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 Amendment
This Office Action is in response to applicant’s communication filed 29 June 2026, in response to the Office Action mailed 7 May 2026. The applicant’s remarks and any amendments to the claims or specification have been considered, with the results that follow.
Claim Objections
A series of singular dependent claims is permissible in which a dependent claim refers to a preceding claim which, in turn, refers to another preceding claim.
A claim which depends from a dependent claim should not be separated by any claim which does not also depend from said dependent claim (see, e.g., claims 15-16). It should be kept in mind that a dependent claim may refer to any preceding independent claim. In general, applicant's sequence will not be changed. See MPEP § 608.01(n).
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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
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) 1-5 and 17-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Converset (US 11,962,957) in view Saeed (US 2018/0075161).
As per claim 1, Converset teaches a large language model (LLM) assisted knowledge base [a language machine learning model (col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) and one or more associated databases with drilling data (col. 6, line 60 to col. 7, line 9; etc.)], comprising:
a processor; and
a memory connected to the processor, the memory including instructions that, when executed by the processor, cause the processor to establish: [the computing device includes a processor and connected storage storing instructions to be executed by the processor (col. 14, lines 28-51; fig. 8; etc.)]
an LLM trained with drilling operations domain knowledge [training a language machine learning model(s) (col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) from one or more associated databases with drilling data (col. 6, line 60 to col. 7, line 9; etc.)];
a [knowledge base] connected to the LLM [a language machine learning model (col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) and one or more associated databases with drilling data (col. 6, line 60 to col. 7, line 9; etc.)],
wherein the LLM receives one or more queries associated with a drilling operation, the drilling operation having current drilling parameters [a language machine learning model (col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.), where the ML model can access one or more additional databases for information about the wellsite (col. 6, line 60 to col. 7, line 9; etc.), from which the data can be queried (col. 7, lines 58-64; etc.); where the wellsite/drill has current drilling parameters], and
the response changing one or more of the current drilling parameters [the model(s) may provide information to the user, including options for the user to validate the data or provide a control response (col. 7, lines 21-43; col. 8, lines 9-32; etc.) which can include changing one or more operating parameters of the wellsite/drilling system (col. 8, line 54 to col. 9, line 2; col. 20, lines 43-58; etc.)].
While Converset teaches a language model trained with drilling data from an associated knowledge base (see above), it has not been relied upon for teaching wherein the knowledge base is a semantic web; a converter, wherein the converter retrieves structured drilling records, converts the structured drilling records into Resource Description Framework (RDF) format and stores the RDF formatted drilling records to the semantic web, wherein the LLM interprets unstructured data and saves the interpreted unstructured data in RDF format to the semantic web, and wherein the LLM accesses the semantic web based on the one or more queries, receives structured data from the semantic web in response to the one or more queries and returns a response based on the received structured data.
Saeed teaches a semantic web knowledge base [disparate data sources are integrated and encoded into a semantic web using semantic web standards (paras. 0031, 0039, 0041, etc.)];
a converter, wherein the converter retrieves structured drilling records, converts the structured drilling records into Resource Description Framework (RDF) format and stores the RDF formatted drilling records to the semantic web [the ontological model is used for converting data into RDF triples (para. 0056, etc.) from structured drilling data (paras. 0063, 0075, etc.) to be integrated into the semantic web (paras. 0031, 0039, 0041, etc.)],
wherein the LLM interprets unstructured data and saves the interpreted unstructured data in RDF format to the semantic web [the model can also access and interpret structured or unstructured drilling data (para. 0075, etc.) and convert it to RDF format (para. 0056, etc.); with the language model of Converset, above], and
wherein the LLM accesses the semantic web based on the one or more queries, receives structured data from the semantic web in response to the one or more queries and returns a response based on the received structured data [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the database (paras. 0031-33, 0036, 0054-56, 0064, etc.), where the model can access and interpret structured or unstructured drilling data (para. 0075, etc.)].
Converset and Saeed are analogous art, as they are within the same field of endeavor, namely using machine learning models for interpreting drilling data for drilling control/predictions.
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to convert the drilling data into a specified format and store in a semantic web, as taught by Saeed, for the drilling data with the language model in the system taught by Converset.
Saeed provides motivation as [storing the drilling data in the semantic web provides more efficient access and improved accuracy of predictions (para. 0075, etc.)].
As per claim 2, Converset/Saeed teaches wherein the LLM includes a SPARQL plug-in, the SPARQL plug-in configured to generate SPARQL to insert new information from unstructured records into the semantic web [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the database (Saeed: paras. 0031-33, 0036, 0054-56, 0064, etc.); for the language machine learning model (Converset: col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) and one or more associated databases with drilling data (Converset: col. 6, line 60 to col. 7, line 9; etc.)], and
wherein the current drilling parameters include formation parameters and drilling tool parameters [The conditions of the equipment of the drilling system 100, the formation 101, the wellbore 102, the drilling fluid, or other part of the wellsite can change during operations. Sensors within the wellsite provide information to make operation decisions for efficiency, safety, and other reasons (Converset: col. 4, lines 60-65; etc.), may have stored thereon or access to a drilling plan for a drill site, well information for a producing wellsite, a drilling interpretation log, drilling rig hours, formation information or other wellsite logs (Converset: col. 18, line 64 to col. 19, line 3; etc.) from one or more sensors of the plurality of sensors located proximate the control cabin (or otherwise at the surface), the drill rig, the drill string, the BHA, the formation, or other locations at the wellsite (Converset: col. 23, lines 43-47; etc.), to change at least one operating parameter of the equipment at the wellsite (Converset: col. 8, lines 62-66; etc.); which includes parameters of both the formation and drilling tool(s)].
As per claim 3, Converset/Saeed teaches wherein generating SPARQL to insert new information from unstructured records into the semantic web includes applying a schema to the unstructured records [the model can access and interpret structured or unstructured drilling data (Saeed: para. 0075, etc.) and, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the database (Saeed: paras. 0031-33, 0036, 0054-56, 0064, etc.); for the language machine learning model (Converset: col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) and one or more associated databases with drilling data (Converset: col. 6, line 60 to col. 7, line 9; etc.)].
As per claim 4, Converset/Saeed teaches wherein the LLM includes a SPARQL plug-in, the SPARQL plug-in configured to generate SPARQL to query the semantic web [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the database (Saeed: paras. 0031-33, 0036, 0054-56, 0064, etc.); for the language machine learning model (Converset: col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) and one or more associated databases with drilling data (Converset: col. 6, line 60 to col. 7, line 9; etc.)].
As per claim 5, Converset/Saeed teaches wherein the LLM includes a SPARQL plug-in, the SPARQL plug-in configured to generate SPARQL to edit the semantic web under user control [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the database (Saeed: paras. 0031-33, 0036, 0054-56, 0064, etc.); for the language machine learning model (Converset: col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) and one or more associated databases with drilling data (Converset: col. 6, line 60 to col. 7, line 9; etc.); where the queries/commands are user control (see, e.g., Converset: col. 7, lines 10-20; Saeed: paras. 0064-65; etc.)].
As per claim 17, Converset teaches a drilling platform controller [a computing device/controller for wellsite control (col. 4, line 66 to col. 5, line 23; col. 17, lines 1-25; etc.)], comprising:
a processor; and
a memory connected to the processor, the memory including instructions that, when executed by the processor, cause the processor to: [the computing device includes a processor and connected storage storing instructions to be executed by the processor (col. 14, lines 28-51; fig. 8; etc.)]
receive current drilling parameters [the computing device/controller receives sensor information including current operating parameters, environmental conditions, etc. (col. 4, line 66 to col. 5, line 23; etc.)];
query a large language model (LLM) assisted knowledge base based on the current drilling parameters [a language machine learning model (col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.), where the ML model can access one or more additional databases for information about the wellsite (col. 6, line 60 to col. 7, line 9; etc.), from which the data can be queried (col. 7, lines 58-64; etc.)], the LLM assisted knowledge base including an LLM trained with drilling operations domain knowledge [training a language machine learning model(s) (col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) from one or more associated databases with drilling data (col. 6, line 60 to col. 7, line 9; etc.)], a [knowledge base] connected to the LLM [a language machine learning model (col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.), where the ML model can access one or more additional databases for information about the wellsite (col. 6, line 60 to col. 7, line 9; etc.)], the query received by the LLM [a language machine learning model (col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.), where the ML model can access one or more additional databases for information about the wellsite (col. 6, line 60 to col. 7, line 9; etc.), from which the data can be queried (col. 7, lines 58-64; etc.)];
receive a response from the LLM [the model(s) may provide information to the user, including options for the user to validate the data or provide a control response (col. 7, lines 21-43; col. 8, lines 9-32; etc.)]; and
change one or more current drilling parameters based on the response [the model(s) may provide information to the user, including options for the user to validate the data or provide a control response (col. 7, lines 21-43; col. 8, lines 9-32; etc.) which can include changing one or more operating parameters of the wellsite/drilling system (col. 8, line 54 to col. 9, line 2; col. 20, lines 43-58; etc.)].
While Converset teaches a language model trained with drilling data from an associated knowledge base (see above), it has not been relied upon for teaching wherein the knowledge base is a semantic web; and a converter, wherein the converter retrieves structured drilling records from a database, converts the structured drilling records into Resource Description Framework (RDF) format and stores the RDF formatted drilling records to the semantic web, [the query] applied by the LLM to the semantic web; [and] the response based on structured data returned from the semantic web.
Saeed teaches a semantic web knowledge base [disparate data sources are integrated and encoded into a semantic web using semantic web standards (paras. 0031, 0039, 0041, etc.)];
[querying] a semantic web knowledge base based on the current drilling parameters [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the database (paras. 0031-33, 0036, 0054-56, 0064, etc.)]; and
a converter, wherein the converter retrieves structured drilling records from a database, converts the structured drilling records into Resource Description Framework (RDF) format and stores the RDF formatted drilling records to the semantic web [the ontological model is used for converting data into RDF triples (para. 0056, etc.) from structured drilling data (paras. 0063, 0075, etc.) to be integrated into the semantic web (paras. 0031, 0039, 0041, etc.)],
[the query] applied by the LLM to the semantic web [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the database (paras. 0031-33, 0036, 0054-56, 0064, etc.)]; [and]
the response based on structured data returned from the semantic web [the ontological model is used for converting data into RDF triples (para. 0056, etc.) from structured drilling data (paras. 0063, 0075, etc.) to be integrated into the semantic web (paras. 0031, 0039, 0041, etc.)].
Converset and Saeed are analogous art, as they are within the same field of endeavor, namely using machine learning models for interpreting drilling data for drilling control/predictions.
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to convert the drilling data into a specified format and store in a semantic web, as taught by Saeed, for the drilling data with the language model in the system taught by Converset.
Saeed provides motivation as [storing the drilling data in the semantic web provides more efficient access and improved accuracy of predictions (para. 0075, etc.)].
As per claim 18, Converset/Saeed teaches wherein the memory further includes instructions that, when executed by the processor, cause the processor to submit an edit to the semantic web through the LLM [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the database (Saeed: paras. 0031-33, 0036, 0054-56, 0064, etc.); for the language machine learning model (Converset: col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) and one or more associated databases with drilling data (Converset: col. 6, line 60 to col. 7, line 9; etc.)].
As per claim 19, Converset/Saeed teaches wherein the current drilling parameters include formation parameters and drilling tool parameters [The conditions of the equipment of the drilling system 100, the formation 101, the wellbore 102, the drilling fluid, or other part of the wellsite can change during operations. Sensors within the wellsite provide information to make operation decisions for efficiency, safety, and other reasons (Converset: col. 4, lines 60-65; etc.), may have stored thereon or access to a drilling plan for a drill site, well information for a producing wellsite, a drilling interpretation log, drilling rig hours, formation information or other wellsite logs (Converset: col. 18, line 64 to col. 19, line 3; etc.) from one or more sensors of the plurality of sensors located proximate the control cabin (or otherwise at the surface), the drill rig, the drill string, the BHA, the formation, or other locations at the wellsite (Converset: col. 23, lines 43-47; etc.), to change at least one operating parameter of the equipment at the wellsite (Converset: col. 8, lines 62-66; etc.); which includes parameters of both the formation and drilling tool(s)], and
wherein changing one or more of the current drilling parameters based on the response includes changing one or more of Weight on Bit (WOB), Rate of Penetration (ROP), RPM or flowrate [the model(s) may provide information to the user, including options for the user to validate the data or provide a control response (Converset: col. 7, lines 21-43; col. 8, lines 9-32; etc.) which can include changing one or more operating parameters of the wellsite/drilling system (Converset: col. 8, line 54 to col. 9, line 2; col. 20, lines 43-58; etc.); including weight-on-bit (WOB), fluid pressure, flow rate, etc. (Converset: col. 5, lines 12-23; col. 6, lines 1-12; etc.)].
As per claim 20, Converset/Saeed teaches wherein the current drilling parameters include formation parameters and drilling tool parameters [The conditions of the equipment of the drilling system 100, the formation 101, the wellbore 102, the drilling fluid, or other part of the wellsite can change during operations. Sensors within the wellsite provide information to make operation decisions for efficiency, safety, and other reasons (Converset: col. 4, lines 60-65; etc.), may have stored thereon or access to a drilling plan for a drill site, well information for a producing wellsite, a drilling interpretation log, drilling rig hours, formation information or other wellsite logs (Converset: col. 18, line 64 to col. 19, line 3; etc.) from one or more sensors of the plurality of sensors located proximate the control cabin (or otherwise at the surface), the drill rig, the drill string, the BHA, the formation, or other locations at the wellsite (Converset: col. 23, lines 43-47; etc.), to change at least one operating parameter of the equipment at the wellsite (Converset: col. 8, lines 62-66; etc.); which includes parameters of both the formation and drilling tool(s)], and
the current drilling parameters include one or more of WOB, ROP, RPM, flowrate or inclination [the sensor information/operating parameters of the wellsite captured and controlled can include weight-on-bit (WOB), fluid pressure, flow rate, etc. (Converset: col. 5, lines 12-23; col. 6, lines 1-12; etc.)].
Claim(s) 6 is/are rejected under 35 U.S.C. 103 as being unpatentable over Converset and Saeed as applied to claim 1 above, and further in view of Newman (US 12,072,918).
As per claim 6, Converset/Saeed teaches the LLM assisted semantic web knowledge base of claim 1, as described above.
While Converset/Saeed teaches a converter to convert to RDF format (see above), it has not been relied upon for teaching wherein the converter is a Relational Database Service (RDS) to RDF converter.
Newman teaches wherein the converter is a Relational Database Service (RDS) to RDF converter [an SQL-to-RDF converter may be used to convert Relational Database tables (a Relational Database Service) to RDF (col. 15, lines 21-32, etc.)].
Converset/Saeed and Newman are analogous art, as they are within the same field of endeavor, namely utilizing a converter to convert data to RDF format for a semantic database.
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to utilize the relational database service to resource description framework converter, taught by Newman, as (or in addition to) the converter for converting structured and unstructured drilling data to RDF format, in the system taught by Converset/Saeed.
Newman provides motivation as [relational database to resource description framework conversion facilitates the mapping of existing relational data to an RDF data model (e.g., a knowledge graph) (col. 15, lines 21-44, etc.)].
Claim(s) 7-15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Converset (US 11,962,957), in view Saeed (US 2018/0075161), and further in view of Liu et al. (Well Logging Based Lithology Identification Model Establishment Under Data Drift: A Transfer Learning Method, June 2020, pgs. 1-17).
As per claim 7, Converset/Saeed teaches wherein the LLM includes domain knowledge obtained from domain knowledge sources [training a language machine learning model(s) (Converset: col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) from one or more associated databases with drilling data (Converset: col. 6, line 60 to col. 7, line 9; etc.)].
While Converset/Saeed teaches wherein the LLM includes domain knowledge obtained from domain knowledge sources (see above), it has not been relied upon for teaching wherein the model includes domain knowledge obtained from transfer learning of domain knowledge sources.
Liu teaches wherein the model includes domain knowledge obtained from transfer learning of domain knowledge sources [a transfer learning method named data drift joint adaptation extreme learning machine (DDJA-ELM) is used to train a model from old models of well data sources (pg. 1, abstract, etc.)].
Converset/Saeed and Liu are analogous art, as they are within the same field of endeavor, namely training a machine learning model from wellsite and drilling domain knowledge sources.
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to include transfer learning for training the model on well site data, from old models, as taught by Liu, in the training of the language model on well site data for controlling a specific drilling operation in the system taught by Converset/Saeed.
Liu provides motivation as [transfer learning improves the model performance and accuracy by incorporating data from old models and applying it to a model for a new well/model (Liu: pg. 1, abstract, etc.)].
As per claim 8, Converset teaches a method comprising:
building a [knowledge base] for drilling records [a language machine learning model (col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) and one or more associated databases with drilling data (col. 6, line 60 to col. 7, line 9; etc.)];
applying [machine] learning to drilling domain knowledge to obtain a large language model (LLM) with drilling domain knowledge [training a language machine learning model(s) (col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) from one or more associated databases with drilling data (col. 6, line 60 to col. 7, line 9; etc.)],
While Converset teaches a language model trained with drilling data from an associated knowledge base (see above), it has not been relied upon for teaching building a semantic web for drilling records, wherein building includes converting drilling records retrieved from a database into a Resource Description Framework (RDF) format and storing the RDF formatted drilling records in the semantic web; applying transfer learning to drilling domain knowledge to obtain a large language model (LLM) with drilling domain knowledge; the LLM configured to query the semantic web and to manipulate RDF formatted data within the semantic web; applying a schema to unstructured data records to extract information from unstructured records; and storing the extracted information in RDF format to the semantic web.
Saeed teaches building a semantic web for drilling records [disparate data sources are integrated and encoded into a semantic web using semantic web standards (paras. 0031, 0039, 0041, etc.)], wherein building includes converting drilling records retrieved from a database into a Resource Description Framework (RDF) format and storing the RDF formatted drilling records in the semantic web [the ontological model is used for converting data into RDF triples (para. 0056, etc.) from drilling data records (paras. 0063, 0075, etc.) to be integrated into the semantic web (paras. 0031, 0039, 0041, etc.)];
the LLM configured to query the semantic web and to manipulate RDF formatted data within the semantic web [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the database (paras. 0031-33, 0036, 0054-56, 0064, etc.); for the language machine learning model of Converset, above];
applying a schema to unstructured data records to extract information from the unstructured records [the model can also access and interpret structured or unstructured drilling data (para. 0075, etc.) and convert it to RDF format (para. 0056, etc.) to be integrated and encoded into a semantic web using semantic web standards (paras. 0031, 0039, 0041, etc.)]; and
storing the extracted information in RDF format to the semantic web [the model can also access and interpret structured or unstructured drilling data (para. 0075, etc.) and convert it to RDF format (para. 0056, etc.) to be integrated and encoded into a semantic web using semantic web standards (paras. 0031, 0039, 0041, etc.)].
Converset and Saeed are analogous art, as they are within the same field of endeavor, namely using machine learning models for interpreting drilling data for drilling control/predictions.
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to convert the drilling data into a specified format and store in a semantic web, as taught by Saeed, for the drilling data with the language model in the system taught by Converset.
Saeed provides motivation as [storing the drilling data in the semantic web provides more efficient access and improved accuracy of predictions (para. 0075, etc.)].
While Converset/Saeed teaches training the LLM with domain knowledge obtained from domain knowledge sources (see above), it has not been relied upon for teaching applying transfer learning to drilling domain knowledge to obtain a large language model (LLM) with drilling domain knowledge.
Liu teaches applying transfer learning to drilling domain knowledge to obtain a large language model (LLM) with drilling domain knowledge [a transfer learning method named data drift joint adaptation extreme learning machine (DDJA-ELM) is used to train a model from old models of well data sources (pg. 1, abstract, etc.); for the language model of Converset/Saeed, above].
Converset/Saeed and Liu are analogous art, as they are within the same field of endeavor, namely training a machine learning model from wellsite and drilling domain knowledge sources.
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to include transfer learning for training the model on well site data, from old models, as taught by Liu, in the training of the language model on well site data for controlling a specific drilling operation in the system taught by Converset/Saeed.
Liu provides motivation as [transfer learning improves the model performance and accuracy by incorporating data from old models and applying it to a model for a new well/model (Liu: pg. 1, abstract, etc.)].
As per claim 9, Converset/Saeed/Liu teaches building a computation model for drilling records, wherein building includes creating a computational framework of algorithms, statistics, and machine learning networks that compute answer products from unstructured or RDF formatted drilling records based on prompts received through the LLM [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the RDF data items of the database (Saeed: paras. 0031-33, 0036, 0054-56, 0064, etc.); for the language machine learning model (Converset: col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) and one or more associated databases with drilling data (Converset: col. 6, line 60 to col. 7, line 9; etc.); where the responses to queries are the answer products]; and
storing information based on the answer products in the semantic web with appropriate classification [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the RDF data items of the database (Saeed: paras. 0031-33, 0036, 0054-56, 0064, etc.); where model(s) may provide information to the user, including options for the user to validate the data or provide a control response (Converset: col. 7, lines 21-43; col. 8, lines 9-32; etc.); where the answer to the query provided to the user is an answer product, and the validation of the response by the user provided back to the database is the information stored with appropriate classification].
As per claim 10, Converset/Saeed/Liu teaches wherein the unstructured records include drilling reports and product manuals [the model can access and interpret structured or unstructured drilling data (Saeed: para. 0075, etc.) and, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the database (Saeed: paras. 0031-33, 0036, 0054-56, 0064, etc.); which drilling data can include current operation parameters (Converset: col. 8, line 54 to col. 9, line 2; col. 20, lines 43-58; etc.); including weight-on-bit (WOB), fluid pressure, flow rate, etc. (Converset: col. 5, lines 12-23; col. 6, lines 1-12; etc.) as well as predefined safety thresholds for the received parameters (Converset: col. 17, lines 1-25; etc.); where the current operating parameters of the drill are drilling reports, and the predefined safety thresholds are product manual information].
As per claim 11, Converset/Saeed/Liu teaches wherein the method further includes:
receiving a query at the LLM [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the RDF data items of the database (Saeed: paras. 0031-33, 0036, 0054-56, 0064, etc.); for the language machine learning model (Converset: col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) and one or more associated databases with drilling data (Converset: col. 6, line 60 to col. 7, line 9; etc.)];
querying the semantic web based on the query received by the LLM [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the RDF data items of the database (Saeed: paras. 0031-33, 0036, 0054-56, 0064, etc.); for the language machine learning model (Converset: col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) and one or more associated databases with drilling data (Converset: col. 6, line 60 to col. 7, line 9; etc.)]; and
delivering a response to one or more of a user or a machine based on the semantic web query [the model(s) may provide information to the user in response to a query, including options for the user to validate the data or provide a control response (Converset: col. 7, lines 21-43; col. 8, lines 9-32; etc.), which can include determining whether to provide the data to a user or another ML model or other computerized model (machine) (Converset: col. 5, lines 45-55; etc.)].
As per claim 12, Converset/Saeed/Liu teaches wherein the received query is a natural language query, and the response is a natural language response [queries and the search response may be provided as keywords or natural language (Saeed: para. 0049, etc.)].
As per claim 13, Converset/Saeed/Liu teaches wherein the received query is a machine language query, and the response is a machine language response [queries and the search response may be provided via an automatic query language generator and/or query execution system for semantic data (Saeed: paras. 0045-46, etc.); which are machine language queries/responses].
As per claim 14, Converset/Saeed/Liu teaches wherein querying the semantic web based on the query received by the LLM includes generating a SPARQL query from the query received by the LLM [the query to the database/model includes generating a SPARQL query (Saeed: para. 0050, etc.); for the language machine learning model (Converset: col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.) and one or more associated databases with drilling data (Converset: col. 6, line 60 to col. 7, line 9; etc.)].
As per claim 15, Converset/Saeed/Liu teaches wherein the method further includes:
receiving an edit at the LLM; and
editing the semantic web based on the edit received by the LLM [the data accessed by the ontological model, once integrated and encoded into the semantic web are accessible via SPARQL queries to read from- or commands to store to the RDF data items of the database (Saeed: paras. 0031-33, 0036, 0054-56, 0064, etc.); and the language model(s) may provide information to the user, including options for the user to validate the data or provide a control response (Converset: col. 7, lines 21-43; col. 8, lines 9-32; etc.) which can include changing one or more operating parameters of the wellsite/drilling system (Converset: col. 8, line 54 to col. 9, line 2; col. 20, lines 43-58; etc.)].
Claim(s) 16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Converset, Saeed, and Liu as applied to claim 8 above, and further in view of Newman (US 12,072,918).
As per claim 16, see the rejection of claim 6, above. [Examiner’s Note: the RDS to RDF converter shows the relational database (RDS) and the converting.]
Response to Arguments
Applicant’s arguments, see the remarks, filed 20 June 2026, with respect to the rejection of claim 16 under 35 U.S.C. 112(b) have been fully considered and are persuasive. The rejection has been withdrawn.
Applicant’s arguments, see the remarks, filed 20 June 2026, with respect to the rejection of claims 1-7 under 35 U.S.C. 101 have been fully considered and are persuasive in view of the amendments filed. The rejection has been withdrawn.
Applicant's arguments filed 29 June 2026, with respect to the rejections under 35 U.S.C. 103, have been fully considered but they are not persuasive.
Applicant argues that the cited art does not teach an LLM that “[assists] a knowledge base.”
However, Converset teaches a language machine learning model (col. 12, line 67 to col. 13, line 21; col. 24, lines 37-61; etc.), where the ML model can access one or more additional databases for information about the wellsite (col. 6, line 60 to col. 7, line 9; etc.), and may provide information to the user, including options for the user to validate the data or provide a control response (col. 7, lines 21-43; col. 8, lines 9-32; etc.) which can include changing one or more operating parameters of the wellsite/drilling system (col. 8, line 54 to col. 9, line 2; col. 20, lines 43-58; etc.). Utilizing an LLM with associated drilling database to provide changes to one or more operating parameters is within the broadest reasonable interpretation of a large language model (LLM) assisted knowledge base.
In response to applicant's arguments against the references individually, regarding Converset not teaching the LLM “[assisting] a semantic web,” having “a semantic web connected to [an] LLM,” and “using an LLM to interpret queries and to supply the queries to the semantic web,” one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986).
In this case, Converset has been relied upon for teaching an LLM querying one or more databases, while Saeed has been relied upon for teaching querying the semantic web to access and interpret structured or unstructured drilling data (see above). Additionally, regarding
Applicant also argues that the cited art does not teach “applying a schema to unstructured data records to extract information from the unstructured records,” or “storing the extracted information in RDF format to the semantic web” and that Saeed “teaches away from the idea.”
However, Saeed teaches the model can also access and interpret structured or unstructured drilling data (para. 0075, etc.) and convert it to RDF format (para. 0056, etc.) to be integrated and encoded into a semantic web using semantic web standards (paras. 0031, 0039, 0041, etc.). Additionally, it is not clear how querying two integrated semantic data sources and integrated and encoding the data into a semantic web teaches away from applying a schema to unstructured data records. Rather, it teaches that applying the schema is better than “having to asses unstructured data and manually combining such data.”
In response to applicant's arguments against the references individually, one cannot show nonobviousness by attacking references individually, regarding Liu applying transfer learning to the drilling domain knowledge to obtain an LLM with drilling domain knowledge, etc., where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986).
In this case, Liu has been relied upon for teaching using transfer learning to improve model performance and accuracy, by incorporating data from old models and applying it to a model for a new well/model (Liu: pg. 1, abstract, etc.) for the training of the language model on well site data for controlling a specific drilling operation in the system taught by Converset/Saeed (see above).
Conclusion
The following is a summary of the treatment and status of all claims in the application as recommended by M.P.E.P. 707.07(i): claims 1-20 are rejected.
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Dixon (US 2021/0019351 – cited in an IDS) – discloses building a relational framework for geologic formation operations, including drilling data converted to RDF using SPARQL queries.
Glesinger (US 2023/0140125 – cited in an IDS) – discloses generating a hierarchical/semantic database where neural network-based trained large language models may be applied to generate the elements of the semantic levels and/or through application of semantic chaining processes.
Dawson (US 2010/0228693) – discloses a system/method of generating document representations in a semantic web.
Xu et al. (Complementing GPT-3 with Few-Shot Sequence-to-Sequence Semantic Parsing over Wikidata, May 2023, pgs. 1-12) – discloses a system/method utilizing semantic parsers with an LLM (GPT-3) to build a knowledge base.
The examiner requests, in response to this Office action, that support be shown for language added to any original claims on amendment and any new claims. That is, indicate support for newly added claim language by specifically pointing to page(s) and line number(s) in the specification and/or drawing figure(s). This will assist the examiner in prosecuting the application.
When responding to this office action, Applicant is advised to clearly point out the patentable novelty which he or she thinks the claims present, in view of the state of the art disclosed by the references cited or the objections made. He or she must also show how the amendments avoid such references or objections. See 37 CFR 1.111(c).
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to GEORGE GIROUX whose telephone number is (571)272-9769. The examiner can normally be reached M-F 10am-6pm.
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, Omar Fernandez Rivas can be reached at 571-272-2589. 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.
/GEORGE GIROUX/Primary Examiner, Art Unit 2128
Prosecution Insights