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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 06/02/2026 has been entered.
Status of the Claims
Claims 1, 2, 8-10, and 16-18 have been amended. Claims 3, 4, 11, 12, 19, and 20 have been canceled. Claims 21-26 are added as new claims. Claims 1, 2, 5-10, 13-18, and 21-26 are pending.
Response to Arguments
Applicant's arguments filed 04/30/2026 regarding 35 U.S.C. 101 have been fully considered but they are not persuasive.
Applicant argues that claim 1 requires formulating a reinforcement learning framework utilizing a “neural network architecture that comprises a recurrent layer, one or more fully connected layers, and an output later…” and that the claims requires training the architecture by “iteratively adjusting one or more weights of nodes within the recurrent layer or the one or more fully connected layers by performing gradient-based optimization”, arguing that “a human mind cannot physically or practically perform gradient based optimization” because these are inherently computer-implemented operations performed on a non-human data architecture. Examiner disagrees, and this argument is unpersuasive. The relevant inquiry under Step 2A Prong One is not whether a human can perform the claimed steps with the same speed, scale, or numerical precision as a computer, but whether the claim limitations, under their broadest reasonable interpretation, cover performance of the limitation in the mind (observation, evaluation, judgment, opinion) or as a mathematical concept, but for the recitation of generic computer components. See MPEP § 2106.04(a)(2)(III). Applicant’s “no human could do this” framing conflates the means of computation with the character of the underlying step. The claims recite a judgment-based evaluation: determining an action of sending an inspection alert or not, by the policy model, based on the state representation; modeling one or more system behaviors in response to the action using preprocessed flight data; calculating a reward for the action under the state representation using a predefined reward structure; etc., and utilizing mathematical concepts: training the policy model using a learning and optimization algorithm with the training data to increase an expected discounted cumulative rewards, wherein the training policy model includes iteratively adjusting one or more weights of nodes within the recurrent layer or the one or more fully connected layers by performing gradient-based optimization to reduce an error between Q value and an optimal Q values, etc., said limitations implemented using a computer, neural network. The claims fall within the enumerated abstract idea groupings regardless of the scale at which a computer performs them. Applicant is reminded that "claiming the improved speed or efficiency inherent with applying the abstract idea on a computer" does not integrate a judicial exception into a practical application or provide an inventive concept.” Intellectual Ventures I LLC v. Capital One Bank (USA), 792 F.3d 1363, 1367, 115 USPQ2d 1636, 1639 (Fed. Cir. 2015); see also MPEP §2106.05(f). As stated in examiner’s previous response, the Federal Circuit has explained that "the 'directed to' inquiry applies a stage-one filter to claims, considered in light of the specification, based on whether 'their character as a whole is directed to excluded subject matter."' Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1335 (Fed. Cir. 2016) (quoting Internet Patents Corp. v. Active Network, Inc., 790 F.3d 1343, 1346 (Fed. Cir. 2015)). It asks whether the focus of the claims is on a specific improvement in relevant technology or on a process that itself qualifies as an "abstract idea" for which computers are invoked merely as a tool. Here it is clear from the specification in light of the claim language that the claims focus on an abstract idea and not an improvement to technology and/or a technical field. For instance, the specification observes in the Background section that Airplane service organizations analyze the collected sensor data, along with flight records, to predict potential component failures. Upon detecting a possible failure, the airplane service organization promptly sends an alert to the affected airline company. In response, the airline company conducts an inspection on the identified component to verify the problem, and, if confirmed, takes corrective actions, such as replacing the component. The action is intended to prevent any compromise in the aircraft’s safety and performance. Further, in the Detailed Description, the invention provides that [0013] “…Traditional approaches to making these decisions often rely on conventional data analysis techniques and manual engineering reviews, which typically require substantial human effort. Furthermore, these traditional methods do not perform well in capturing patterns within flight records and sensor data, especially given the large data volume and the presence of noise. As a result, this can lead to suboptimal alert decisions”. [0014] “To address these issues, the present disclosure introduces techniques that leverage deep reinforcement learning to automate the airplane component failure prognostic process”. The cited portions of the specification further supports the examiner’s position that the claimed invention merely used a computer (including deep reinforcement learning programmed on the computer) to automate a process that humans would typically do manually. This amounts to “apply it” or merely using a computer as a tool to implement the judicial exception. The alleged improvement is, at best, an improvement in the judicial exception itself, and not an improvement computers ore technology. Computers and deep learning are “leveraged” in order to automate the failure prognostic process, as stated in applicants disclosure. Technical improvement focuses on enhancing the tools, software, or machinery, while business process improvement focuses on streamlining the steps, workflows, and methodologies people use to do their work. Applicant’s claims fall in the latter. Claims 1, 9, and 17 gives the detailed limitations of the airplane failure prognostic process and recites limitations that details deep reinforcement learning techniques to automate detecting airplane component failures and the claim limitations amount to mental processes (observation, evaluation, judgment, opinion) as evidenced by the limitations detailing identifying operational characteristics, defining state representation comprising one or more parameters, determining an action to send an alert or not based on the state representation, modeling behaviors in response, calculating a reward for the action, collecting training data by simulating airplane component failure, and training the policy model using a learning and optimization algorithm to increase an expected discount cumulative reward. The limitations directly involve observing and evaluating data in order to make a determination (judgment/opinion) based on the observed and evaluated data. The use of computing techniques to assist in these steps do not take the claims out of the mental processes grouping. The claims further correspond to mathematical concepts (mathematical relationships, mathematical formulas or equations, mathematical calculations) as evidenced by the claim limitations disclosing training the policy model using a learning and optimization algorithm with the training data to increase an expected discounted cumulative reward, wherein training the policy model includes iteratively adjusting one or more weights of nodes within the recurrent layer or the one or more fully connected layers by performing gradient- based optimization to reduce an error between an estimated Q value and an optimal Q value. The claims in light of the specification provides the reinforcement learning framework and policy models as optimization algorithms and the use of algorithms (mathematical equations, formulas/equations, relationships). The claims recite an abstract idea. It is important to note, the judicial exception alone cannot provide the improvement. See MPEP §2106.05(a). It is important to keep in mind that an improvement in the judicial exception itself is not an improvement in technology (emphasis added). For example, in Trading Technologies Int’l v. IBG LLC, the court determined that the claim simply provided a trader with more information to facilitate market trades, which improved the business process of market trading but did not improve computers or technology. Similarly, the Applicant’s claim recitations are an improvement in the judicial exception, not an improvement in technology. Utilizing general computer components that processes machine learning techniques for airplane component prognostics which ultimately impacts aircraft safety and maintenance since, upon detecting a possible failure, the airplane service organization can promptly send an alert to the affected airline company, does not constitute an improvement in computers or technology. Instead, the improvement is at best in the business process and judicial exception itself.
Applicant further argues that the claimed system receives downstream, real-world, component statuses resulting from its alerts and uses that data to iteratively update and retune its own system models. Applicant argues that this creates a self-updating dynamic technological framework representing a specific technical improvement to the functioning of the computer-implemented system itself. Examiner disagrees. Applicant’s argument regarding real-world component statuses is unpersuasive. This an argument in utility, not patent eligibility. Eligibility (subject matter eligibility) asks "what" is being claimed; whether the invention falls into a category that the patent system protects. Utility asks "if" the invention works and provides a specific, substantial, and credible benefit. Applicant’s utility argument is unpersuasive in arguing eligibility. Further, applicant’s argument regarding iteratively updating and retuning its system models is also unpersuasive. First, applicant’s claimed invention leverages generic machine learning (deep reinforcement learning) in an airplane component prognostics, and does not provide an improvement to machine learning technology. "[P]atents that do no more than claim the application of generic machine learning to new data environments, without disclosing improvements to the machine learning models to be applied, are patent ineligible under § 101." Recentive Analytics, Inc. v. Fox. Corp., Fed Cir. No. 2023-2437 (Apr. 18, 2025). Additionally, "The requirements that the machine learning model be 'iteratively trained' or dynamically adjusted in the Machine Learning Training patents do not represent a technological improvement." Id., slip op. at 12. The judicial exception is not integrated into a practical application simply because the claims recite the additional elements of: an airplane component, a reinforcement learning framework, a policy model comprising a neural network architecture, one or more system models, one or more memories (claim 9), one or more computer processors (claim 9), one or more non-transitory computer- readable media (claim 17), and a computer system (claim 17). The additional elements of a reinforcement learning framework, a policy model comprising a neural network architecture, one or more system models, one or more memories, one or more computer processors, one or more non-transitory computer- readable media, and a computer system are computer components recited at a high-level of generality performing the above mentioned limitations. The combination of the additional elements are no more than mere instructions to apply the judicial exception using a generic computer. Further, the airplane component, reinforcement learning framework, and policy model comprising a neural network architecture amount to generally linking the judicial exception to a particular field of use (prognostics of airplane component failures). Accordingly, in combination, 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 directed to an abstract idea.
Applicant’s argument under Step 2B is unpersuasive. Whether the additional elements amount to well-understood, routing, and conventional activity (“WURC”) is only one consideration under Step 2B, the absence of which does not automatically result in “significantly more” or patent eligibility. 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 amount to no more than mere instructions to apply the exception using a generic computer, and generally linking the judicial exception to a particular field of use. Mere instructions to apply an exception using a generic computer cannot provide an inventive concept. Thus, when viewed as an ordered combination, nothing in the claims add significantly more (i.e. an inventive concept) to the abstract idea. The claims are not patent eligible.
The 35 U.S.C. 101 rejection is maintained.
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, 2, 5-10, 13-18, and 21-26 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e. an abstract idea) without significantly more.
Claims 1, 2, 5-8, 21, and 22 recite a method (i.e. process), claims 9, 10, 13-16, 23, and 24 recite a system (i.e. machine), and claims 17, 18, 25, and 26 recite non-transitory computer-readable media (i.e. machine or article of manufacture). Therefore claims 1, 2, 5-10, 13-18, and 21-26 fall within one of the four statutory categories of invention.
Independent claims 1, 9, and 17 recite the limitations: preprocessing flight data to identify a plurality of parameters representing operational characteristics of an [airplane component]; formulating, based on the preprocessed flight data, a [reinforcement learning framework] for airplane component failure prognostics, comprising: defining a state representation as an input to a [policy model], wherein the state representation comprises one or more parameters from the plurality of parameters, and wherein the policy model comprises a [neural network] architecture that comprises a recurrent layer, one or more fully connected layers, and an output laver; determining an action of sending an inspection alert or not, by the [policy model], based on the state representation; modeling one or more system behaviors in response to the action using preprocessed flight data; calculating a reward for the action under the state representation using a predefined reward structure; an collecting training data by simulating an airplane component failure prognostic procedure training the [policy model] using a learning and optimization algorithm with the training data to increase an expected discounted cumulative reward by choosing the action under the state representation, wherein training the [policy model] includes iteratively adjusting one or more weights of nodes within the recurrent laver or the one or more fully connected layers by performing gradient-based optimization to reduce an error between an estimated Q value and an optimal Q value and increase a likelihood of selecting the action with a higher expected discounted cumulative reward under the state representation; and deploying the [policy model] in a real-time prognostic environment, wherein the deploying comprises: receiving state data representing real- time operational characteristics of the [airplane component]; outputting the action of sending the inspection alert for the [airplane component] based at least in part on the state data; collecting data on one or more responses from one or more airlines and one or more airline component statuses after outputting the action of sending the inspection alert; and iteratively adjusting internal parameters of one or more [system models] based on the collected data. The invention and claims are drawn towards using deep reinforcement learning techniques to automate detecting airplane component failures and the claim limitations amount to mental processes (observation, evaluation, judgment, opinion) as evidenced by the limitations detailing identifying operational characteristics, defining state representation comprising one or more parameters, determining an action to send an alert or not based on the state representation, modeling behaviors in response, calculating a reward for the action, collecting training data by simulating airplane component failure, and training the policy model using a learning and optimization algorithm to increase an expected discount cumulative reward. The limitations directly involve observing and evaluating data in order to make a determination (judgment/opinion) based on the observed and evaluated data. The use of computing techniques to assist in these steps do not take the claims out of the mental processes grouping (computing techniques/components discussed further below in Step 2A Prong Two). The claims further correspond to mathematical concepts (mathematical relationships, mathematical formulas or equations, mathematical calculations) as evidenced by the claim limitations disclosing training the policy model using a learning and optimization algorithm with the training data to increase an expected discounted cumulative reward, wherein training the policy model includes iteratively adjusting internal parameters of the policy model by performing gradient- based optimization to reduce an error between an estimated Q value and an optimal Q value. For instance, the claims in light of the specification provides the reinforcement learning framework and policy models as optimization algorithms and the use of algorithms (mathematical equations, formulas/equations, relationships). The claims recite an abstract idea.
Note: The features or elements in brackets in the above section are inserted for reading clarity, but are analyzed as “additional elements” under Step 2A Prong Two and Step 2B below.
The judicial exception is not integrated into a practical application simply because the claims recite the additional elements of: an airplane component, a reinforcement learning framework, a policy model comprising a neural network, one or more memories (claim 9), one or more computer processors (claim 9), one or more non-transitory computer-readable media (claim 17), and a computer system (claim 17). The additional elements of a reinforcement learning framework, a policy model comprising a neural network, one or more memories, one or more computer processors, one or more non-transitory computer-readable media, and a computer system are computer components recited at a high-level of generality performing the above-mentioned limitations. The combination of the additional elements are no more than mere instructions to apply the judicial exception using a generic computer. Further, the airplane component, reinforcement learning framework, and policy model comprising a neural network amount to generally linking the judicial exception to a particular field of use (prognostics of airplane component failures). Accordingly, in combination, 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 directed to an abstract idea.
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 amount to no more than mere instructions to apply the exception using a generic computer, and generally linking the judicial exception to a particular field of use. Mere instructions to apply an exception using a generic computer cannot provide an inventive concept. Thus, when viewed as an ordered combination, nothing in the claims add significantly more (i.e. an inventive concept) to the abstract idea. The claims are not patent eligible.
Dependent claims 21, 23, and 25 recites the limitation that the reinforcement learning framework is implemented through a Markov Decision Process (MDP). The limitation is further directed to the judicial exception grouping of mathematical concepts, said MDP being a mathematical framework used to model decision-making, and allows for the algorithms to optimize actions based on rewards and penalties. Claims recite an abstract idea.
Dependent claims 2, 5-8, 10, 13-16, 18, 22, 24, and 26 recite additional limitations that are further directed to the abstract idea analyzed in the rejected claims above. The claims also recite additional elements that have been analyzed in the rejected claims above. Thus, claims 2, 5-8, 10, 13-16, 18, 22, 24, and 26 are also rejected under 35 U.S.C. 101.
Allowable Subject Matter
Claims 1, 2, 5-10, 13-18, and 21-26 would be allowable if rewritten or amended to overcome the rejection(s) under 35 U.S.C. 101, set forth in this Office action.
The closest patent or patent application prior art reference(s) found that is relevant to the applicant’s invention includes Auerbach (2022/0363405), Chopra (2019/0147670), Ito (2022/0319240), and Moy (US 11,386,800). Auerbach discloses a system and method of monitoring health of an electric vertical take-off and landing vehicle including at least a flight component display. A computing device is communicative with a first sensor, and is configured to receive a first characteristic, analyze the first characteristic, and determine a condition of the at least a flight component as a function of the first characteristic. Chopra discloses real time streaming analytics for flight data processing and receives a plurality of data streams acquired for a respective parameter of a plurality of parameters indicating an operating condition of the aircraft, selects at least one data stream corresponding to at least one parameter of the respective plurality of parameters, selects a portion of data from the at least one data stream, compares the portion of data to a model determined for the at least one parameter based on historical data, determines that a failure has occurred or is likely to occur during operation of the aircraft based on the comparing, and transmits aircraft health monitoring information indicative of occurrence or likelihood of occurrence of the failure. Ito discloses a management system that acquires, as information related to components of an analysis device analyzing a performance of a vehicle or a specimen which is a portion of the vehicle, at least one of warning information indicating an event leading to a failure of the component, information related to a sensitivity or performance of the component, and information related to a life limit of the component from the analysis device; and a determination unit that uses a trained model which has been trained so as to output an efficient maintenance schedule in response to an input of the information related to the components of the analysis device and inputs the information acquired by the acquisition unit to the trained model to determine the maintenance schedule of the analysis device. Moy discloses a system for flight control of a vertical take-off and landing (VTOL) aircraft includes a flight simulator communicatively coupled to a VTOL aircraft, wherein the flight simulator is configured to generate a model for at least a flight component and a flight controller, wherein the flight controller is configured to receive the model for the at least a flight component, determine a command for the at least a flight component as a function of the model, and initiate the command for the at least a flight component. Neither reference, individually nor in combination, appears to disclose the amended limitations of the applicant’s invention, particularly: training the policy model including iteratively adjusting one or more weights of nodes within the recurrent laver or the one or more fully connected layers by performing gradient- based optimization to reduce an error between an estimated Q value and an optimal Q value, increasing a likelihood of selecting the action with a higher expected discounted cumulative reward under the state representation and, deploying the policy model in a real-time prognostic environment. The claims appear to overcome the prior art.
The closest non-patent literature prior art reference found that is relevant to the applicant’s invention incudes the publication “Prognostic and Health Management of Critical Aircraft Systems and Components: An Overview” (Fu, Avdelidis; 2023) which analyzes the current state of research advancements in prognostics for aircraft systems, with a specific focus on prominent algorithms and their practical applications and challenges. The reference does not appear to explicitly disclose the detailed limitations of the applicant’s invention. The claims appear to overcome the prior art.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DIONE N SIMPSON whose telephone number is (571)272-5513. The examiner can normally be reached M-F; 7:30 a.m.-4:30 p.m..
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, Sarah Monfeldt can be reached at (571) 270-1833. 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.
DIONE N. SIMPSON
Primary Examiner
Art Unit 3628
/DIONE N. SIMPSON/Primary Examiner, Art Unit 3629
Prosecution Insights