DETAILED ACTION
This Final Office action is in response to Applicant’s Amendment filed on 05/26/2026. Claims 1, 2, 4, 6-12, 14, and 16-25 are pending. The effective filing date of the claimed invention is 04/10/2024.
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 § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1, 2, 4, 6-12, 14, 16-25 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
The term “optimize” in claim 1 is a relative term which renders the claim indefinite. The term “optimize” is not defined by the claim, the specification does not provide a standard for ascertaining the requisite degree, and one of ordinary skill in the art would not be reasonably apprised of the scope of the invention. Appropriate correction is required.
Claim 1 line 19 (and similarly claims 11 and 25) recites “a predicted processed output” and line 25-26 recites “a predicted level of processed output.” This renders the claim indefinite. First, it is unclear based on the lack of antecedent basis if these are the same. Assuming they are the same, and the “level” is a level within the previously claimed predicted processed output, the examiner recommends amending claim 1, lines 25-26 to include “a level within the predicted processed output” if that is what is intended in the claim language. If this is not intended and these are different, then different wording needs to be used. Appropriate correction is required.
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, 4, 6-12, 14, 16-25 are rejected under 35 U.S.C. 101 because the claims are directed to abstract idea without significantly more.
Step 1 – Claims 1, 2, 4, 6-10, 21-25 relate to machine claims; and, claims 11, 12, 14, 16-20 relate to process claims. Claims 1, 2, 4, 6-12, 14, 16-25 satisfy Step 1.
Step 2A, Prong 1 – Exemplary claim 11 (and similarly claim 1, 25) recites the following abstract idea:
A computer implemented method for managing a processing system, the method comprising:
Access, 1 (see e.g. MPEP 2106.04(a)(2)(III)(C)(2) performing a mental process in a computer environment citing Fairwarning; see also MPEP 2106.04(a)(2)(II)(A));
monitoring, (see e.g. MPEP 2106.04(a)(2)(III)(D) citing Electric Power Grp.);
generating, more of: operation of the D/H, the processed output produced, or the feedstock input (see MPEP 2106.04(a)(2)(I)(A) citing Digitech; see also MPEP 2106.04(a)(2)(II)(A)); and
train, (See Recentive Analytics Inc. v. Fox Corp., Appeal No. 2023-2437 (Fed. Cir. 04/18/2025); MPEP 2106.04(a)(2)(I); July 2024 USPTO Subject Matter Eligibility Examples, Example 47, claim 2; for the filtering/modifying see MPEP 2106.04(a)(2)(II)(A-C) Other examples of managing personal behavior recited in a claim include: i. filtering content, BASCOM Global Internet v. AT&T Mobility, LLC, 827 F.3d 1341, 1345-46, 119 USPQ2d 1236, 1239 (Fed. Cir. 2016) (finding that filtering content was an abstract idea under step 2A));
define, (see MPEP 2106.04(a)(2)(III) pen and paper);
execute, (see MPEP 2106.04(a)(2)(I) execute model/algorithm; See e.g. Recentive Analytics Inc. v. Fox Corp., Appeal No. 2023-2437 (Fed. Cir. 04/18/2025); MPEP 2106.04(a)(2)(I); July 2024 USPTO Subject Matter Eligibility Examples, Example 47, claim 2.) configured to predict material need for the D/H and output the candidate operational parameters to optimize operation of the D/H to achieve a predicted processed output (see e.g. MPEP 2106.04(a)(2)(II)(A) citing Bancorp, The court described the claims as an “attempt to patent the use of the abstract idea of [managing a stable value protected life insurance policy] and then instruct the use of well-known [calculations] to help establish some of the inputs into the equation.” 687 F.3d at 1278, 103 USPQ2d at 1433 (alterations in original) (citing Bilski).);
executing, by the at least one processor, a second machine learning model (see MPEP 2106.04(a)(2)(I) execute model/algorithm; See e.g. Recentive Analytics Inc. v. Fox Corp., Appeal No. 2023-2437 (Fed. Cir. 04/18/2025); MPEP 2106.04(a)(2)(I); July 2024 USPTO Subject Matter Eligibility Examples, Example 47, claim 2.) configured to predict an optimal distribution schedule for transportation, routing, and allocation of feedstock sources based on a factor/quality to a plurality of locations of the one or more D/H components responsive to generation of the candidate operational parameters by the first machine learning model (see e.g. MPEP 2106.04(a)(2)(II)(A) citing OIP Techs, Inc. v. Amazon.com, Inc., 788 F.3d 1359, 1364 (Fed Cir. 2015) (a new method of price optimization was found to be a fundamental economic concept), where the process of optimizing known variables is similar to both the present claim set (predicting an optimal distribution) and to the abstract idea of OIP (a new method of price optimization based on known variables));
control, (see e.g. MPEP 2106.04(a)(2)(II)(B) ii. using an algorithm for determining the optimal number of visits by a business representative to a client, In re Maucorps, 609 F.2d 481, 485, 203 USPQ 812, 816 (CCPA 1979) where the claim only requires the calculating of data to reach some target output, such as an algorithm that determines an optimal number of visits by a business representative; see also MPEP 2106.04(a)(2)(II)(C) citing Interval Licensing, The patentee claimed an attention manager for acquiring content from an information source, controlling the timing of the display of acquired content, displaying the content, and acquiring an updated version of the previously-acquired content when the information source updates its content. 896 F.3d at 1339-40, 127 USPQ2d at 1555. The Federal Circuit concluded that “[s]tanding alone, the act of providing someone an additional set of information without disrupting the ongoing provision of an initial set of information is an abstract idea,” observing that the district court “pointed to the nontechnical human activity of passing a note to a person who is in the middle of a meeting or conversation as further illustrating the basic, longstanding practice that is the focus of the [patent ineligible] claimed invention.” 896 F.3d at 1344-45, 127 USPQ2d at 1559.).
When these limitations are viewed alone and in ordered combination, the examiner finds that claim 11 (and similarly claim 1, 25) recite abstract idea.
Step 2A, Prong 2 –
Exemplary claim 1 (and similarly claim 11, 25) do not integrate the recited abstract idea with practical application. Claim 1 recites the additional limitations of “at least one processor operatively connected to a memory, the at least one processor configured to [implement the recited abstract idea]; accessing operational parameters; and control operation of the D/H component to execute a minimum level of output. In general, the additional limitation(s) are recited at a very high level of generality and act as tools to implement the recited abstract idea. For the claimed “at least one processor” the examiner refers to MPEP 2106.05(f), apply it rationale. For the accessing of the parameters, the examiner refers to MPEP 2106.05(f)(1)(i) Other examples where the courts have found the additional elements to be mere instructions to apply an exception, because they recite no more than an idea of a solution or outcome include: i. Remotely accessing user-specific information through a mobile interface and pointers to retrieve the information without any description of how the mobile interface and pointers accomplish the result of retrieving previously inaccessible information, Intellectual Ventures v. Erie Indem. Co., 850 F.3d 1315, 1331, 121 USPQ2d 1928, 1939 (Fed. Cir. 2017).
For the monitoring limitations, examiner refers to MPEP 2106.05(h) field of use and technological environment, iv. Specifying that the abstract idea of monitoring audit log data relates to transactions or activities that are executed in a computer environment, because this requirement merely limits the claims to the computer field, i.e., to execution on a generic computer, FairWarning v. Iatric Sys., 839 F.3d 1089, 1094-95, 120 USPQ2d 1293, 1295 (Fed. Cir. 2016).
For the controlling operations limitation(s), the examiner refers to Interval Licensing LLC, v. AOL, Inc., 896 F.3d 1335, 127 USPQ2d 1553 (Fed. Cir. 2018), where such generic controlling instructions provided by a computer were found to “generic sets of instructions” and not found to be an improvement to the technology. Further, the examiner refers to MPEP 2106.05(g)(3)(mere data gathering, citing to OIP Tech. Further for the control limitation, the examiner refers to Applicant’s Remarks, 01/26/2026, page 8, where Applicant states the following:
PNG
media_image1.png
218
663
media_image1.png
Greyscale
Importantly, Applicant states that “it stands to reason that the POSA would then use those operational parameters as stated in the originally filed claimed to then “control operation.” Applicant admits that there is no explicit support for the specifics of the control operation limitation, and that a POSA would understand that sending a communication and then controlling based on that would be well known to POSA. This is similar to apply it, where a processor is used to implement the abstract idea relating to communicating data and notifications based on gathered data. See MPEP 2106.05(f).
As for the first machine learning model and the second machine learning model, the examiner has addressed under Step 2A Prong 1. Under prong 2, there is no improvement to the underlying machine learning/AI models and/or training. The claims recite executing the models, which is similar to “apply it” rationale. Simply applying/executing the models to reach a result. See MPEP 2106.05(f).
When these additional limitations are viewed alone and in ordered combination, the examiner finds that the claims are directed to abstract idea.
Step 2B –
Exemplary claim 1 (and similarly claim 11, 25) do not recite significantly more. The additional element analysis from Step 2A Prong 2 is equally applied to Step 2B. Another consideration when determining whether a claim recites significantly more than a judicial exception is whether the additional element(s) are well-understood, routine, conventional activities previously known to the industry. This consideration is only evaluated in Step 2B of the eligibility analysis. See MPEP 2106.05(d).
The courts have recognized the following computer functions as well‐understood, routine, and conventional (WURC) functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity.
For the controlling step, while not provided with any real support, could just be sending an instruction from the processor to the one or more DH components, where this has been found to be well-understood, routine, and convention at MPEP 2106.05(d)(II)(i) i. Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); TLI Communications LLC v. AV Auto. LLC, 823 F.3d 607, 610, 118 USPQ2d 1744, 1745 (Fed. Cir. 2016) (using a telephone for image transmission); OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network); but see DDR Holdings, LLC v. Hotels.com, L.P., 773 F.3d 1245, 1258, 113 USPQ2d 1097, 1106 (Fed. Cir. 2014) (“Unlike the claims in Ultramercial, the claims at issue here specify how interactions with the Internet are manipulated to yield a desired result‐‐a result that overrides the routine and conventional sequence of events ordinarily triggered by the click of a hyperlink.” (emphasis added)). There is no discussion in Applicant’s Specification nor the claims that overrides the routine and conventional sequence of events when a processor sends an instruction over a network. Further, in regards to the WURC analysis, the fact that there is no written description (except for the original claim 1 and 11) of the claimed “control operation,” this is further Berkheimer evidence that the limitation of controlling operations is well-understood, routine, and conventional. See MPEP 2106.05(d)(I)(2) where when the Specification is silent on something, this is further Berkheimer evidence that the limitation is likely WURC. Again, Applicant is silent regarding how the controlling operations is being performed other than do it on a computer.
For the accessing operational parameters, see the following WURC finding: iv. Storing and retrieving information in memory, Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015); OIP Techs., 788 F.3d at 1363, 115 USPQ2d at 1092-93.
For the monitoring steps, see the following WURC finding: ii. Performing repetitive calculations, Flook, 437 U.S. at 594, 198 USPQ2d at 199 (recomputing or readjusting alarm limit values); Bancorp Services v. Sun Life, 687 F.3d 1266, 1278, 103 USPQ2d 1425, 1433 (Fed. Cir. 2012) (“The computer required by some of Bancorp’s claims is employed only for its most basic function, the performance of repetitive calculations, and as such does not impose meaningful limits on the scope of those claims.”); iii. Electronic recordkeeping, Alice Corp. Pty. Ltd. v. CLS Bank Int'l, 573 U.S. 208, 225, 110 USPQ2d 1984 (2014) (creating and maintaining “shadow accounts”); Ultramercial, 772 F.3d at 716, 112 USPQ2d at 1755 (updating an activity log);
For the control step that is based on communication, see the following WURC finding: i. Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); TLI Communications LLC v. AV Auto. LLC, 823 F.3d 607, 610, 118 USPQ2d 1744, 1745 (Fed. Cir. 2016) (using a telephone for image transmission); OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network); but see DDR Holdings, LLC v. Hotels.com, L.P., 773 F.3d 1245, 1258, 113 USPQ2d 1097, 1106 (Fed. Cir. 2014) (“Unlike the claims in Ultramercial, the claims at issue here specify how interactions with the Internet are manipulated to yield a desired result‐‐a result that overrides the routine and conventional sequence of events ordinarily triggered by the click of a hyperlink.” (emphasis added)).
When these additional limitations are viewed alone and in ordered combination, the examiner finds that the claims are directed to abstract idea.
Dependent claims – Claims 2 and 12 recite more abstract idea. See MPEP 2106.04(a)(2)(I)(C) Examples of mathematical calculations recited in a claim include: i. performing a resampled statistical analysis to generate a resampled distribution, SAP America, Inc. v. InvestPic, LLC, 898 F.3d 1161, 1163-65, 127 USPQ2d 1597, 1598-1600 (Fed. Cir. 2018). Claims 4, 6, 8, 10, 14, 16, 18, and 20 recites more abstract idea. See e.g. Recentive Analytics Inc. v. Fox Corp., Appeal No. 2023-2437 (Fed. Cir. 04/18/2025); MPEP 2106.04(a)(2)(I); July 2024 USPTO Subject Matter Eligibility Examples, Example 47, claim 2. Claims 7 and 17 recite a set of sensors that monitor internal operating parameters of the D/H component, where this is using sensors in an “apply it” manner (see MPEP 2106.05(f)). See also Yu v. Apple Inc., Appeal No. 2020-1760 and 1803 (Fed. Cir. 06/11/2021). Claims 9 and 19 recite more abstract idea. See MPEP 2106.04(a)(2)(I) citing Digitech Image Techs., LLC v. Elecs. for Imaging, Inc., 758 F.3d 1344, 1350, 111 USPQ2d 1717, 1721 (Fed. Cir. 2014) (holding that claims to a ‘‘process of organizing information through mathematical correlations’’ are directed to an abstract idea).
For the sensor limitations, the examiner recommends attempting to amend the claims into the facts of Thales, Examples that the courts have indicated may be sufficient to show an improvement in existing technology include:
MPEP 2106.05(a)(II)(vii) Particular configuration of inertial sensors and a particular method of using the raw data from the sensors, Thales Visionix, Inc. v. United States, 850 F.3d 1343, 1348-49, 121 USPQ2d 1898, 1902 (Fed. Cir. 2017).
For claims 21-23, these are also abstract idea concepts. See MPEP 2106.05(f) citing IV I, The claims were found to be directed to the abstract idea of “collecting, displaying, and manipulating data.” 850 F.3d at 1340; MPEP 2106.04(a)(2)(II)(A); See also WURC analysis at MPEP 2106.05(d)(II) ii. Performing repetitive calculations, Flook, 437 U.S. at 594, 198 USPQ2d at 199 (recomputing or readjusting alarm limit values). For claim 24, this is the abstract idea of converting input to output, using the mature industry known D/H in “apply it” manner. 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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1, 2, 4, 6-12, 14, and 16-25 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Pat. Pub. No. 2020/0074307 to Kent et al. (“Kent”) in view of U.S. Pat. Pub. No. 2022/0292434 to Khadivi et al. (“Khadivi”) in view of Sun et al. 05/18/2023). Prediction of biogas production in anaerobic digestion of a full-scale wastewater treatment plant using ensembled machine learning models. Wiley Online library, retrieved from https://pubmed.ncbi.nlm.nih.gov/37202836/ (referred to as “Sun”).
With regard to claims 1 and 11, 21, 22, 23, 24, 25, Kent discloses the claimed processing system comprising exemplary claim 1:
at least one processor operatively connected to a memory, the at least one processor (see e.g. [0054], [0072] etc.) configured to:
access operational parameters for a digester/hydrolyzer ("D/H"), the D/H configured to accept a feedstock input and generate a processed output (see e.g. Fig. 1, where get input parameter values accessing such parameters; see input/outputs at Fig. 2; inputs - Fig. 5 54a, b, c, d; outputs at e.g. [0023] [0051] [0061-69] etc.);
monitor the processed output produced by the D/H (see e.g. [0070] In other alternative embodiments, the system 10 may be directly interfaced with the plant controller 30 and the AD plant (not shown) such that the system 10 receives real-time data about the operation of the AD plant. This real-time data may be obtained from various sensors that are used at the AD plant. Sensors may include but are not limited to, biogas production sensors, mass flow rate sensors, electricity generation sensors, feedstock and digestate characteristic sensors.);
generate candidate operational parameters, including distribution and allocation of feedstock sources based on emulation/prediction2 of one or more of: operation of the D/H, the processed output produced, or the feedstock input (see e.g. [0070] The system 10 then performs simulations to achieve or maintain certain optimal target operating conditions, which may include generating updated input material settings and/or operational process settings to achieve the optimal target operating conditions. The system 10 then sends the input material settings and/or the operational process settings, which may be updated, to the plant controller 30 which then controls the operation of the AD plant to achieve the optimal target operating conditions. The measurements from the AD plant and simulations to generate the settings can be done periodically such as, but not limited to, every 0.5, 1, 4, 8, 16 or 24 hours, for example, in order to maintain the optimal target operating conditions.); and
train a first machine learning model on D/H material need, the first machine learning model based on (see e.g. [0064] [0194] [0222]; Kent does not disclose where the first model is based on gradient boosting architecture. See Sun below);
define the set of training data to include at least a set of top contributing parameters in captured historical data (see e.g. [0194] where the defined set of training data could be “if the error is too high” and/or the fitness function gets tuned over time based “the constant factors” where the constant factors could be the set of top contributing parameters, as claimed);
execute the first machine learning model configured to predict material need for the D/H and output the candidate operational parameters to optimize operation of the D/H to achieve a predicted processed output (see e.g. [0017] [0019] [0030] [0032] [0042] Fig. 8, etc; [0019] In at least one embodiment, the simulation engine is further configured to operate in an online simulation mode where a machine learning model is used to simulate the operation of the AD plant to generate online simulator predictions, compare the online simulator predictions to actual results from the AD plant and to send the plant controller subsequent recipes for the optimal operation of the AD plant for subsequent time intervals.);
executing, by the at least one processor, a second machine learning model configured to predict a distribution schedule for transportation, routing, and allocation of feedstock sources based on quality and volume/mass to a plurality of locations of the one or more D/H components responsive to generation and input of the candidate operational parameters from the first machine learning model (see also claim 23) (see e.g. [0193] where each time the training occurs a new machine learning model is created, and the new machine learning model is based on the first machine learning model as it is a trained version of it; [0222], each responsive to the last and so on, as the model is continuously trained and changed based on the parameters etc.; for the based on quality/volume/mass aspect and claim 23, see e.g. [0094-95] quality and [0096] volume [0161-162]);
control operation of the D/H to execute according to a predicted level of processed output defined by the candidate parameters and the schedule for transportation, routing, and allocation (see e.g. [0011], [0013] In at least one embodiment, the at least one optimization goal comprises maximizing biogas production, maximizing electricity production, minimizing greenhouse gas emissions and minimizing feedstock leftover or a weighted combination of those options.; see e.g. Fig. 2 60 [Wingdings font/0xE0] 62
PNG
media_image2.png
226
271
media_image2.png
Greyscale
; see [0107] [0069] [0070] where the controlling of the AD plant can be done automatically where the server is directly connected to the plant controller 30 and various communications between the plant and the system 10).
Kent may not expressly disclose where the second machine learning model predicts “optimal” distribution schedule for transport route allocate of items to a plurality of locations; and where the ML model is based on gradient boosting architecture.
Khadivi teaches at e.g. [0034], [0038], [0042] that it would have been obvious to one of ordinary skill in the resource planning for delivery of goods art before the effective filing date of the claimed invention the limitation of a machine learning model predicts optimal distribution schedule for transport route allocate of items to a plurality of locations, as taught by Khadivi, where the combination of Khadivi with Kent solves one of the challenges in logistics for delivering goods to multiple locations utilizing multiple vehicles, drivers, and determining the best route and distribution for them to take to delivery said goods to said plurality of location(s), where the savings may arise reducing staff expenses, reducing the number of resources used in the distribution, or reducing the distance traveled by the drivers to deliver the goods. See Kent, [0002-3] problems addressed by said machine learning system. Further, for claims 21-22, see Khadavi at e.g. [0065] where the parameters can be adjusted and then used in the output, where the benefit in this combination of Khadavi with Kent, allows the users to tune the system for trade-offs (including and not limited to greenhouse gases) between various parameters and situations.
As for the use of specific machine learning models in the anaerobic disgestion plant area, see Sun.
Abstract
Anaerobic digestion (AD) of sludge is a key approach to recover useful bioenergy from wastewater treatment and its stable operation is important to a wastewater treatment plant (WWTP). Because of various biochemical processes that are not fully understood, AD operation can be affected by many parameters and thus modeling AD processes becomes a useful tool for monitoring and controlling their operation. In this case study, a robust AD model for predicting biogas production was developed using ensembled machine learning (ML) model based on the data from a full-scale WWTP. Eight ML models were examined for predicting biogas production and three of them were selected as metamodels to create a voting model. This voting model had a coefficient of determination (R2 ) at 0.778 and a root mean square error (RMSE) of 0.306, outperformed individual ML models. The Shapley additive explanation (SHAP) analysis revealed that returning activated sludge and temperature of wastewater influent were important features, although they affected biogas production in different ways. The results of this study have demonstrated the feasibility of using ML models for predicting biogas production in the absence of high-quality data input and improving model prediction through assembling a voting model. PRACTITIONER POINTS: Machine learning is applied to model biogas production from anaerobic digesters at a full-scale wastewater treatment plant. A voting model is created from selected individual models and exhibits better performance of predication. In the absence of high quality data, indirect features are identified to be important to predicting biogas production.
Machine learning (ML) has emerged as a promising computational tool for developing alternative models for ADs.
Therefore, there is a need for investigating ML models that can simulate process performance with limited data or gaps in data parameters.
Model description
After feature selection, all input features were normalized using Min-Max scaling to improve the performance and stability of the ML models by reducing the effect of differences in the scales of input features. The present study employed eight ML models, namely ANN (Sun et al., 2022), Light Gradient-Boosting Machine (LightGBM) (Jin et al., 2020), XGBoost (Chen & Guestrin, 2016), Gradient Boosting Decision Tree (GBDT) (Kumar Ganti et al., 2022), RF (Li et al., 2018), KNN (Song et al., 2017), support vector machine (SVM) (Xu et al., 2022), Stochastic gradient descent (SGD) (Ighalo et al., 2020), and voting algorithms (Atallah & Al-Mousa, 2019), to build predictive models. A detailed introduction to each of these models can be found in the supporting information.
PNG
media_image3.png
798
725
media_image3.png
Greyscale
Model selection
The results of R2 and RMSE for eight models are shown in Figure 4. One can see that XGBoost, GBDT, and RF exhibited good performance on the training data with R2 of 0.999, 0.979, and 0.883, respectively, but such performance was not maintained on the testing data, indicating the problems with generalization of the models trained using the default hyperparameters. Therefore, there is a need for further hyperparameters optimization to improve the performance of those ML models. Among the eight models, GBDT, LightGBM, and XGBoost models had R2 above 0.5 with both the training and testing sets. Thus, those three models were chosen for further hyperparameters optimization, which is presented in Table 1. Figure 5a–c show the performance of each model with the optimal hyperparameters that were determined using the grid search technique.
Therefore, it would have been obvious to one of ordinary skill in the AD art before the effective filing date to modify Kent to include such specific ML models such as the gradient boosting architecture, which was used tested and shown to work well in and throughout Sun. Therefore, it would have been obvious to make this modification because Sun identified there is a need for investigating ML models that can simulate process performance with limited data or gaps in data parameters, and higher prediction accuracy. The results would help create alternative modeling solutions to traditional models, especially for WWTPs that do not regularly or accurately monitor all the key features. The use of an ensembled ML algorithm to model complex AD systems was effective because it reduces the prediction error of individual models. Based on cross-validation and grid search to validate the hyperparameters of the model, the overfitting problem of the model was reduced. It was shown that returning activated sludge, as an indirect feature for AD, had a positive effect on biogas and a lower temperature of influent wastewater could increase the accumulation of organic matter that led to more biogas production from AD. See Sun, Conclusions.
With regard to claims 2 and 12, Kent further discloses where the emulation includes emulating/predicting physical properties of the processed output produced, physical properties of the feedstock input, operational parameters associated with the D/H component (see e.g. [0006] [0020] [0064] [0067] etc.).
With regard to claims 4 and 14, Kent further discloses where the first machine learning model is trained on material consumption and processed output data, and once trained the first machine learning model is configured to predict an anticipated material need for one or more D/H components having one or more locations (see e.g. [0064] [0194] [0222]).
With regard to claims 6 and 16, Kent further discloses where the second machine learning model is trained on material need and resource utilization for distribution, and once trained the second machine learning model is configured to predict the optimal distribution schedule upon input of a predicted material need for one or more D/H components having one or more locations (see e.g. [0222] [02225]).
With regard to claims 7 and 17, Kent further discloses a set of sensors configured to monitor internal operating parameters of the D/H component (see [0220]).
With regard to claims 8 and 18, Kent further discloses where the system is configured to update training of one or more of the first or second machine learning models with data returned from the set of sensors (see e.g. [0221]).
With regard to claims 9 and 19, Kent further discloses where the system is configured to correlate external parameters with data from the set of sensors (see e.g. [0221]).
With regard to claims 10 and 20, Kent further discloses where the system is configured to update training of one or more of the first or second machine learning models with data returned from the set of sensors and the external parameters (see e.g. [0221]).
Response to Arguments
Applicant’s arguments with respect to the claims have been considered but are not persuasive.
The examiner has withdrawn the previously made 112 rejections.
Applicant argues that the claims recite eligible subject matter. The examiner respectfully disagrees. The examiner has amended the 101 rejection above to cover the amended limitations, and previous limitations. The examiner does not find anything in the claims that is eligible, and cannot recommend anything for eligibility currently.
Applicant argues that the claims are distinguished over the prior art. The examiner respectfully disagrees. The examiner has found some very solid NPL documents that are attached. The one cited to, Sun, is very much on point regarding the amendments and the specific recitation of the gradient boost architecture usage. Please refer to Sun as referenced above. The other NPL documents attached are also relevant, but not currently needed for the rejections.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Peter Ludwig whose telephone number is (571)270-5599. The examiner can normally be reached Mon-Fri 9-5.
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, Fahd Obeid can be reached at 571-270-3324. 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.
/PETER LUDWIG/Primary Examiner, Art Unit 3627
1 Strike-through language is addressed in Step 2A Prong 2; Step 2B.
2 See Applicant’s originally-filed Specification at page 13, lines 30-31, “emulate/predict”. Accordingly, the terms emulate and predict are found to, under the BRI in light of the specification, equate in meaning.
Prosecution Insights