Regarding claim 10, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 1. Ibemere further discloses that the geological data comprise depositional environment (pg. 1, Abstract, ¶3: “The results of the analysis suggest that 50% of the analyzed samples were from oxic paleoenvironment”).
Regarding claim 11, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 1. Chen further teaches that a convolutional neural network is useful for avoiding overfitting and improving reliability and accuracy in oil fingerprinting (pg. 10, column 2: “In the future, we would like to utilize different state-of-the-art NN in deep learning. For example, applying a combination of feature selection, model interpretation algorithms and NN, especially the convolutional neural network (CNN), for classification tasks to avoid overfitting and improve reliability and accuracy in oil fingerprinting.”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Chen with the invention of Ibemere in view of Chen and Knight and Walters by causing the ML network to comprise a CNN. Doing so would avoid overfitting and improve reliability and accuracy.
Regarding claim 16, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 15. Considering that Ibemere is interested in oil exploration and production (see pg. 1, Abstract, ¶1), it would have been obvious to one of ordinary skill practicing the invention of Ibemere in view of Chen and Knight and Walters to cause the system to comprise a production system configured to extract the oil sample from at least one of the two or more wells. Doing so would enable one to analyze a subterranean region of interest for purposes of oil exploration and production.
Regarding claim 17, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 15. Furthermore, it would have been obvious to one of ordinary skill in the art practicing the invention of Ibemere in view of Chen and Knight and Walters to communicably couple the gas chromatography system and the computer processor in order to enable the computer to obtain geochemical data for an oil sample when the sample is analyzed by the gas chromatography system.
Regarding claim 18, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 15. Ibemere further discloses that the gas chromatography system comprises a chromatographic column (pg. 1, Abstract, ¶2: “The analysis was done with Gas Chromatographic instrument equipped with flame ionization detector (GC-FID) and HP-PONA capillary column.”).
Regarding claim 19, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 15. Ibemere further discloses that the gas chromatography system comprises a flame ionization detector (pg. 1, Abstract, ¶2: “The analysis was done with Gas Chromatographic instrument equipped with flame ionization detector (GC-FID) and HP-PONA capillary column.”).
Regarding claim 20, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 15, but does not explicitly teach the limitations of claim 20. However, Ibemere teaches that gas chromatography with a mass spectrometer may be used to perform crude oil fingerprinting (pg. 1, Introduction, ¶1: “A crude oil fingerprint is the signature produced when a given volume of oil sample is injected into the Gas Chromatograph equipped with Flame Ionization Detector (GC-FID) or Gas Chromatograph equipped with Mass Spectrophotometer (GC-M/S) under certain conditions.”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Ibemere with the invention of Ibemere in view of Chen and Knight and Walters by using a gas chromatography system comprising a mass spectrometer in place of a gas chromatography system comprising a flame ionization detector. Both systems are recommended for fingerprinting oil samples to determine their geochemistry, therefore one would reasonably expect to obtain similar results by substituting GS-FID with GC-M/S (see MPEP 2143(I)(B)).
Regarding claim 21, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 1. Furthermore, Chen teaches training its ML algorithms using training sets (pg. 3, Section 2.3, ¶1: “before feeding input values into ML algorithms, the preprocessed datasets were divided into training sets (80%) and test sets (20%)”). Note also that all of the ML algorithms used by Chen are or may include supervised learning algorithms (see Abstract).
In supervised learning, an ML algorithm is trained on multiple pieces of data for which each input is associated with a particular correct output. In the context of Ibemere in view of Chen and Knight and Walters, the input would be geochemical data of an oil sample obtained by GC analysis and including abundances of the hydrocarbon molecules, where those abundances are presented on first multi-dimensional plots (cf. Ibemere, Fig. 4 and rejection of claim 1), while the output would be an estimate of whether the sample is from a particular depositional environment (i.e. oxic or anoxic paleoenvironment; see rejection of claim 1). Thus, to train an ML algorithm to perform this classification, it would have been obvious to one of ordinary skill in the art to:
obtain training geochemical data and training geological data for a plurality of oil samples, wherein each oil sample among the plurality of oil samples comprises the hydrocarbon molecules, wherein the training geochemical data comprises abundances of the hydrocarbon molecules presented on first multi-dimensional plots, and wherein the training geological data comprises at least one of training origin data, training depositional environment data, or training organofacies data; and
training the machine learning network using the training geochemical data and the training geological data, wherein the trained machine learning network produces the geological data in response to the geochemical data.
Regarding claim 22, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 21. Furthermore, Chen teaches determining training input and output data from a database using an artificial intelligence algorithm (Fig. 1, Feature extraction from Entire datasets; pg. 3, top of column 1: “Fig. 1 showed the workflow of the framework, where datasets were preprocessed…Each square represents a machine learning operator”; machine learning is a kind of artificial intelligence algorithm. See also pg. 3, column 2, under Section 2.3: “the preprocessed datasets were divided into training sets (80%) and test sets (20%).” From this it follows that the dataset from which data was taken and processed included input and output data.); and determining if data is missing using the artificial intelligence algorithm (pg. 3, ¶ under 2.2 Data entry and preprocessing: “The raw datasets might not be desirable to train ML algorithms directly because of missing data, outliers, or merely heavy computations…To simplify the feature input process while maintaining the most information of datasets, principal component analysis (PCA) was introduced in this framework. It decreased hundreds of features to dozens and hence significantly reduced computational time. By reducing dimensionality through PCA, the datasets could be denoised as well.”).
Chen does not explicitly recite separating training input and output data from the database, however it is reasonable to expect that the process of determining data from the database would have involved a separation step to distinguish data meant to be input into a ML model from data meant to be compared to the ML model’s output. Doing so would have been useful to ensure output data is not treated as input data and vice versa.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Chen with the invention of Ibemere in view of Chen and Knight and Walters by causing obtaining the training geochemical data and the training geological data to further comprise:
separating the training geochemical data and the training geological data from a database using an artificial intelligence algorithm; and
determining if at least one of the training geochemical data or the training geological data for each oil sample is missing using the artificial intelligence algorithm.
Doing so would enable one to automatically obtain the data needed to train the ML network, and to remove noise caused by oil samples that have insufficient data to properly train the ML algorithm.
Regarding claim 23, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 22. Chen further teaches using a random forest algorithm to classify oil samples (pg. 3, below Table 2: “This study introduces ML as a new analysis tool by proposing a new binary classification framework to aid source identification in oil fingerprinting by distinguishing weathered crude oil (WCO) and CDO. The framework comprises six ML algorithms…The ML algorithms considered in the study include Random Forest.”).
Considering that the algorithm for determining the data and training on the data can be considered a single artificial intelligence algorithm, it would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Chen with the invention of Ibemere in view of Chen and Knight and Walters by causing the artificial intelligence algorithm to comprise a random forest algorithm. Doing so would enable one to use a random forest algorithm as the ML network for determining depositional environment of an oil sample given its geochemical data.
Regarding claim 24, Ibemere in view of Chen and Knight and Walters teaches the limitations of claim 22. As seen in the rejection of claim 21, Chen teaches determining outliers using artificial intelligence (see pg. 3, ¶ under 2.2 Data entry and preprocessing, quoted above).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the teachings of Chen with the invention of Ibemere in view of Chen and Knight and Walters by causing the obtaining the training geochemical data and the training geological data to further comprise determining a training geological data outlier using the artificial intelligence algorithm. Doing so would enable one to remove unnecessary noisy data improper for training the ML algorithm.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Jung (“ChartSense: Interactive Data Extraction from Chart Images”) describes using deep learning to determine a chart type then uses semi-automatic methods to extract underlying data (Abstract). The method works with star plots (called “radar” plots in Jung; see for example Fig. 1).
Liu (“Data Extraction from Charts via Single Deep Neural Network”) supplies an automatic data extraction model usable on bar and pie charts, and extendable to other charts (Abstract).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ETHAN WESLEY EDWARDS whose telephone number is (571)272-0266. The examiner can normally be reached Monday - Friday, 7:30am-5pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Andrew Schechter can be reached at (571) 272-2302. 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.
ETHAN WESLEY EDWARDS
Examiner
Art Unit 2857
/E.W.E./ Examiner, Art Unit 2857
/ANDREW SCHECHTER/ Supervisory Patent Examiner, Art Unit 2857