Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 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 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1:
All claims are directed towards either a method, a system or a non-transitory computer-readable medium and thus satisfies Step 1 as falling into one of the statutory categories.
Step 2A, Prong One:
Independent Claim 1 recites (the same analysis applies to similar independent Claims 11 and 20):
based on the one or more subject areas associated with the query and a respective domain-specific knowledge of each domain-specific language model from a plurality of domain-specific language models, selecting one or more domain-specific language models from the plurality of domain-specific language models to answer the query;
this limitation, under its broadest reasonable interpretation, covers concepts that can be performed in the human mind and therefore would fall under the “Mental Processes” groupings of abstract ideas. That is a person can select the most pertinent language models from a plurality of language models to answer a query based on judgement and evaluation.
and generating, by the control language model, a response to the query based on one or more responses to the query received from the one or more domain-specific language models.
This limitation, under its broadest reasonable interpretation, also covers concepts that can be performed in the human mind and therefore would fall under the “Mental Processes” groupings of abstract ideas. That is a person is capable of generating a response to a query using judgement and evaluation (the language model is considered as using a machine learning model as a tool to perform an abstract idea - see MPEP 2106.05(f)).
Step 2A, Prong Two:
Claim 1 recites the additional elements of (the same analysis applies to similar independent Claims 11 and 20):
receiving, by a control language model configured to perform natural language processing, a query related to one or more subject areas;
sending, to the one or more domain-specific language models, a request to answer the query;
these limitations are considered as adding insignificant extra-solution activity (sending and receiving data) to the judicial exception - see MPEP 2106.05(g).
The further additional elements of “processors” as recited in independent Claims 11 and 20 are recited at a high-level of generality such that they amount to no more than mere instructions to apply the exception using a generic computer component. Accordingly, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. The claims are therefore directed to an abstract idea.
Step 2B:
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements are considered as appending well-understood, routine, conventional activities (sending and receiving data) previously known to the industry, specified at a high level of generality, to the judicial exception - see MPEP 2106.05(d). The further additional elements of “processors” as recited in independent Claims 11 and 20 amounts to no more than mere instructions to apply the exception using a generic computer component. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. The claims are therefore not patent eligible.
Dependent Claims 2-5, 11-15 are considered as using a language model as a tool to perform the abstract idea which includes its training/learning - see MPEP 2106.05(f).
Dependent Claims 6, 16 are also considered, under their broadest reasonable interpretation, to cover concepts that can be performed in the human mind and therefore would fall under the “Mental Processes” groupings of abstract ideas. That is a person is capable of combining information from different sources to formulate a response using judgement and evaluation (the language model is considered as using a machine learning model as a tool to perform an abstract idea - see MPEP 2106.05(f)).
Dependent Claim 7 and the first part of Claim 17 are considered as using a model as a tool to perform the abstract idea - see MPEP 2106.05(f).
Dependent Claim 8 and the second part of Claim 17 are considered as linking the use of the judicial exception to a particular technological environment or field of use – see MPEP 2106.05(h).
Dependent Claims 9, 18 are considered as using a model as a tool to perform the abstract idea - see MPEP 2106.05(f).
Dependent Claims 10, 19 are also considered, under their broadest reasonable interpretation, to cover concepts that can be performed in the human mind and therefore would fall under the “Mental Processes” groupings of abstract ideas. That is a person is capable of looking up information in a knowledge base for formulating a response to a query (the language model is considered as using a machine learning model as a tool to perform an abstract idea - see MPEP 2106.05(f)).
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.
Claims 1, 6, 11, 16, 20 are rejected under 35 U.S.C. 103 as being unpatentable over Walsh, US 2023/0018116 A1, in view of Cho, US 2025/0013633 A1.
Regarding Claim 1, Walsh teaches:
A method comprising: receiving, by a control language model configured to perform natural language processing, a query related to one or more subject areas (paragraph 14: “The collaborative knowledge system may receive, from a user device, a query associated with one or more of the first knowledge domain or the second knowledge domain. The collaborative knowledge system may generate a response to the query based on processing the query with the trained centralized model and may provide the response to the query to the user device”; And, paragraph 36: “the centralized model may include a neural language model. The neural language model may include a model based on a neural network (e.g., a convolutional neural network (CNN) and/or a recurrent neural net language model (RNNLM), among other examples). In some implementations, the centralized model may include a transformer-based deep learning neural network architecture. For example, the central model may include an autoregressive language model that uses deep learning to produce text in response to a query”);
based on the one or more subject areas associated with the query and a respective domain-specific knowledge of each domain-specific language model from a plurality of domain- specific language models, selecting one or more domain-specific language models from the plurality of domain-specific language models to answer the query (paragraph 21: “the domain specific language model may automatically select the set of private documents. As an example, the knowledge worker may input (e.g., via a user interface) information identifying a particular topic and/or a particular project, among other examples. The domain specific language model may identify a set of private documents from the private knowledge database based on the information input by the knowledge worker”. Examiner’s note: see also Bhat, US 2025/0005276 A1, for example paragraph 22);
and generating, by the control language model, a response to the query based on one or more responses to the query received from the one or more domain-specific language models (paragraph 14: “the collaborative knowledge system may receive a first set of privatized embeddings and a second set of privatized embeddings. The first set of privatized embeddings may be generated by a first local model based on a first set of private documents associated with a first knowledge domain. The second set of privatized embeddings may be generated by a second local model based on a second set of private documents associated with a second, different knowledge domain. The collaborative knowledge system may train, based on the first set of privatized embeddings and the second set of privatized embeddings, a centralized model to generate a trained centralized model. The collaborative knowledge system may receive, from a user device, a query associated with one or more of the first knowledge domain or the second knowledge domain. The collaborative knowledge system may generate a response to the query based on processing the query with the trained centralized model and may provide the response to the query to the user device”).
With Walsh teaching the above, Cho more directly shows:
sending, to the one or more domain-specific language models, a request to answer the query (paragraph 105: “implementing a method for providing domain-specific large language models to facilitate automated analytics of unstructured streams of natural language data. In FIG. 5, the language model may be converted into a domain-specific language model to provide responses to user requests that are received in a natural language format”).
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to use the teachings of Cho with that of Walsh for sending, to the one or more domain-specific language models, a request to answer the query.
The ordinary artisan would have been motivated to modify Walsh in the manner set forth above for the purposes of having a language model configured to perform domain-specific tasks [Cho: paragraph 106].
Regarding Claim 6, Walsh further teaches:
The method of claim 1, wherein the one or more domain-specific language models comprise a subset of domain-specific language models from the plurality of domain-specific language models and the one or more responses comprise multiple responses from different domain-specific language models of the subset of domain-specific language models, and wherein generating the response to the query comprises combining information from the multiple responses into a combined response (paragraph 50: “The collaborative knowledge system may apply the trained centralized model to the query to generate an output (e.g., a response to the query). The output may indicate a new knowledge insight generated by the centralized model by combining knowledge from multiple, separate knowledge domains”; And, paragraph 32: “each domain specific language model, of the group of domain specific language models, may generate a respective set of knowledge space embeddings” . Examiner’s note: see also Ferrucci, US 2023/0401467 A1, for example 187).
Claims 2-5, 7-10, 12-15, 17-19 are rejected under 35 U.S.C. 103 as being unpatentable over Walsh, US 2023/0018116 A1, in view of Cho, US 2025/0013633 A1, and further in view of Michalopulos, US 2025/0111336 A1.
Regarding Claim 2, Walsh teaches:
The method of claim 1, wherein each domain-specific language model from the plurality of domain-specific language models comprises a neural network that is at least one of trained with information
and configured to answer queries (paragraph 17: “a domain specific language model may include a transformer-based deep learning neural network architecture. For example, the domain specific language model may include an autoregressive language model (e.g., a generative pre-trained transformer (GPT) language prediction model) that uses deep learning to produce text in response to a query (e.g., to produce an answer to a question and/or to produce a reply to a comment, among other examples)”).
Neither Walsh nor Cho may have taught the following, however, Michalopulos shows
about a different petrophysics domain
associated with the different petrophysics domain (paragraph 165: teaches the use of large language models to answer question in different domains of petrophysics).
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to use the teachings of Michalopulos with that of Walsh and Cho for using a language model to answer queries associated with different petrophysics domain.
The ordinary artisan would have been motivated to modify Walsh and Cho in the manner set forth above for the purposes of using a large language model or other AI/ML tools for providing additional information, showing backup information, or providing a more detailed explanation to a drilling engineer query [Michalopulos: paragraph 165].
Regarding Claim 3, Walsh further teaches:
The method of claim 1, wherein each domain-specific language model from the plurality of domain-specific language models comprises a neural network that is at least one of trained with different context-specific information and configured to answer queries associated with the different context-specific information (paragraph 17: “a domain specific language model may include a transformer-based deep learning neural network architecture. For example, the domain specific language model may include an autoregressive language model (e.g., a generative pre-trained transformer (GPT) language prediction model) that uses deep learning to produce text in response to a query (e.g., to produce an answer to a question and/or to produce a reply to a comment, among other examples)”; And, paragraph 22: “The domain specific language model may obtain the set of documents from the private knowledge database based on the set of documents being associated with the topic”),
And Michalopulos further teaches:
wherein the different context-specific information comprises at least one of tool-specific information and wellbore-specific information (paragraph 45: “Sensor system 528 may be for obtaining sensor data about the drilling operation and drilling system 100, including the downhole equipment. For example, sensor system 528 may include MWD or logging while drilling (LWD) tools for acquiring information, such as toolface and formation logging information, which may be saved for later retrieval, transmitted with or without a delay using any of various communication means (e.g., wireless, wireline, or mud pulse telemetry), or otherwise transferred to steering control system 168. As used herein, an MWD tool is enabled to communicate downhole measurements without substantial delay to the surface 104, such as using mud pulse telemetry, while a LWD tool is equipped with an internal memory that stores measurements when downhole and can be used to download a stored log of measurements when the LWD tool is at the surface 104. The internal memory in the LWD tool may be a removable memory, such as a universal serial bus (USB) memory device or another removable memory device. It is noted that certain downhole tools may have both MWD and LWD capabilities. Such information acquired by sensor system 528 may include information related to hole depth, bit depth, inclination angle, azimuth angle, true vertical depth, gamma count, standpipe pressure, mud flow rate, rotary rotations per minute (RPM), bit speed, ROP, WOB, among other information”; And paragraph 123: “A core of the report can be a detailed description of the drilling progress made during that day (i.e., the time period covered by the report). This includes data about the drilled depth, wellbore diameter, drilling rate (rate of penetration), and details about the formations encountered”).
Regarding Claim 4, Walsh further teaches:
The method of claim 1, wherein the control language model is trained with
information
and each domain-specific language model from the plurality of domain-specific language models is trained with more specific
information than the control language model (paragraph 14: “the collaborative knowledge system may receive a first set of privatized embeddings and a second set of privatized embeddings. The first set of privatized embeddings may be generated by a first local model based on a first set of private documents associated with a first knowledge domain. The second set of privatized embeddings may be generated by a second local model based on a second set of private documents associated with a second, different knowledge domain”; And, paragraph 17: “a group of domain specific language models may each receive a respective set of private documents from a private knowledge database”. And, paragraph 14: “The collaborative knowledge system may train, based on the first set of privatized embeddings and the second set of privatized embeddings, a centralized model to generate a trained centralized model”. Examiner’s note: see also Biadsy, US 2018/0053502 A1, for example paragraph 9).
And Michalopulos further teaches:
petrophysics
petrophysics (paragraph 200: “The content of a drilling plan can vary depending on factors such as the type of well-being drilled (oil, gas, geothermal, etc.), the geological conditions”).
Regarding Claim 5, with Michalopulos teaching petrophysics information as previously pointed out, Walsh further teaches:
The method of claim 1, wherein the response to the query is generated based on information from the one or more responses and petrophysics information learned by the control language model to generate the response (paragraph 17: “the domain specific language model may include an autoregressive language model (e.g., a generative pre-trained transformer (GPT) language prediction model) that uses deep learning to produce text in response to a query”; And, paragraph 51: “The centralized model may process the query to determine a response (e.g., an answer to the question posed by the user). The response may include a new knowledge insight generated by the centralized model based on the centralized model being trained on the privatized embeddings associated with different knowledge domains”).
Regarding Claim 7, with Walsh and Cho teaching those limitations of the claim as previously pointed out, Michalopulos further teaches:
The method of claim 1, wherein the control language model further comprises a model configured to convert image data into text data describing the image data, wherein the query comprises visual information collected from a wellbore site, and wherein at least one of selecting the one or more domain-specific language models and generating the response is further based on the visual information (paragraph 165: “The large language models can take that input, output the correct language response, and then the video reporting tool would generate a video report including the audio narration of the answer, together with the 3D representation of the well, with the display of the current well provided with displays of the earlier three reference wells so that the video report visually shows the similarities and differences of the current well and the three earlier reference wells that were the subject of the question. The engineer could then refine the request after reviewing the generated video report and say “now take wells MB14 and X well and also compare the dip angles on each well as well as direction the wells were drilled”. The large language model would then provide a response, which the video reporting engine can use to provide an updated report with updated text and/or audio narration including the response to the second input query”; And, paragraph 152: “The system can include and use a three-dimensional (3D) rendering and visualization program (e.g., 3DViz) to provide to the end user a video showing the work done in 3D, with a summary of the pertinent information embedded in the video report. Three-dimensional (3D) rendering and visualization displays can be a powerful tool for depicting drilling data in a visually informative and engaging way. In some embodiments, the report generator engine may include a 3D rendering and visualization engine, which can receive the relevant data from the well plan, offsetting wells, or synthetic wells and the preceding drilling operations to be covered by the video report, then generate a 3D video display showing, for example, the progress of the wellbore and its trajectory during the drilling operations during the covered time interval. The video display can also show the wellbore trajectory's progress as compared to that of the well plan, offsetting wells, or synthetic wells for that time period and can also display one or more geological formations or conditions relative to the wellbore's trajectory. If the well plan does not include geological information, the 3D rendering and visualization engine can be provided with such information and then use it to generate the 3D display for the report”).
Regarding Claim 8, with Walsh and Cho teaching those limitations of the claim as previously pointed out, Michalopulos further teaches:
The method of claim 7, wherein the visual information comprises at least one of a petrophysical log, a petrophysical map, a petrophysical chart, and a petrophysical image (paragraph 153: “The three-dimensional (3D) rendering and visualization routine can receive and gather relevant drilling data, including parameters like drill bit position, depth, speed, torque, pressure, weight on bit, rate of penetrations, geological formations, bit type and characteristics, BHA type and characteristics, location of drill pipe and heavyweight drill pipe and collars, depth of cut planned versus actual on the bits run, and more. In addition, the 3D engine can receive logging information, such as gamma ray logs, resistivity logs, mud logs, and other logs, which may be included in the video display as desired. In addition, one or more of the logs displayed may be modified, such as by converting a log from a measured depth format to a true vertical depth format. Moreover, portions of one or more of the logs may be displayed against portions of one or more logs from one or more reference wells previously drilled that are included in the well plan. In addition, portions of one or more logs may be displayed and may be manipulated to show the correlation of a well log obtained during drilling with a corresponding portion of one or more reference or offset well logs or a synthetic well log. The 3D rendering and visualization routine can ensure the data is accurate, complete, and properly formatted for visualization to be able to visualize all the different pieces at once”).
Regarding Claim 9, Walsh further teaches:
The method of claim 1, wherein at least one of the control language model and one or more of the plurality of domain-specific language models comprises a transformer network (paragraph 17: “In some implementations, a domain specific language model may include a transformer-based deep learning neural network architecture. For example, the domain specific language model may include an autoregressive language model (e.g., a generative pre-trained transformer (GPT) language prediction model) that uses deep learning”),
And Michalopulos further teaches:
and wherein the plurality of domain-specific language models comprises at least one of a resistivity model configured to answer questions relating to resistivity, a nuclear magnetic resonance (NMR) model configured to answer questions relating to NMR, a porosity model configured to answer questions relating to porosity, a fluid contact model configured to answer questions relating to fluid contact, a permeability model configured to answer questions relating to permeability, a formation testing model configured to answer questions relating to formation testing, and a continuity model configured to answer questions relating to continuity (paragraph 165: “The system would be able to use large language models (like ChatGPT) or other AI/ML tools to be interactive, such as by accepting users' input questions and then outputting the short video clips with answers to the questions, such as by providing additional information, showing backup information, or providing a more detailed explanation. An example would be the drilling engineer could ask “how does this current H2-MB14 well intermediate section and curve compared to the last three MB14 wells we drilled, include trajectories, bit and BHA's, hours on each of those during that time, geological formations and faults, mud type and mud additives added at this stage, mud properties?” The large language models can take that input, output the correct language response, and then the video reporting tool would generate a video report including the audio narration of the answer, together with the 3D representation of the well, with the display of the current well provided with displays of the earlier three reference wells so that the video report visually shows the similarities and differences of the current well and the three earlier reference wells that were the subject of the question. The engineer could then refine the request after reviewing the generated video report and say “now take wells MB14 and X well and also compare the dip angles on each well as well as direction the wells were drilled”. The large language model would then provide a response, which the video reporting engine can use to provide an updated report with updated text and/or audio narration including the response to the second input query”).
Regarding Claim 10, Walsh further teaches:
The method of claim 1, further comprising: based on the query, performing, by each of the one or more domain-specific language models, a respective lookup in a knowledge base (paragraph 22: “The domain specific language model may obtain the set of documents from the private knowledge database based on the set of documents being associated with the topic”);
and generating, by the one or more domain-specific language models, the one or more responses based on each respective lookup in the knowledge base (paragraph 23: “a domain specific language model, of the group of domain specific language models, may generate a set of knowledge space embeddings based on the private documents”; And, paragraph 51: “a user may input (e.g., via a user interface associated with the collaborative knowledge system) a query. The query may indicate a question posed by the user. The collaborative knowledge system may provide the query to the centralized model as an input. The centralized model may process the query to determine a response”).
And Michalopulos further teaches:
of petrophysics information
of petrophysics information (paragraph 200: “The content of a drilling plan can vary depending on factors such as the type of well-being drilled (oil, gas, geothermal, etc.), the geological conditions”).
Claims 11-16 are similar to Claims 1-6 and are rejected under the same rationale as stated above for those claims.
Claim 17 is a combination of Claims 7 and 8 and is rejected under the same rationale as stated above for those claims.
Claims 18-19 are similar to Claims 9-10 and are rejected under the same rationale as stated above for those claims.
Claim 20 is similar to Claim 1 and is rejected under the same rationale as stated above for that claim.
Examiner's Note:
The Examiner cites particular pages, sections, columns, line numbers, and/or paragraphs in the references as applied to the claims above for the convenience of the applicant. Although the specified citations are representative of the teachings in the art and are applied to the specific limitations within the individual claim, other passages and figures may apply as well. It is respectfully requested that, in preparing responses, the applicant fully consider the references in its entirety as potentially teaching all or part of the claimed invention, as well as the context of the passage as taught by the prior art or disclosed by the examiner and the additional related prior arts made of record that are considered pertinent to applicant's disclosure to further show the general state of the art. The Examiner's interpretations in parenthesis are provided with the cited references to assist the applicants to better understand how the examiner interprets the prior art to read on the claims. Such comments are entirely consistent with the intent and spirit of compact prosecution.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. See PTO-892 for the relevant prior art where for example Bhat, US 2025/0005276 A1, teaches large language models (LLMs) used in various complex natural language processing (NLP) tasks in a variety of applications.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DAVE MISIR whose telephone number is (571)272-5243. The examiner can normally be reached M-R 8-5 pm, F some hours.
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, Abdullah Al Kawsar can be reached at 5712703169. 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.
/DAVE MISIR/Primary Examiner, Art Unit 2127
Prosecution Insights