DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
Claims 1 and 7 are amended, claims 2, 6, 10-12, and 14-20 are cancelled, and claims 21-25 are new. Claims 1, 3-5, 7-9, 13, and 21-25 are pending.
Response to Arguments
Applicant's arguments filed 9/22/2025 have been fully considered.
Regarding the rejections of claim 1 under 101, Applicant contends on pages 6-7 of the response that the claim does not merely recite abstract activities and instead integrates ideas into a specific and practical application. In support, Applicant notes that the claim requires the hypothesis be based on simulations produced by a virtual research model and further that the claim requires use of cluster analysis to identify smart farms for data collection and real-time data collection from sensors located around the livestock and that the data is then analyzed using statistical models and a customized report is generated, transmitted, and either accepted or denied by a smart farm operator.
The Examiner submits that the foregoing functions includes some functions that have been found to fall within the mental processes judicial exception because they may be performed, individually or in combination, via mental processes and Applicant’s arguments do not appear to directly address which elements and/or combinations of elements that have been found to fall within the judicial exception may have been incorrectly categorized in this way. The Examiner further notes that while some of the foregoing functions cited by Applicant (e.g., generating and transmitting a report) fall outside the judicial exception and therefore qualify as “additional elements,” Applicant’s arguments do not specifically address the manner in which such additional elements, either individually or in combination, may integrate the functions falling within the judicial exception into a practical application such as via improvements to the functioning of a computer, or to any other technology or technical field; applying the judicial exception with, or by use of, a particular machine; and/or effecting a transformation or reduction of a particular article to a different state or thing.
On page 7 of the response, Applicant contends that features such as implementation of recommendations followed by transmission of implementation data to a cloud configuration, wherein the transmitted data becomes historical data for future virtual modeling forms a feedback loop between the digital and virtual research systems, enabling continuous improvement and machine learning. Applicant further asserts that the feedback mechanism, coupled with real-time sensor integration, distinguishes the claim from abstract data analysis and reporting and demonstrate that claim 1 is not directed to a judicial exception but rather to a practical application of digital technology for a specific and useful purpose.
The Examiner respectfully disagrees that the various components and functions, individually and in combination, that fall outside the judicial exception act/function to integrated the steps falling within the judicial exception into a practical application. Applicant generally alludes to an overall practical application, but has not described a particular manner in which the additional elements in claim 1 are not merely extra solution activity but rather functions/structures that play a significant functional role in the manner in which the steps falling within the judicial exception are performed (in contrast to being routine, convention data processing activities/components that serve the substantially a similar role with respect to any type of underlying computer processing).
In further regard to claim 1, Applicant contends on page 7 of the response that even if claim 1 is found to recite an abstract idea, claim 1 recites additional features that amount to significantly more than the judicial exception. In support Applicant cites the use of real-time livestock sensors, wireless data transmission, integration with a cloud-based virtual research model, and application of artificial intelligence models to real-world data as providing an inventive concept that is not routine, conventional, or generic.
The Examiner submits that the recited additional elements, either individually or in combination do not result in the claim as a whole amounting to more than the judicial exception because the additional elements constitute extra-solution activity (i.e., activity/components having no particularized functional relation to the elements falling within the judicial exception). Furthermore, as explained in the grounds for rejecting amended claim 1 under 103, the combination of elements falling within the judicial exception and elements falling outside the judicial exception are taught or otherwise rendered obvious by the set of prior arts utilized in the rejection such that no inventive concept is presented by claim 1.
Regarding the rejection of independent claim 7 under 101, Applicant notes that claim 7 recites that the data analysis is performed by applying artificial intelligence to livestock data using multiple data processing modules within a cloud computing environment and that these modules are executed in conjunction with a processor and are configured to continuously evaluate data from a plurality of producers to estimate/predict health improvements by modifying variables. Applicant contends on page 8 of the response that this configuration is not a generic computing environment, but rather a specific infrastructure for managing and analyzing complex real-world livestock data requiring implementation in a distributed, cloud-based system.
The Examiner acknowledges that the combination of all of the additional elements recited in claim 7 may fall outside what may be considered a “generic” computing environment. However, as explained in the grounds for rejecting claim 7 many of these elements and in particular the processor/processing, sensor-based collection, cloud processing/networking environment appear to be substantially generic in the similar applications disclosed in the prior arts including Kuper (US 2018/0350010 A1), Stroman (US 2008/0059264 A1), Singh (US 2019/0053470 A1), Richt (US 2018/0075546 A1), and Kopic (US 2014/0116341 A1). Furthermore, the additional elements in claim 7, individually and in combination, appear to constitute extra solution activity that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception.
In further regard to claim 7, Applicant contends on page 8 that claim 7 recites a non-abstract improvement to an existing technological process. In support, Applicant notes that the claimed method overcomes efficiency problems encountered by traditional method by enabling scalable, predictive simulations and feedback-based learning that reduce the need for repetitive trials and enhance the efficiency and accuracy of research outcomes. Applicant further asserts that the claim recites additional features that amount to significantly more than an abstract idea, including integration of cloud-based computing, real-time data input, artificial intelligence model training and retraining, and application-specific improvements for agricultural systems.
The Examiner submits that these arguments are somewhat generalized and do not specifically describe the manner in which one or more specific additional elements, individual or in combination, act to integrate the elements falling within the judicial exception into a practical application or result in the claim as a whole amounting to significantly more than the judicial exception. The Examiner submits that as set forth in the current grounds for rejecting claim 7, cloud-based computing represents routine, convention data processing infrastructures for storing and distributing data, real-time data input represents high-level data collection, and artificial intelligence modeling and training represents routine, convention program instruction implementation of underlying functions, such that these elements, individually or in combination, represent extra solution activity that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception.
Regarding the rejection of claim 7 under 103, Applicant contends on page 9 of the response that the combination of references fails to teach or suggest the structured and iterative framework set forth in claim 7. Specifically, Applicant asserts that Kuper ‘010 does not describe a formalized virtual research process that includes simulation, hypothesis generation, report generation, user feedback, field trail validation, and model training.
The Examiner submits that Applicant’s argument regarding what the “combination” of references teaches does not set forth the specific reasons as to why/in what manner the previous combination of prior art applied to reject claim 7 fails to teach the combination of recited elements and is therefore essentially conclusory. For example, Applicant contends, on page 9, that the step of implementing a recommendation in a controlled field trial following acceptance of a report is not disclosed by any of the references, without supporting this contention by showing how the particular references fails to actually disclose particular features and/or why the references are not properly combinable. Therefore, the Examiner can only point to the grounds for rejecting claim 7 to indicate why the combination of Kuper (US 2018/0350010 A1), “Kuper ‘010,” in view of Stroman (US 2008/0059264 A1), and in further view of Singh (US 2019/0053470 A1), Kopic (US 2014/0116341 A1), and Park (WO 2019216647A1) teaches or otherwise renders obvious the combination of elements in claim 7.
Regarding the amendments to claim 7, Applicant contends on page 9 that the previously cited references do not teach the application of artificial intelligence in a distributed, modular, cloud-based system designed for continuous real-time evaluation and adaptation. Applicant further contends that the cited references to not disclose a model retraining step in connection with specific field trial outcomes resulting from a recommendation generated by earlier virtual research. These arguments are again essentially conclusory and generalized in terms of the cited prior art references failing to teach combinations of elements (in which there appears to be some inference of features that are not actually recited in the claim), and the Examiner notes that as set forth in the current grounds for rejecting claim 7, the combination of Kuper ‘010, Stroman, Singh, Kopic, and Park teach or otherwise render obvious these combinations of elements. The Examiner further notes that, as explained in the grounds for rejecting claim 7 under 112(a), Applicant’s original disclosure does not appear to support, with reasonable clarity, “an artificial intelligence supervised model trained on livestock-specific real-time data obtained from the field trial.”
Regarding the grounds for rejecting claim 1 under 103, Applicant sets forth arguments on pages 10-11 including that the Kuper references “do not describe a simulation-based research framework that generates a hypothesis later tested through real-time sensor data collection at selected farms,” that the combination of references do not disclose “a cluster analysis of historical data be performed to determine which smart farms are appropriate for real-time data collection,” that the cited references do not teach or suggest “the integration of real-time livestock data analysis with a feedback loo that includes wireless transmission of a customized research report, confirmation or denial of acceptance by the receiving smart farm, and the implementation of recommendations derived from the initial simulation-based hypothesis,” and that the cited references fail to teach or suggest “the final limitations of claim 1.”
The Examiner submits that these arguments are generalized and incorporate inferences that are not recited in the claim language. Furthermore, and similar to the arguments relating to claim 7, these arguments do not address with any specificity how/why the cited prior art references fail to teach particular elements and/or how/why the references are not properly combinable as set forth in the rejections and are therefore essentially conclusory. Therefore, the Examiner cites the current grounds for rejecting claim 1 under 103 as indicating why the subject matter of claim 1 is unpatentable over Kuper ‘010, Stroman, Singh, Kopic, Park, and Perry (US 2019/0050948 A1).
Note: The Examiner recognizes that the application of multiple references for the various aspects of the claims results in numerous complexities and has undertaken to make the rejections as clear as possible. Given the relative complexity of the rejections, the Examiner earnestly welcomes the Applicant’s representative to contact the Examiner for the purpose of scheduling an Examiner’s Interview to discuss any questions or concerns that may arise in review of the rejections such that prosecution of this application may be advanced as efficiently as possible.
Claim Objections
Claim 24 is objected to because of the following informalities:
In claim 24 line 1, “the step of storing” should read “a step of storing.”
In claim 24 line 3, “the updated model version” does not have sufficient antecedent basis and should be amended to recite “the retrained model.”
Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
Claims 1, 3-5, 7-9, 13, and 21-25 are rejected under 35 U.S.C. 112(a) as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor had possession of the claimed invention.
Claim 7 lines 9-12 recites “analyzing the livestock data … with one or more analysis models, including a trained artificial intelligence model configured to generate predictive outcomes by applying statistical correlations, normalization, and aggregation techniques to the livestock data,” which does not appear to be adequately supported by Applicant’s original disclosure. The support for this element appears in Applicant’s specification on page 6 lines 27-29, which discloses that analyzing the livestock data includes “correlation, normalization, aggregation, and modeling of the data” (i.e., modeling in addition to correlation, normalization, and aggregation is used for analyzing the data). While this disclosure generally associates modeling and the functions of statistical correlation, normalization, and data aggregation, Applicant’s original disclosure does not appear to disclose with reasonable clarity that the model itself applies statistical correlations, normalization, and aggregation techniques to the livestock data.
Claims 1, 3-5, 8-9, 13, and 21-25 depend from claim 7 and are likewise rejected for the same reasons.
Claim 7 lines 19-20 recites “an artificial intelligence supervised model trained on livestock-specific real-time data obtained from the field trial,” which does not appear to be adequately supported by Applicant’s original disclosure. The support for this element appears in Applicant’s specification on page 2 lines 30-32 and in original claim 12, which discloses analyzing field trial data and retraining the one or more analysis models. While this disclosure appears to associates field trail data and model training in terms of an overall process, and may suggest to one of ordinary skill in the art that the field trial data may be used for training the model, this disclosure does not expressly disclose that the field trial data is used for training a model. Moreover, Applicant’s original disclosure does not appear to disclose using real-time data obtained from a field trial for training a model. Therefore, Applicant’s original disclosure does not appear to disclose with reasonable clarity an artificial intelligence supervised model trained on livestock-specific real-time data obtained from the field trial.”
Claims 1, 3-5, 8-9, 13, and 21-25 depend from claim 7 and are likewise rejected for the same reasons.
Claim 24 recites “storing the retrained model in a version-controlled model registry,” which does not appear to be adequately supported by Applicant’s original disclosure. The support for this element appears to be the “database” function/component referenced throughout Applicant’s specification such as on page 8 lines 13-15, page 8 line 15 through page 9 line 12, and page 9 lines 29-33, which disclose use a database for storing information relating to modeling (e.g., inputs/outputs) and other storage functions. However, Applicant’s original disclosure does not appear to disclose, with reasonable clarity, storing any model (trained and/or retrained) “in a version-controlled model registry.”
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.
Claims 1, 3-5, 7-9, 13, and 21-25 are rejected under 35 U.S.C. 112(b) as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor regards as the invention.
In claim 7 lines 9-12, “analyzing the livestock data … with one or more analysis models, including a trained artificial intelligence model configured to generate predictive outcomes by applying statistical correlations, normalization, and aggregation techniques to the livestock data” renders claim 7 indefinite because this language potentially suggests that the “model” itself applies statistical correlations, normalization, and aggregation techniques to the livestock data. This potential interpretation does not appear entirely consistent with the specification description on page 6 lines 27-29, which discloses that analyzing the livestock data includes “correlation, normalization, aggregation, and modeling of the data” (i.e., modeling in addition to correlation, normalization, and aggregation is used for analyzing the data). As noted in the grounds for rejecting claim 7 under 103, correlation and aggregation may be considered inherently incident to artificial intelligence modeling but normalization implemented by the model itself is not. It appears that the intent of the language in lines 9-12 of claim 7 is to characterize the “analyzing” as entailing using “one or more analysis models, including a trained artificial intelligence model configured to generate predictive outcomes in which the overall modeling process, including data processes not strictly implemented by the model itself, applies statistical correlations, normalization, and aggregation techniques, which is how claim 7 is interpreted for the purposes of prior art examination.
Claims 1, 3-5, 8-9, 13, and 21-25 depend from claim 7 and are likewise rejected for the same reasons.
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, 3-5, 7-9, 13, and 21-25 are rejected under 35 U.S.C. 101 because the claimed invention in each of these claims is directed to the abstract idea judicial exception without significantly more.
Claim 7 recites:
“[a] method for conducting virtual research to predict the outcome of a selected variable of interest and improve livestock health outcomes, the method comprising the steps of:
(a) automatically obtaining livestock data from a plurality of livestock producers, the livestock data related to one or more variables pertaining to livestock nutrition, livestock environment, livestock health or livestock performance;
(b) transmitting the livestock data to a cloud computing environment comprising at least one server, at least one processor, and a memory component;
(c) analyzing the livestock data in the cloud computing environment with one or more analysis models, including a trained artificial intelligence model configured to generate predictive outcomes by applying statistical correlations, normalization, and aggregation techniques to the livestock data;
(d) generating a customized report that includes the results of the analysis, the results including an estimate or prediction of an expected improvement in livestock health outcome by changing one or more variables;
(e) transmitting the report to an end user, wherein the end user may accept or deny the report;
(f) implementing the change to one or more variables in a field trial;
(g) building an artificial intelligence supervised learning model trained on livestock-specific real-time data obtained from the field trial;
(h) analyzing the data produced from the field trial using at least one trained machine learning model selected from the group consisting of ensemble learning, deep learning, time series modeling, and Bayesian statistics; and
(i) retraining the one or more of the analysis models,
wherein the step of analyzing the livestock data includes applying artificial intelligence to the livestock data in a plurality of data processing modules within a cloud computing environment in which the plurality of data processing modules are executed in conjunction with at least one processor, the data processing modules and artificial intelligence configured to continuously evaluate data from a plurality of livestock producers to estimate or predict an expected improvement in livestock health outcome by changing one or more variables,
wherein the cloud computing environment executes a plurality of modularized data processing components, each configured to perform a specific transformation on the data, and wherein the data analysis is performed in conjunction with at least one processor to generate a livestock-specific output that enables automated feedback and iterative improvement in livestock health outcomes.”
Claim 1 recites:
“[a] method of conducting digital research to improve livestock health outcomes, the method comprising
forming, using a processor, a hypothesis regarding what changes may be made to one or more variables related to livestock nutrition, environment, health status, or animal performance to improve livestock health outcomes for a particular livestock producer, the hypothesis based on simulations produced according to the virtual research method of claim 7;
performing, using a statistical or machine learning algorithm, a cluster analysis of historical data related to the one or more variables to determine one or more smart farms for real-time livestock data collection;
obtaining real-time livestock data from one or more sensors or cameras located on or around a plurality of livestock animals at the determined one or more smart farms;
analyzing the real-time livestock data using a trained artificial intelligence model to test the hypothesis;
generating, via a cloud-based processing module, a customized research report that includes results of the analysis of the real- time livestock data including one or more recommendations to improve livestock health outcomes,
transmitting the research report wirelessly to the one or more smart farms;
confirming or denying acceptance of the research report by the one or more smart farms;
implementing the one or more recommendations by the smart farm; and
transmitting data related to an outcome of implementation of the one or more recommendations to a cloud configuration or central database, wherein the transmitted data is made available for use as historical data in a virtual research model,
wherein upon implementation of the one or more recommendations by the smart farm, livestock health outcomes are improved.”
The claim limitations considered to fall within in the abstract idea are highlighted in bold font above and the remaining features are “additional elements.”
Step 1 of the subject matter eligibility analysis entails determining whether the claimed subject matter falls within one of the four statutory categories of patentable subject matter identified by 35 U.S.C. 101: process, machine, manufacture, or composition of matter. Claim 7 recites a method and therefore falls within a statutory category.
Step 2A, Prong One of the analysis entails determining whether the claim recites a judicial exception such as an abstract idea. Under a broadest reasonable interpretation, the highlighted portions of claims 7 and 1 fall within the abstract idea judicial exception. Specifically, under the 2019 Revised Patent Subject matter Eligibility Guidance, the highlighted subject matter falls within the mental processes category (including an observation, evaluation, judgment, opinion) and the mathematical concepts category (mathematical relationships, mathematical formulas or equations, mathematical calculations). MPEP § 2106.04(a)(2).
Regarding claim 7, the recited functions:
“conducting virtual research to predict the outcome of a selected variable of interest and improve livestock health outcomes”
“automatically obtaining livestock data from a plurality of livestock producers, the livestock data related to one or more variables pertaining to livestock nutrition, livestock environment, livestock health or livestock performance”
“analyzing the livestock data in the cloud computing environment”
“generate predictive outcomes by applying statistical correlations, normalization, and aggregation techniques to the livestock data”
“estimate or prediction of an expected improvement in livestock health outcome by changing one or more variables,”
“implementing the change to one or more variables in a field trial”
“analyzing the data produced from the field trial,” and
“continuously evaluate data from a plurality of livestock producers to estimate or predict an expected improvement in livestock health outcome by changing one or more variables” may be performed, individually and in combination, as mental processes.
Conducting virtual research to predict an outcome of a selected variable of interest and improve livestock health outcomes may be performed via mental processes (e.g., evaluation of data, including data from online sources, relating to factors affecting livestock and applying judgment to such evaluations to formulate outcome predictions). Automatically obtaining livestock data from a plurality of livestock producers, the livestock data related to one or more variables pertaining to livestock nutrition, livestock environment, livestock health or livestock performance may also be performed via mental processes (e.g., observation, automatically in terms of some prompting, of data generated in some manner from livestock producers). Analyzing the livestock data in the cloud computing environment may also be performed via mental processes (e.g., evaluation of livestock data in a manner entailing mental comparison of the data with example livestock data (analysis models) or, mental evaluation of livestock data that is generated using a computer-implemented analysis model). Generating predictive outcomes by applying statistical correlations, normalization, and aggregation techniques to the livestock data may be performed via mental processes (e.g., applying analytic techniques, possibly aided by pen-and-paper, including statistical correlation, normalization, and aggregation and evaluation of results to determine (via judgement) predictions). Estimating or predicting an expected improvement in livestock health outcome by changing one or more variables may be performed via mental processes (e.g., evaluation and judgement). Implementing the change to one or more variables in a field trial (e.g., implementation beginning with ascertaining course of action based on a changed variable involved in field trial implementation), and analyzing the data produced from the field trial (e.g., evaluation) may also be performed via mental processes. Continuously evaluating data from a plurality of livestock producers to estimate or predict an expected improvement in livestock health outcome by changing one or more variables may be performed via mental processes (e.g., evaluation and judgment).
Regarding claim 7, the recited functions “applying statistical correlations, normalization” are further found by the Examiner to fall within the mathematical relations subcategory of the mathematical concepts judicial exception because statistical correlations and normalization are each fundamentally characterized by mathematical relations/calculations.
Regarding claim 1, the recited functions:
“conducting digital research to improve livestock health outcomes”
“forming” “a hypothesis regarding what changes may be made to one or more variables related to livestock nutrition, environment, health status, or animal performance to improve livestock health outcomes for a particular livestock producer, the hypothesis based on simulations produced according to the virtual research method of claim 7;
performing a cluster analysis of historical data related to the one or more variables to determine one or more smart farms for real-time livestock data collection;
obtaining real-time livestock data from one or more sensors or cameras located on or around a plurality of livestock animals at the determined one or more smart farms;
analyzing the real-time livestock data to test the hypothesis,”
“confirming or denying acceptance of the research report by the one or more smart farms,” and
“implementing the one or more recommendations by the smart farm,” may be performed as mental processes.
Conducting digital research to improve livestock health outcomes and forming hypotheses regarding changes to variables related to livestock nutrition, environment, health status, or animal performance to improve livestock health outcomes for a particular livestock producer in which the hypothesis based on simulations produced according to the virtual research method of claim 7 may be performed via mental processes (e.g., evaluation of data, including data from digital sources, relating to factors affecting livestock and applying judgment to such evaluations to formulate hypotheses). Performing a cluster analysis of historical data related to the one or more variables to determine one or more smart farms for real-time livestock data collection may also be performed via mental processes (e.g., evaluation/judgment of historical variable data to mentally formulate patterns such as commonalities defining clusters of such information relating to smart farms). Obtaining real-time livestock data from one or more sensors or cameras located on or around a plurality of livestock animals at one or more smart farms may also be performed via mental processes (e.g., observation either of data in a real-time stream from sensors or observation of data that may be stored and was collected by sensors in real-time). Analyzing real-time livestock data to test the hypothesis may also be performed via mental processes (e.g., evaluation and judgment applied to livestock data conducted for the purpose of testing a related hypothesis). Confirming or denying acceptance of the research report by the one or more smart farms may also be performed via mental processes (e.g., evaluation/review of research report and judgement regarding the same). Implementing the one or more recommendations by the smart farm may also be performed via mental processes (e.g., implementation beginning with ascertaining course of action based on a recommendation and performed by a smart farm personnel).
The type of high-level information analysis and deduction recited in the foregoing elements of claims 7 and 1 has been found by the Federal Circuit to constitute patent ineligible matter (see Electric Power Group v. Alstom, S.A., 830 F.3d 1350, 1353-54, 119 USPQ2d 1739, 1741-42 (Fed. Cir. 2016), a claim to "collecting information, analyzing it, and displaying certain results of the collection and analysis," where the data analysis steps are recited at a high level of generality such that they could practically be performed in the human mind).
Step 2A, Prong Two of the analysis entails determining whether the claim includes additional elements that integrate the recited judicial exception into a practical application. “A claim that integrates a judicial exception into a practical application will apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception, such that the claim is more than a drafting effort designed to monopolize the judicial exception” (MPEP § 2106.04(d)).
MPEP § 2106.04(d) sets forth considerations to be applied in Step 2A, Prong Two for determining whether or not a claim integrates a judicial exception into a practical application. Based on the individual and collective limitations of claims 7 and 1 and applying a broadest reasonable interpretation, the most applicable of such considerations appear to include: improvements to the functioning of a computer, or to any other technology or technical field (MPEP 2106.05(a)); applying the judicial exception with, or by use of, a particular machine (MPEP 2106.05(b)); and effecting a transformation or reduction of a particular article to a different state or thing (MPEP 2106.05(c)).
Regarding improvements to the functioning of a computer or other technology, none of the “additional elements” of claims 7 and 1 in any combination appear to integrate the abstract idea in a manner that technologically improves any aspect of a device or system that may be used to implement the highlighted step or a device for implementing the highlighted step such as a signal processing device or a generic computer. For example, “transmitting the livestock data to a cloud computing environment comprising at least one server, at least one processor, and a memory component,” “one or more analysis models, including a trained artificial intelligence model configured to generate predictive outcomes” for performing the livestock data analysis, “generating a customized report that includes the results of the analysis,” and “transmitting the report to an end user, wherein the end user may accept or deny the report” in claim 7, and “using a processor” to form a hypothesis, “using a statistical or machine learning algorithm” to perform a cluster analysis, “generating, via a cloud-based processing module, a customized research report that includes results of the analysis of the real-time livestock data including one or more recommendations to improve livestock health outcomes,” “transmitting the research report wirelessly to the one or more smart farms,” and “transmitting data related to an outcome of implementation of the one or more recommendations to a cloud configuration or central database, wherein the transmitted data is made available for use as historical data in a virtual research model,” in claim 1, represent well-known data processing functions that do not influence or rely on the functions entailing the abstract idea and do not otherwise technologically improve any aspect of a device or system. Therefore, these elements constitute extra-solution activity (post-solution activity for “generating a customized research report).
Similarly, “building an artificial intelligence supervised learning model trained on livestock-specific real-time data obtained from the field trial,” “using at least one trained machine learning model selected from the group consisting of ensemble learning, deep learning, time series modeling, and Bayesian statistics” to analyze the data produced from the field trial, and “retraining the one or more of the analysis models, wherein the step of analyzing the livestock data includes applying artificial intelligence to the livestock data in a plurality of data processing modules within a cloud computing environment in which the plurality of data processing modules are executed in conjunction with at least one processor, the data processing modules and artificial intelligence configured to continuously evaluate data from a plurality of livestock producers” in claim 7, individually and collectively with the other additional elements represent standard data collection, transmission, processing functions (high-level model generation and training) that are well-known and do not directly influence the functions entailing the abstract idea and therefore constitute extra solution activity that does not otherwise technologically improve any aspect of a device or system.
Regarding claim 1, the Examiner notes that “wherein upon implementation of the one or more recommendations by the smart farm, livestock health outcomes are improved” recites an intended result/purpose that does not further characterize the function performed by the method and is therefore does not constitute an additional element that integrates the abstract idea into a practical application.
Regarding application of the judicial exception with, or by use of, a particular machine, the foregoing additional elements of claims 7 and 1 are configured and implemented in a conventional rather than a particularized manner of implementing livestock monitoring.
Regarding a transformation or reduction of a particular article to a different state or thing, claims 7 and 1 do not include any such transformation or reduction. Instead, each of these claims as a whole entails steps and components for receiving/collecting livestock input information, applying standard processing techniques (i.e., computer-implemented execution of instruction) to the information to obtain and deduce livestock health-related information with the additional elements failing to provide a meaningful integration of the abstract idea in an application that transforms an article to a different state. Instead, the additional elements represent extra-solution activity that does not integrate the judicial exception into a practical application. In view of the various considerations encompassed by the Step 2A, Prong Two analysis, claims 7 and 1 do not include additional elements that integrate the recited abstract idea into a practical application.
Therefore, each of claims 7 and 1 is directed to a judicial exception and requires further analysis under Step 2B.
Regarding Step 2B, the additional elements in claims 7 and 1 represent extra-solution activity as set forth in the Step 2A Prong Two analysis and therefore do not result in the claim as a whole amounting to significantly more than the judicial exception. Furthermore, the additional elements appear to be generic and well understood as evidenced by the disclosures of Kuper (US 2018/0350010 A1), “Kuper ‘010”, Stroman (US 2008/0059264 A1), Singh (US 2019/0053470 A1), Richt (US 2018/0075546 A1), and Kopic (US 2014/0116341 A1) each of which disclose livestock/farming monitoring methods/systems utilizing substantially the same data collection, transmission, and processing operations.
Regarding claims 7 and 1, Kuper ‘010 teaches generating an output report that includes analysis results (FIG. 1 output data 170 that includes average daily gain 171, growth rate predictions 172, and livestock recommendations 173) as does Stroman (FIG. 9 block 350).
Regarding claims 7 and 1, Singh teaches using a cloud computing environment for transmitting and processing livestock data ([0035]) as does Richt ([0217] analogously applied for plant farming rather than livestock farming).
Regarding claim 7, “transmitting the report to an end user” is well-understood and conventional in the art as taught by Stroman (FIG. 1 depicting networked systems including client computers 110 and server 107 providing access to data within central database 109 via links 115, [0070]; [0129] the adaptive reasoning system 140 (that is part of overall platform depicted in FIG. 1) electronically available at ranches; [0130] output from adaptive reasoning system 140 in the form of a notification; [0133]-[0134] user may retrieve analyzed information from the database via the Internet; [0104] “users” may include ranchers) and Singh (FIG. 1B depicting networked connection between remote server 108 and users 110a-110b; [0035], FIG. 3A block 314).
In claim 7, the functions “building an artificial intelligence supervised learning model,” “retraining the one or more of the analysis models,” and “applying artificial intelligence to the livestock data in a plurality of data processing modules within a cloud computing environment in which the plurality of data processing modules are executed in conjunction with at least one processor” (effectively conveying application of artificial intelligence to the livestock data in a plurality of data processing modules in a cloud computing environment in conjunction with a processor) are well-understood and conventional in the art as taught by Kuper ‘010 (FIG. 1 AI layer 146 and adaptive machine learning methods 147 configured to processing variable input data), Stroman (FIG. 1 adaptive reasoning system 140 and other networked components, [0132]), Singh (FIG. 3A block 312; [0035] cloud computing environment), and Richt (FIG. 1A, [0125] machine learning applied, [0217] cloud computing environment). Building/training models including AI models is further well-known and conventional in the art as disclosed by Stroman ([0129] analytic and empirical models may be refined, improved, recalibrated. Examiner notes that improving and recalibrating models of the system that per [0128] is “adaptive” and has “learning capabilities” entails retraining the model), Kuper ‘010 (FIG. 1 adaptive machine learning methods 147 (Examiner notes that a machine learning method is inherently implemented by some form of data structure (model) that has been built), [0033]), and Kopic ([0076] discloses a method for analytic management of dairy production that includes generating a production model (Abstract) that may be a supervised machine learning model).
Regarding claim 1, transmitting the research report wirelessly to the one or more smart farms and transmitting data related to an outcome of implementation of the one or more recommendations to a cloud configuration or central database, are well-understood and conventional in the art as taught by Stroman (FIG. 1 depicting networked systems including client computers 110 and server 107 providing access to data within central database 109 via links 115 that per [0073] may be wireless, [0070]; [0129] the adaptive reasoning system 140 (that is part of overall platform depicted in FIG. 1) electronically available at ranches; [0130] output from adaptive reasoning system 140 in the form of a notification; [0133]-[0134] user may retrieve analyzed information from the database via the Internet; [0104] “users” may include ranchers) and Singh (FIG. 1B depicting networked connection between remote server 108 and users 110a-110b; [0035], FIG. 3A block 314).
Therefore, the additional elements in claims 7 and 1 are insufficient to amount to significantly more than the judicial exception and consequently claims 7 and 1 are not patent eligible.
Regarding claim 7, The Examiner notes that even if “automatically obtaining livestock data from a plurality of livestock producers, the livestock data related to one or more variables pertaining to livestock nutrition, livestock environment, livestock health or livestock performance” is interpreted more narrowly to fall outside the mental processes exception, this element represents high-level data collection and therefore constitutes extra-solution activity that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception. Similarly, even if “implementing the change to one or more variables in a field trial” is interpreted more narrowly to fall outside the mental processes exception, this element represents convention, routine data processing activity having no particularized functional relation to the steps falling within the judicial exception and therefore constitutes extra-solution activity that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception.
Regarding claim 3, the characterization of the sensors as transmitting a particular code associated with an animal neither integrates the abstract idea into a practical application nor results in claim 3 as a whole amounting to significantly more than the judicial exception because it constitutes extra solution activity (high-level data collection) and is furthermore well-understood and conventional in the art as taught by Stroman (FIG. 1 tracking system 145 that per [0097]-[0099] includes livestock sensors; FIG. 2 block 215 depicting electronic identification of animals based on RFID detection at block 210, [0028] RFID detection/identification/data collection; [0097]-[0099] RFID tagging using RFID tags on animals. Examiner notes that RFID tag sensing inherently entails sending an RF identification code to a reader), and Singh (FIG. 1B tag assemblies 102 including sensor 114, [0047]).
Claim 3 therefore also constitutes ineligible subject matter.
Regarding claim 4, using one or more statistical models to implement the analysis of real-time livestock data constitutes extra solution activity (well-known, conventional data processing for implementing the judicial exception) that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception.
Claim 5 does not recite any further additional elements because it only characterizes the data that is processed (cluster analysis).
Claim 5 therefore also constitutes ineligible subject matter for the same reasons as set forth for claim 1.
Claims 8-9 do not recite any further additional elements because each only characterizes the data that is processed and therefore also constitute ineligible subject matter for the same reasons as set forth for claim 7.
Claim 13 further recites that steps (a)-(d) are repeated upon denial of the report. As set forth in the grounds for rejecting claim 7, steps (a) and (c) fall within the judicial exception and the additional elements (b) and (d) do not integrate the judicial exception into a practical application or result in the claim overall as amounting to significantly more than the judicial exception. Therefore, the additional element newly introduced by claim 13 is essentially repeating processing cycles in response to a state or condition, which constitutes well-understood computer data processing that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception.
Claim 21 further recites “the step of obtaining livestock data comprises receiving real-time sensor data from a plurality of smart farms, the sensor data including at least one of environmental temperature, humidity, water pH, water temperature, ammonia levels, or animal body weight,” which represents high-level data collection having no particularized relation to the steps falling within the judicial exception and therefore constitutes extra solution activity that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception.
Claim 22 further recites “the artificial intelligence supervised learning model comprises one or more machine learning models selected from the group consisting of decision trees, regression trees, random forests, gradient boosting machines, support vector machines, neural networks, and Bayesian networks,” which represents application of known artificial intelligence modeling program configurations having no particularized relation to the steps falling within the judicial exception and therefore constitutes extra solution activity that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception.
Claim 23 further recites “the step of retraining one or more of the analysis models comprises modifying at least one model parameter based on performance data from the field trial using a semi-supervised or transfer learning approach,” which represents program instruction preparation using known machine learning training techniques having no particularized relation to the steps falling within the judicial exception and therefore constitutes extra solution activity that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception.
Claim 24 further recites “storing the retrained model in a version-controlled model registry accessible within the cloud computing environment, wherein subsequent analyses utilize the updated model version,” which represents conventional, routine data processing activity in terms of storing model data in an accessible registry (e.g., searchable database) having no particularized relation to the steps falling within the judicial exception and therefore constitutes extra solution activity that neither integrates the judicial exception into a practical application nor results in the claim as a whole amounting to significantly more than the judicial exception.
Claim 25 further recites “the customized report includes a prediction of improvement in at least one performance metric selected from feed conversion ratio, mortality rate, average daily gain, or carcass yield, based on changing one or more variables,” which only describes the nature of the data processed by/resulting from the steps falling within the judicial exception and therefore itself falls within the judicial exception.
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 7-9, 13, 22-23, and 25 are rejected under 35 U.S.C. 103 as being unpatentable over Kuper (US 2018/0350010 A1), “Kuper ‘010,” in view of Stroman (US 2008/0059264 A1), and in further view of Singh (US 2019/0053470 A1), Kopic (US 2014/0116341 A1), and Park (WO 2019216647A1).
As to claim 7, as best understood in view of the grounds for rejecting claim 7 under 112(b), Kuper ‘010 teaches “[a] method for conducting virtual research to predict the outcome of a selected variable of interest and improve livestock health outcomes (Abstract disclosing adaptive framework method for modeling livestock growth including optimizing and recommending feed operations (feed operation being the variable of interest and outcome being growth); FIG. 2 including block 290, [0008] nutrition recommendations are predictive), the method comprising the steps of:
(a) automatically obtaining livestock data from” “livestock producers (FIG. 1 input data 110 relating to livestock includes nutrition, weather, animal-specific data, location, etc.; [0037] and [0039]-[0040] data inputs relating to livestock conditions are provided automatically), the livestock data related to one or more variables pertaining to livestock nutrition, livestock environment, livestock health or livestock performance (FIG. 1 input data 110 relating to livestock includes nutrition, weather, animal-specific data, location, etc.);”
“(c) analyzing the livestock data” “with one or more analysis models, including a trained artificial intelligence model configured to generate predictive outcomes (FIG. 1 depicting data processing components 140 configured to analyze via modeling input data 110 including AI layer 146, adaptive machine learning methods 147 (application of machine learning inherently entails the model having been trained), and selected growth model 160; [0015] AI layer 146 and selected model 160 generate output data 170 such as predicted growth rate 172, average gain 171 and recommendations 173 based on processing input data described in [0019]) by applying statistical correlations (statistical correlation inherent to machine learning modeling),” “, and aggregation techniques to the livestock data (aggregation inherent to machine learning modeling such as is entailed in generating (training) a model and also in AI layer 146, adaptive machine learning methods, and selected model 160 processing the multi-variate/aggregate input data 110);
(d) generating a customized report that includes the results of the analysis (FIG. 1 output data 170 that includes average daily gain 171, growth rate predictions 172, and livestock recommendations 173 (Examiner notes the output data constitutes a report that is customized in terms of being dynamically adjusted in accordance with changing input data 110 and per [0026] in accordance with model weight adjustments)), the results including an estimate or prediction of an expected improvement in livestock health outcome by changing one or more variables (FIG. 1 output data 170 that includes growth rate predictions 172 and corresponding livestock recommendations 173; [0033] modeling including machine learning used to predict and recommend livestock feeding and management (Examiner notes the predictions are determined based on (dependent on) modeling/processing of variable input data such that the predictions themselves constitute an estimate/prediction of expected improvement in which the “changing one or more variables” aspect of the predictions are manifested in corresponding recommendations).”
Kuper ‘010 further suggests automatically obtaining livestock data from “a plurality of” livestock producers ([0005], [0023], and [0035] disclosing that producer-specific information utilized to optimize workflow, which suggests discerning among multiple producers (farms)).”
Stroman teaches automatically obtaining livestock data from multiple livestock producers ([0002] Internet-based platform shared by members of livestock production chain including producers; FIGS. 1 and 8 tracking system 145; [0016] tracking system used to document information from different ranches).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have combined Stroman’s teaching of collecting livestock data from multiple farms/livestock producers with the method taught by Kuper ‘010.
The motivation would have been to expand the comprehensive coverage of livestock data available for processing such as to more accurately determine monitoring processing results from such varied information as disclosed by Stroman.
Neither Kuper ‘010 nor Stroman expressly teaches “transmitting the livestock data to a cloud computing environment comprising at least one server, at least one processor, and a memory component,” analyzing the livestock data “in the cloud computing environment,” generating predictive outcomes by applying “normalization,” and “wherein the cloud computing environment executes a plurality of modularized data processing components, each configured to perform a specific transformation on the data, and wherein the data analysis is performed in conjunction with at least one processor to generate a livestock-specific output.”
Singh discloses a method for tracking health in animal populations (Abstract) in which livestock data ([0036]-[0038] tag assemblies 102 may include sensors 114 that collect livestock information) is transmitted to a cloud computing environment comprising at least one server, at least one processor, and a memory component (FIGS. 1A-1B depicting network-based computing environment comprising network 106 as information portal between user devices 110, tag assemblies 102 via concentrator 104, and remote server 108; [0043] tag assembly information provided to concentrator 104; [0050] concentrator 104 communicatively coupled with remote server 108 via network 106; [0052] network interface of concentrator transmits livestock data to remote server; [0035] network 106 may comprise a cloud-based architecture. Examiner notes that the computer-based cloud computing environment depicted in FIGS. 1A and 1B that implements data processing function such as performed by remote server 108 inherent includes a processor and memory for execution of functions) and analyzing the livestock data in the cloud computing environment (FIG. 1B remoter server 108 forming part of cloud computing environment in terms of being provisioned data via network 106 and including processor 130 configured to process data received in the cloud computing environment; [0052]), and wherein the cloud computing environment executes a plurality of modularized data processing components (FIG. 1B environment includes multiple data processing components including remote server 108, user devices 110a-110b, concentrator 104; [0052] remote server 108 itself may include multiple processors (modularized data processing components)), each configured to perform a specific transformation on data (each of the processing components within the cloud environment depicted in FIGS. 1A and 1B perform some particularized data processing), and wherein the data analysis is performed in conjunction with at least one processor to generate a livestock-specific output (each of the processing components, including the multiple processors within remote server 108, within the cloud environment depicted in FIGS. 1A and 1B inherently perform processor-driven processing in conjunction to determine livestock condition per [0052]).
Singh further teaches normalizing an input dataset (aggregated data) for use with machine learning ([0064] and [0180] processors scale/normalize the dataset to facilitate machine learning in terms of prediction accuracy).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Singh’s teaching of transmitting livestock data to a cloud computing environment in which the livestock data is analyzed to the method taught by Kuper ‘010 as modified by Stroman, such that in combination the method includes transmitting the livestock data from a plurality of livestock producers to a cloud computing environment (e.g., cloud-based architecture as an input portal for Kuper ‘010 input data 110 and analyzes the data (e.g., modeling within Kuper ‘010 adaptive framework 100) within the cloud computing environment (e.g., the model-based processing would be within a cloud-based environment by virtue of receiving the inputs through a cloud-based network architecture).
The motivation would have been to provide the combined network connectivity and processing-resource provisioning flexibility characteristics of cloud computing to enhance livestock data processing as suggested by Singh.
It would also have been obvious to one of ordinary skill in the art before the filing date, to have applied Singh’s teaching of normalizing the input dataset to the method taught by Kuper ‘010 as modified by Stroman, such that in combination the method includes generating predictive outcomes by applying normalization in addition to the statistical correlations and aggregation incident to machine learning modeling.
The motivation would have been to optimize accuracy of output predictions by the artificial intelligence modeling as disclosed by Singh.
While as set forth above, the preamble language “to predict the outcome of a selected variable of interest and improve livestock health outcomes” is taught by Kuper ‘010, this feature recites an intended result/purpose that does not further characterize the function performed by the method and is therefore not given patentable weight.
Kuper ‘010 suggests, but does not expressly teach that the method further includes “transmitting the report to an end user ([0036] output data 170 (constituting “report”) including recommendations regarding livestock feeding and veterinary care. The formulation of a recommendation inferentially suggests a need for communicating such recommendation to a user (user at farm) that may implement the recommendation),” and does not appear to teach steps (e), (f), (h), and (i).
Regarding “transmitting the report to an end user” and steps (e), (f), (h), and (i) Stroman discloses (e) transmitting the report to an end user (FIG. 1 depicting networked systems including client computers 110 and server 107 providing access to data within central database 109 via links 115 that per [0073] may be wireless, [0070]; [0129] the adaptive reasoning system 140 (that is part of overall platform depicted in FIG. 1) electronically available at ranches; [0130] output from adaptive reasoning system 140 in the form of a notification; [0133]-[0134] user may retrieve analyzed information from the database via the Internet; [0104] “users” may include ranchers),” and “wherein the end user may accept or deny the report ([0133] recommendation sent to user may be reviewed by the user (i.e., user may confirm acceptance via reviewing or deny acceptance via not reviewing; [0129] adaptive reasoning system 140 recommends actions (transmission inherently required) and the system processes results of implementation of the recommendations (Examiner notes that user implementation of recommendations entails a form of “acceptance” of the recommendation/report));
(f) implementing the change to one or more variables in a field trial ([0022] user performs action to implement recommendation from adaptive reasoning system (i.e., implements some change in variable factor); FIG. 15 depicting adaptive reasoning system 140 as providing recommendation 1015 with corresponding action 1020 and 1030; [0133]-[0134]);”
“(h) analyzing the data produced from the field trial ([0129] and [0134] outcome of the recommended actions monitored and compared (analyzed) with respect to expected results) using at least one trained machine learning model ([0128]-[0129], and [0132] adaptive reasoning system implemented via artificial intelligence learning (training/trained model))” and
“(i) retraining the one or more of the analysis models ([0129] analytic and empirical models may be refined, improved, recalibrated. Examiner notes that improving and recalibrating models of the system that per [0128] is “adaptive” and has “learning capabilities” entails retraining the model).”
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Stroman’s teaching of transmitting an output report an end user (e.g., smart farm) wherein the user may accept or deny the report, and wherein upon acceptance of the report, the method includes implementing the change to one or more variables in a field trial, analyzing the data produced from the field trial, and retraining the one or more of the analysis models to the method taught by Kuper ‘010 as modified by Stroman and Singh such that Kuper ‘010 output data (output data 170) that includes livestock health-related recommendations is transmitted to a user (e.g., client at a smart farm) in a manner in which the user may accept or deny the report, and whereupon acceptance of the report, the method includes implementing the change to one or more variables in a field trial, analyzing the data produced from the field trial, and retraining the one or more of the analysis models.
The motivation would have been to provide the recommendation within the report to the onsite user (user at ranch/farm) to effectuate consideration of the recommendation by the user, who may accept or deny acceptance of the report and undertake logically corresponding actions such as implementing the change to one or more variables in a field trial, analyzing the data produced from the field trial, and retraining the one or more of the analysis models as disclosed by Stroman.
Regarding the trained machine learning model being “selected from the group consisting of ensemble learning, deep learning, time series modeling, and Bayesian statistics,” Stroman teaches that the model may be a neural network which frequently includes multiple layers (deep learning) [0132]) but does not disclose the type of neural network. Park discloses use of a deep learning (multi-layer) type of neural network that may be utilized for estimating livestock-related data (page 1, Claims (13), paragraph beginning with “An oil production forecasting method”; page 3, Description, paragraph beginning with “According to an embodiment of the present disclosure, the first prediction model is based on a Long-Short-Term Recurrent Neural Network (RNN) algorithm”. Examiner notes that an RNN is known as a multi-layer/deep learning type neural network model).
It would have been obvious to one of ordinary skill in the art before the filing date, to have applied Park’s teaching of an RNN as a type of neural network modeling (deep learning) to the method disclosed by Kuper ‘010 as modified by Stroman and Singh to include using neural network modeling, such that in combination the neural network modeling implements multi-layer modeling (e.g., RNN).
Such a combination would amount to selecting a known design option for implementing neural network modeling to achieve predictable results.
The combination of Kuper ‘010, Stroman, and Singh further teaches “wherein the step of analyzing the livestock data includes applying artificial intelligence to the livestock data in a plurality of data processing modules (Kuper ‘010: FIG. 1 adaptive framework 100 that processes input data 110 includes AI layer 146 and adaptive machine learning methods 147) within a cloud computing environment in which the plurality of data processing modules are executed (Singh: cloud-computing environment as combined with Kuper ‘010 for claim 7) in conjunction with at least one processor (Kuper ‘010: FIG. 1 computing environment 130 that includes AI layer 146 and machine learning methods 147 includes processors 132; [0016]), the data processing modules and artificial intelligence configured to continuously evaluate data from a plurality of livestock producers (Kuper ‘010: [0006]-[0007] real-time assessments of data inputs (Examiner notes that real-time processing of input data implicitly entails continuousness of such processing); claim 1 reciting continuous evaluation of input data) to estimate or predict an expected improvement in livestock health outcome by changing one or more variables (Kuper ‘010: FIG. 1 output data 170 that includes growth rate predictions 172 and corresponding livestock recommendations 173; [0033] modeling including machine learning used to predict and recommend livestock feeding and management (Examiner notes the predictions are determined based on (dependent on) modeling/processing of variable input data such that the predictions themselves constitute an estimate/prediction of expected improvement in which the “changing one or more variables” aspect of the predictions are manifested in corresponding recommendations).”
It should be noted that while the combination of Kuper ‘010, Stroman, and Singh teaches “to estimate or predict an expected improvement in livestock health outcome by changing one or more variables,” this feature recites an intended result/purpose (“to estimate or predict …” may be interpreted as “for the purpose of estimating or predicting …”) that does not further characterize the function performed by the method and is therefore not given patentable weight. Similarly, claim 7 recites a livestock-specific output “that enables automated feedback and iterative improvement in livestock health outcomes,” which also recites an intended result/purpose that does not further characterize the function performed by the method and is therefore not given patentable weight.
Claim 7 further recites a step “(g) building an artificial intelligence supervised learning model trained on livestock-specific real-time data obtained from the field trial” that does not appear to have any functional or structural relation with any other element in the claim. There is no clear antecedent relation between the “supervised learning model trained on livestock-specific real-time data” in element (g) and “at least one trained machine learning model” in element (h) or the “one or more analysis” models in element (c). Kuper ‘010 effectively discloses building an artificial intelligence model (FIG. 1 adaptive machine learning methods 147 (Examiner notes that a machine learning method is inherently implemented by some form of data structure (model) that has been built), [0033]), but is silent regarding whether the model is supervised or unsupervised and therefore does not teach that the artificial intelligence model “supervised learning model,” and furthermore does not itself expressly teach that the model has been “trained on livestock-specific real-time data obtained from the field trial.”
Supervised learning models were a well-known type of artificial intelligence model prior to the effective filing date. For example, Kopic discloses a method for analytic management of dairy production that includes generating a production model (Abstract) that may be a supervised machine learning model ([0076]).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Kopic’s teaching of generating a production model that may be a supervised machine learning model to the method taught by Kuper ‘010 as modified by Stroman, Singh, and Park such that the combined method includes building an artificial intelligence supervised learning model.
The motivation would have been to generate a type of artificial intelligence model that may be usefully employed in the context of monitoring livestock activity as disclosed by Kopic.
Regarding model being/having been “trained on livestock-specific real-time data obtained from the field trial,” Kuper ‘010 teaches obtaining real-time livestock data (FIG. 1 input data 110 relating to livestock includes nutrition, weather, animal-specific data, location, etc.; [0006]-[0007] real-time assessments of data; [0037] and [0039]-[0040] data inputs relating to livestock conditions are provided in real-time) to be processed by machine learning to determine livestock condition (FIG. 1 depicting data processing components 140 configured to analyze via modeling input data 110 including AI layer 146, adaptive machine learning methods 147 (machine learning inherently entails training/trained model), and selected growth model 160; [0015] AI layer 146 and selected model 160 generate output data 170 such as predicted growth rate 172, average gain 171 and recommendations 173 based on processing input data described in [0019]).
Stroman teaches that field trial data may be used to improve rules, logic of an adaptive reasoning system that may be implemented via machine learning ([0022] user performs action to implement recommendation from adaptive reasoning system (i.e., implements some change in variable factor) and resulting outcome information used to improve rules, logic; FIG. 15 depicting adaptive reasoning system 140 as providing recommendation 1015 with corresponding action 1020 and result 1025 that is provided as feedback 1030 to update rules 1010; [0129], [0132] adaptive reasoning system may implement artificial intelligence (learning) models such as neural networks; [0133]-[0134]).
It would have been obvious to one of ordinary skill in the art before the effective filing date, in view of Kuper ‘010 teaching of processing real-time livestock data by machine learning modeling and Stroman’s teaching of using field trial data for updating/training adaptive data processing models to have modified the method taught by Kuper ‘010 as modified by Stroman, Singh, Kopic, and Park such that the field trial data is collected as real-time data and used to train the supervised learning model.
The use of real-time collection of the field trial data would amount to selecting a known design option for collecting model-relevant data to achieve predictable results, and the motivation for using the real-time field trial data for model training would have been to leverage field trial data to improve modeling accuracy as suggested by Stroman.
As to claim 8, the combination of Kuper ‘010, Stroman, Singh, Kopic, and Park teaches “[t]he method of claim 7, wherein the selected variable of interest is selected from the group consisting of microbiome, feeding method (Kuper ‘010: Abstract disclosing adaptive framework method for modeling livestock growth including optimizing and recommending feed operations (feed operation being the variable of interest and outcome being growth); FIG. 2 including block 290, [0008] nutrition recommendations include feed rations and feed additives that are part of “feeding method”), feeding schedule, medical history, breed, breeding status, age, body condition, genetics, gender, environmental temperature, humidity, water pH, water temperature, water quality, ammonia levels in environment, air quality, barn flooring quality/type, feeder type, feed quality, feed size, nutritional levels and requirements, disease, medication, medication delivery method, distance of livestock from other barns and/or packing plants, growth performance, and carcass characteristics.”
As to claim 9, the combination of Kuper ‘010, Stroman, Singh, Kopic, and Park teaches “[t]he method of claim 7,” and Stroman further teaches that livestock data may be available from an Internet source serving animal information (“Internet of Animals”) (FIG. 26 depicting Internet Service Provider (hence Internet server architecture) in/over which livestock information is available; [0088] animal information may be obtained via the Internet).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Stroman’s teaching of using animal/livestock information stored at and served from the Internet (Internet of Animals) to provision at least some of the livestock data used for livestock analysis to the method taught by Kuper ‘010 as modified by Stroman and Singh, such that the combined method entails the livestock data further comprising data obtained from such Internet source (Internet of Animals).
The motivation would have been to provide a widely utilized networking source (the Internet) to which relevant animal/livestock information may be stored and readily obtained using standard networking protocols as suggested by Stroman.
As to claim 13, the combination of Kuper ‘010, Stroman, Singh, Kopic, and Park teaches “[t]he method of claim 7,” and as set forth in the grounds for rejecting claim 7 Kuper ‘010 teaches generating an output report including a recommendation (FIG. 1 output data 170 that includes average daily gain 171, growth rate predictions 172, and livestock recommendations 173).
In view of Kuper ‘010 teaching of generating an output report including recommendation, it would have been obvious to one of ordinary skill in the art before the effective filing date, to have sent the output report including recommendation to an end user for acceptance or denial as this step is inferentially a part of making a recommendation and would be required to effectuate the intent of making a recommendation (i.e., the motivation for sending the report/recommendation is inherent in the generating of the recommendation disclosed by Kuper ‘010).
Kuper ‘010 further teaches repeating steps in the acquisition and analysis of livestock data (obtaining, transmitting, and analyzing livestock data as recited in steps (a)-(c)) until one or more variables dependent on improvement through several simulations is fulfilled (FIG. 1 feedback loop 162, [0024] framework 100 incorporates a feedback loop for validating models and consequently obtaining improved predictions from the model(s) such that the framework is self-improving over time (i.e., as more data is processed); [0007] feedback is provided in an ongoing manner as real-time information received).
In view of Kuper ‘010 teaching of a repeating process via feedback in which the analysis process, which includes collection and transmission of new livestock information, is repeated on an ongoing manner, it would have been obvious to one of ordinary skill in the art before the effective filing date, to have also repeated the steps including the reporting step (step (d)) as part of the overall process regarding of whether the user accepts or denies the report.
Such a combination would be a logical extension of the data collection/analysis/reporting process that is performed in an ongoing manner using real-time data to improve and/or otherwise adjust recommendations based on changes in the input data necessitating modeling adjustments as suggested by Kuper ‘010.
As to claim 22, the combination of Kuper ‘010, Stroman, Singh, Kopic, and Park teaches “[t]he method of claim 7,” and as set forth in the grounds for rejecting claim 7, the combination of Kuper ‘010 and Kopic teaches building the “artificial intelligence supervised learning model.”
Kuper ‘010 and Kopic are largely silent regarding the particular type of supervised learning model and therefore neither expressly discloses “wherein the artificial intelligence supervised learning model comprises one or more machine learning models selected from the group consisting of decision trees, regression trees, random forests, gradient boosting machines, support vector machines, neural networks, and Bayesian networks.”
Stroman teaches application of neural networks as a type of machine learning model/algorithm applicable in the general context of livestock monitoring ([0132], claim 5).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Stroman’s teaching of the utility of using neural networks for monitoring/processing livestock related data to the method taught by Kuper ‘010 as modified by Stroman, Singh, Kopic, and Park, such that in combination the artificial intelligence supervised learning model comprises a neural network.
Such a combination would amount to selecting a known design option for an artificial intelligence supervised learning model to achieve predictable results.
As to claim 23, the combination of Kuper ‘010, Stroman, Singh, Kopic, and Park teaches “[t]he method of claim 7, wherein the step of retraining one or more of the analysis models comprises modifying at least one model parameter (Stroman: [0129] analytic and empirical models may be refined, improved, recalibrated. Examiner notes that modifying the models necessarily and inherent entails modifying a model parameter) based on performance data from the field trial using a semi-supervised or transfer learning approach.”
Regarding the analysis models being trained (parameters modified) “based on performance data from the field trial,” Stroman teaches that field trial performance data may be used to improve rules, logic of an adaptive reasoning system that may be implemented via machine learning ([0022] user performs action to implement recommendation from adaptive reasoning system (i.e., implements some change in variable factor) and resulting outcome information used to improve rules, logic; FIG. 15 depicting adaptive reasoning system 140 as providing recommendation 1015 with corresponding action 1020 and result 1025 that is provided as feedback 1030 to update rules 1010; [0129], [0132] adaptive reasoning system may implement artificial intelligence (learning) models such as neural networks; [0133]-[0134]).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Stroman’s teaching of using field trial data for updating/training adaptive data processing models to the method taught by Kuper ‘010 as modified by Stroman, Singh, Kopic, and Park such that the retraining including modifying of a model parameter is based on performance data from the field trial.
The motivation for using the field trial performance data for model training/parameter updating would have been to leverage field trial data to improve modeling accuracy as suggested by Stroman.
Regarding the training/learning technique being “a semi-supervised or transfer learning approach,” Kopic further teaches that semi-supervised learning is an option for training/retraining models ([0076) semi-supervised learning for updating model].
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Kopic’s teaching of using semi-supervised learning/training for updating/training models to the method taught by Kuper ‘010 as modified by Stroman, Singh, Kopic, and Park such that a semi-supervised learning approach is used for training/parameter update for the model.
Such a combination would amount to selecting a known design option for updating/training a model to achieve predictable results.
As to claim 25, the combination of Kuper ‘010, Stroman, Singh, Kopic, and Park teaches “[t]he method of claim 7, wherein the customized report includes a prediction of improvement in at least one performance metric selected from feed conversion ratio, mortality rate, average daily gain (Kuper ‘010: [0005]; FIG. 1 output data 170 that includes average daily gain 171, growth rate predictions 172, and livestock recommendations 173 (Examiner notes the output data constitutes a report that is customized in terms of being dynamically adjusted in accordance with changing input data 110 and per [0026] in accordance with model weight adjustments)), or carcass yield, based on changing one or more variables (FIG. 1 output data 170 that includes growth rate predictions 172 and corresponding livestock recommendations 173; [0033] modeling including machine learning used to predict and recommend livestock feeding and management (Examiner notes the predictions are determined based on (dependent on) modeling/processing of variable input data such that the predictions themselves constitute an estimate/prediction of expected improvement in which the “changing one or more variables” aspect of the predictions are manifested in corresponding recommendations).
Claims 1, 3, 5, and 21 are rejected under 35 U.S.C. 103 as being unpatentable over Kuper ‘010, Stroman, Singh, Kopic, and Park as applied to claim 7 above, and further in view of Perry (US 2019/0050948 A1).
As to claim 1, Kuper ‘010 teaches “[a] method of conducting digital research to improve livestock health outcomes (Abstract disclosing adaptive framework method for modeling livestock growth including optimizing and recommending feed operations; FIG. 2), the method comprising
forming, using a processor (FIG. 1 processors 132), a hypothesis regarding what changes may be made to one or more variables related to livestock nutrition, environment, health status, or animal performance to improve livestock health outcomes for a particular livestock producer (FIG. 1 ensemble of models 150 in combination with AI layer 146 and adaptive machine learning methods 147, [0015] input data applied to models 150 for determining and predicting livestock growth rate; [0023] adaptive machine learning methods used to predict and recommend livestock feeding and management based on input data 110 (Examiner notes the models themselves constitute a formulated hypothesis regarding how livestock outcomes are affected by livestock-related input variables such that the models by their function inherently demonstrate/predict how a change in variable input data affects livestock health outcomes); [0028] results of comparison of predicted performance and actual performance may be used to adjust the weights of the model (the premise/hypothesis represented by the model) to improve modeling accuracy),”
“performing, using a statistical or machine learning algorithm, a” “analysis of historical data related to the one or more variables ([0005] and [0023] workflow processing is optimized using machine learning methods in accordance with particular facility and producer (farm) that are evaluated based on historical performance, genetics, and management practices of producer) to determine one or more” “farms for real-time livestock data collection ([0023] and [0035] workflow processing optimized in accordance with particular facility and producer (farm). In this manner of optimizing/customizing workflow in accordance with producer (farm) historical data, the producers/farms are themselves determined as part of performing the analysis/evaluation);
obtaining real-time livestock data (FIG. 1 input data 110 relating to livestock includes nutrition, weather, animal-specific data, location, etc.; [0006]-[0007] real-time assessments of data; [0037] and [0039]-[0040] data inputs relating to livestock conditions are provided in real-time) from one or more sensors (FIG. 1 input data 110 including scale data 127, weather information 115, and GPS/tracking 121)” “at the determined one or more” “farms ([0023] and [0035] workflow processing optimized in accordance with particular facility and producer (farm));
analyzing the real-time livestock data using a trained artificial intelligence model (FIG. 1 depicting data processing components 140 configured to analyze via modeling input data 110 including AI layer 146, adaptive machine learning methods 147 (machine learning inherently entails training/trained model), and selected growth model 160; [0015] AI layer 146 and selected model 160 generate output data 170 such as predicted growth rate 172, average gain 171 and recommendations 173 based on processing input data described in [0019]) to test the hypothesis ([0026] the predictions resulting from modeling analysis may be tested such as by comparison with actual results; [0028] results of comparison of predicted performance and actual performance may be used to adjust the weights of the model (the premise/hypothesis represented by the model));
generating” “a customized research report that includes results of the analysis of the real-time livestock data (FIG. 1 output data 170 that includes average daily gain 171, growth rate predictions 172, and livestock recommendations 173 (Examiner notes the output data constitutes a report that is customized in terms of being dynamically adjusted in accordance with changing input data 110 and per [0026] in accordance with model weight adjustments)) including one or more recommendations to improve livestock health outcomes (FIG. 1 livestock recommendations 173; [0008] recommendations include recommended nutrition; [0032]).”
Kuper ‘010 does not expressly teach that the customized research report is generated “via a cloud-based processing module.”
Singh discloses a method for tracking health in animal populations (Abstract) in which generating and processing of livestock data report data is performed via a cloud-based processing module (FIGS. 1A-1B depicting network-based computing environment comprising network 106 as information portal between user devices 110, tag assemblies 102 via concentrator 104, and remote server 108; [0052] remote server generates livestock data that per FIG. 4 blocks 414 and 416 is reported; [0035] network 106 may comprise a cloud-based architecture).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Singh’s teaching that livestock report data may be generated via a cloud-based processing module to the method taught by Kuper ‘010 such that in combination the method includes generating the customized research report via a cloud-based processing module.
The motivation would have been to provide the combined network connectivity and processing-resource provisioning flexibility characteristics of cloud computing to enhance livestock data processing as suggested by Singh.
Kuper ‘010 teaches utilizing artificial intelligence and machine learning for analyzing livestock related input data such as location data and facility information (FIG. 1 feedlot location 120, GPS tracking 121, and facility type 117) but does not appear to expressly teach that the analysis of historical data related to the one or more variables is a “cluster” analysis, characterized in Applicant’s specification as potentially entailing artificial intelligence and/or machine learning.
Perry discloses a machine learning based agricultural method for evaluating historical agricultural data to predict particular outcomes (Abstract) that includes using grouping/clustering analysis to determine the collection of agricultural input information to be used for predictive analysis (Abstract disclosing selective access of agricultural condition data (i.e., data relevant to crop health) corresponding to request including parameters such as location, weather, and soil composition; [0004] and [0007] crop growth information including multiple growth-related characteristics collected (historical) and then grouped/clustered in accordance with similarities based on a particular request; [0009] growth information collected from a plurality of fields/locations associated with (grouped/clustered in accordance with) a threshold geographic and/or environmental diversity; [0017]-[0019] accessed field information (historical because it has been collected) mapped by prediction model to a selected set of farming operations (i.e., clustered in accordance with farming operation parameters)).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Perry’s teaching of selecting sets of collected agricultural input information based on threshold matching with agricultural farm parameters (grouping/clustering) as a selection means for collecting relevant input information to the method taught by Kuper ‘010, which teaches real-time collection of livestock information from one or more farms, such that in combination the method includes performing a “cluster analysis” of historical data related to the one or more variables to determine one or more smart farms for real-time livestock data collection.
The motivation would have been to enable selective collection of data across one or more farms having shared characteristics that are determined either by request or otherwise to be most relevant to the subsequent processing analysis of agricultural/livestock health including growth as disclosed by Perry.
Regarding determining one or more “smart” farms for real-time livestock data collection and obtaining real-time livestock data from the selected one or more “smart” farms, Kuper ‘010 does not appear to clearly characterize the subject “producers” (farms) as being “smart farms.”
Perry further discloses that the agricultural monitoring method including collecting agricultural information from farm regions equipped with automated monitoring/sensing devices and thus constitute a “smart” farm ([0010] and [0017]).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Perry’s teaching of applying agricultural monitoring processes for smart farms to the method taught by Kuper ‘010 such that the farms (producers) from which livestock data is obtained are smart farms. The motivation would have been to leverage the automated data gathering capabilities of smart farms to improve data gathering efficiency as disclosed by Perry.
It should be noted that, while as set forth above, the preamble language “to improve livestock health outcomes” is taught by Kuper ‘010, this feature recites an intended result/purpose that does not further characterize the function performed by the method and is therefore not given patentable weight.
It should be noted that, while as set forth above, “to determine one or more smart farms for real-time livestock data collection” is taught by the combination of Kuper ‘010 and Perry, this feature recites an intended result/purpose (“to determine …” may be interpreted as “for the purpose of determining …”) that does not further characterize the function performed by the method and is therefore not given patentable weight. The claim may intend to convey that the method include a step of determining/selecting one or more smart farms for real-time livestock data collection that is based on “performing a cluster analysis of historical data related to the one or more variables.” If so, the claim language must be amended to clearly convey the “determining” step as a positive limitation. For example, claim 1 may be amended to recite
“performing a cluster analysis of historical data related to the one or more variables”
“determining (or selecting) one or more smart farms for real-time livestock data collection based on the cluster analysis.”
As set forth above, the combination of Kuper ‘010 and Perry teaches “obtaining real-time livestock data from one or more sensors (Kuper ‘010: FIG. 1 input data 110 including scale data 127, weather information 115, and GPS/tracking 121),” and Stroman discloses a livestock monitoring method (Abstract) in which the ranches (animal farms) subject to (determined for) livestock monitoring in which the livestock management method includes obtaining real-time livestock data from “one or more sensors” “located on or around a plurality of livestock animals” at a smart farm (FIG. 2 blocks 210 and 215, [0081] RFID tags used to detect, identify, and otherwise collect information about one or more animals. Examiner notes that RFID tag detection inherently utilizes tag and reader for sensing such that each pair formed by a respective tag on an animal and the reader constitutes an individual “sensor” such that for multiple sensors result from multiple tagged animals; [0097]-[0099] rectal thermometers and a digital camera may be included as sensors along with RFID sensors).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Stroman’s teaching of using multiple, on-site sensors (sensors on or around the animals) to collect livestock information for livestock management processing to the method taught by Kuper ‘010 as modified by Perry.
The motivation would have been to provide more comprehensive sensor coverage such as in terms of geographical coverage and data type coverage that may be individually correlated to each of multiple individual animals as disclosed by Stroman.
Regarding “transmitting the research report wirelessly to the one or more smart farms;” “confirming or denying acceptance of the research report by the one or more smart farms,” and “implementing the one or more recommendations by the smart farm,” Kuper ‘010 suggests, but does not expressly teach that the method further includes, “transmitting the research report” “to the one or more smart farms ([0036] output data 170 (constituting “report”) including recommendations regarding livestock feeding and veterinary care. The formulation of a recommendation inferentially suggests a need for communicating such recommendation to a place (farm) at which the recommendation may be performed).”
Stroman discloses transmitting an output report wirelessly to the one or more smart farms (FIG. 1 depicting networked systems including client computers 110 and server 107 providing access to data within central database 109 via links 115 that per [0073] may be wireless, [0070]; [0129] the adaptive reasoning system 140 (that is part of overall platform depicted in FIG. 1) electronically available at ranches; [0130] output from adaptive reasoning system 140 in the form of a notification; [0133]-[0134] user may retrieve analyzed information from the database via the Internet; [0104] “users” may include ranchers), confirming or denying acceptance of the report by the one or more smart farms ([0133] recommendation sent to user may be reviewed by the user (i.e., user may confirm acceptance via reviewing or deny acceptance via not reviewing; [0129] adaptive reasoning system 140 recommends actions (transmission inherently required) and the system processes results of implementation of the recommendations (Examiner notes that user implementation of recommendations entails a form of “acceptance” of the recommendation/report)), and implementing the one or more recommendations by the smart farm ([0022] user performs action to implement recommendation from adaptive reasoning system (i.e., implements some change in variable factor); FIG. 15 depicting adaptive reasoning system 140 as providing recommendation 1015 with corresponding action 1020 and 1030; [0133]-[0134]).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Stroman’s teaching of transmitting an output report wirelessly to the one or more smart farms, confirming or denying acceptance of the research report by the one or more smart farms, and implementing the one or more recommendations (in the report) to the method taught by Kuper ‘010 as modified by Perry and Stroman such that Kuper ‘010 output data (output data 170) that includes livestock health-related recommendations is transmitted wireless to a smart farm in a manner in which the smart farm may confirm or deny acceptance of the report and further determine to implement the recommendation(s).
The motivation would have been to provide the recommendation within the report to the onsite user (user at ranch/farm) who may be remote (hence wireless transmission) to effectuate consideration of the recommendation by the user, who may accept or deny acceptance of the report such as by or including implementation as disclosed by Stroman.
Regarding “transmitting data related to an outcome of implementation of the one or more recommendations to a cloud configuration or central database, wherein the transmitted data is made available for use as historical data in a virtual research model,” Stroman further teaches transmitting data related to an outcome of implementation of one or more recommendations (FIG. 15 depicting adaptive reasoning system 140 in which Result 1025 data resulting from an Action 1020 that is based on a Recommendation 1015 is transmitted as Feedback 1030 to step 1010)” to a central database (step 1010 depicting Expert Rules, Corrective Actions, Recommendations, Best Practices in bi-directional communicative contact with a Central Database; [0128] “Specific health regimens and their causes and effects may entered into … adaptive reasoning system 140”), “wherein the transmitted data is made available for use as historical data ([0128] “Specific health regimens and their causes and effects may entered into … adaptive reasoning system 140, which allows the system to learn and pass the information on to the users; [0129] historical information archived in the central database) in a virtual research model (FIG. 15 adaptive reasoning system 140 includes remotely accessible data storage (Central Database), processing functionality (Expert Rules, Diagnosis, Corrective Actions, Recommendations, Best Practices), and multivariate inputs (step 105) that constitute and implement a virtual research model in terms of the step 1010 implementing recommendations and modifying actions based on results (i.e., data modeling that implements research).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Stroman’s teaching of transmitting data related to an outcome of implementation of one or more recommendations to a central database, wherein the transmitted data is made available for use as historical data in a virtual research model to the method taught by Kuper ‘010 as modified by Stroman, Singh, Kopic, Park, and Perry such that in combination the method includes transmitting data related to an outcome of implementation of the one or more recommendations to a cloud configuration or central database, wherein the transmitted data is made available for use as historical data in a virtual research model.
The motivation would have been to store implementation outcome data to a widely accessible storage from which it may be accessed and usefully applied to improve future recommendations as suggested by Stroman.
The element “wherein upon implementation of the one or more recommendations by the smart farm, livestock health outcomes are improved” recites an intended result/purpose that does not further characterize the function performed by the method and is therefore not given patentable weight.
Regarding “the hypothesis based on simulations produced according to the virtual research method of claim 7”, Kuper ‘010 teaches adjustment to the models (hypothesis) based on the simulations in a feedback manner ([0028] results of comparison of predicted performance and actual performance may be used to adjust the weights of the model (the premise/hypothesis represented by the model) to improve modeling accuracy).
Furthermore, and per the grounds for rejecting claim 7, simulations based on hypotheses performed according to virtual research method of claim 7 are taught by the combination of Kuper ‘010, Stroman, Singh, Kopic, and Park.
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have combined the teaching of Kuper ’010 of determining (e.g., by adjustment) the hypothesis based on simulations with the teachings of Kuper ‘010, Stroman, Singh, and Kopic such that the simulations produced according to the virtual research method of claim 7 are utilized to form the hypothesis.
Such a combination would amount to applying a known design option for formulating and applying a hypothesis regarding changes to variables may affect livestock performance to achieve predictable results.
As to claim 3, the combination of Kuper ‘010, Stroman, Singh, Kopic, Park, and Perry teaches “[t]he method of claim 1, wherein the one or more sensors transmits a particular code associated with an individual livestock animal (Stroman: FIG. 2 block 215 depicting electronic identification of animals based on RFID detection at block 210, [0028] RFID detection/identification/data collection; [0097]-[0099] RFID tagging using RFID tags on animals. Examiner notes that RFID tag sensing inherently entails sending an RF identification code to a reader) or individual smart farm.”
As to claim 5, the combination of Kuper ‘010, Stroman, Singh, Kopic, Park, and Perry teaches “[t]he method of claim 1,” and Kuper ‘010 further suggests “wherein the cluster analysis utilizes historic data obtained from a plurality of smart farms located throughout one or more global regions ([0005], [0023], and [0035] disclosing that producer-specific information utilized to optimize workflow, which suggests discerning among multiple producers (farms)).”
Furthermore, Stroman teaches collecting historical information relating to factors affecting livestock condition from multiple smart farms ([0002] Internet-based platform shared by members of livestock production chain including producers; FIGS. 1 and 8 tracking system 145; [0016] tracking system used to document information from different ranches (Examiner notes that the information relates to events that have occurred and are thus “historical”); [0098]-[0099] ranches are smart farms in terms of using automated livestock information collection) located throughout one or more global regions (Examiner notes that ranches are inherently located on at least one or more global regions).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have combined Stroman’s teaching of collecting historical health-related livestock data from multiple smart farms with Kuper ‘010 teaching of analyzing historical livestock data from potentially differing farms, such that the combined method includes the cluster analysis utilizing historic data obtained from a plurality of smart farms located throughout one or more global regions.
The motivation would have been to expand the comprehensive coverage of historical livestock data available for processing and to provide access within a centralized livestock monitoring platform as disclosed by Stroman.
As to claim 21, the combination of Kuper ‘010, Stroman, Singh, Kopic, and Park teaches “[t]he method of claim 7, wherein the step of obtaining livestock data comprises receiving real-time sensor data (Kuper ‘010: FIG. 1 input data 110 relating to livestock includes nutrition, weather, animal-specific data, location, etc.; [0006]-[0007] and [0020] real-time assessments of data; [0037] and [0039]-[0040] data inputs relating to livestock conditions are provided in real-time; FIG. 1 input data 110 including scale data 127, weather information 115, and GPS/tracking 121) from a” “farm[s] (Kuper ‘010: [0023] and [0035] workflow processing optimized in accordance with particular facility and producer (farm)), the sensor data including at least one of environmental temperature (Kuper ‘010: [0020] real-time weather information may include temperature), humidity (Kuper ‘010: [0021] weather information may include humidity), water pH, water temperature, ammonia levels, or animal body weight.”
Regarding the sensor data being received from “a plurality” of smart farms, Kuper ‘010 further suggests automatically obtaining livestock data from “a plurality of” livestock producers ([0005], [0023], and [0035] disclosing that producer-specific information utilized to optimize workflow, which suggests discerning among multiple producers (farms)), and Stroman teaches automatically obtaining livestock data from multiple livestock producers ([0002] Internet-based platform shared by members of livestock production chain including producers; FIGS. 1 and 8 tracking system 145; [0016] tracking system used to document information from different ranches).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have combined Stroman’s teaching of collecting livestock-related data from multiple farms/livestock producers with the method taught by Kuper ‘010 in which the data is real-time sensor data, such that in combination the real-time sensor data is received from a plurality of farms.
The motivation would have been to expand the comprehensive coverage of livestock data available for processing such as to more accurately determine monitoring processing results from such varied information as disclosed by Stroman.
Regarding the farms being “smart” farms from which the real-time sensor data is received, Kuper ‘010 does not appear to clearly characterize the subject “producers” (farms) as being “smart farms.”
Perry further discloses that the agricultural monitoring method including collecting agricultural information from farm regions equipped with automated monitoring/sensing devices and thus constitute a “smart” farm ([0010] and [0017]).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Perry’s teaching of applying agricultural monitoring processes for smart farms to the method taught by Kuper ‘010 such that the farms (producers) from which the real-time sensor data is received are smart farms. The motivation would have been to leverage the automated data gathering capabilities of smart farms to improve data gathering efficiency as disclosed by Perry.
Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Kuper ‘010 in view of Stroman, Singh, Kopic, Park, and Perry as applied to claim 1 above, and further in view of Kuper (US 10,628,756 B1), “Kuper ‘756.”
As to claim 4, the combination of Kuper ‘010, Stroman, Singh, Kopic, Park, and Perry teaches “[t]he method of claim 1, wherein the step of analyzing the real-time livestock data is accomplished by one or more statistical models (Kuper ‘010: FIG. 1 data processing components for processing/analyzing input data 110 include AI layer 146 and adaptive machine learning methods 147 (Examiner notes that AI inherently entails statistical processing (probabilities) consistent with description of statistical models in Applicant’s specification).”
Furthermore, even if Kuper ‘010 disclosure is considered not to expressly teach analyzing livestock data by statistical models, Kuper ‘756 explicitly teaches using statistical analysis for analyzing input livestock data (FIG. 1 statistical analysis 153 applied to input data, col. 2 lines 15-24, col. 4 lines 40-53, col. 8 lines 14-23).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Kuper ‘756 teaching of using statistical-type modeling for analyzing livestock data to the method taught by Kuper ‘010 as modified by Stroman, Singh, Kopic, Park, and Perry such that the analysis of livestock data includes using statistical analysis. Such a combination would amount to selecting a known design option for analyzing livestock data to achieve predictable results.
Claim 24 is rejected under 35 U.S.C. 103 as being unpatentable over Kuper ‘010 in view of Stroman, Singh, Kopic, and Park as applied to claim 7 above, and further in view of Guim Bernat (US 2019/0228326 A1).
As to claim 24, the combination of Kuper ‘010, Stroman, Singh, Kopic, and Park teaches “[t]he method of claim 7,” but does not appear to teach “storing the retrained model in a version-controlled model registry accessible within the cloud computing environment, wherein subsequent analyses utilize the updated model version.”
As set forth in the grounds for rejecting claim 7, Stroman teaches retraining the analysis models. Stroman further teaches using a form of registry in terms of a centralized database (FIG. 1 central database 109; FIG. 15 central database coupled with Expert Rules) from which users may track various information (FIG. 9 block 340), and which serves a comprehensive storage service ([0070] used for storing performance and history data and any other suitable information), and which may be characterized as “version-controlled” ([0088] users having particularized access to entries for particular animals; [0095] tracking system (used by the database) uses versioning in terms of animal and stage of development).
Stroman does not appear to specifically discloses that models (trained/retrained) may be stored in the database.
Storing models in registries/databases such that the models are available for subsequent access was well-known in the art prior to the effective filing date.
For example, Guim Bernat discloses a method for distributing stored data (Abstract) in a cloud computing environment ([0022]) in which artificial intelligence/machine learning models are trained ([0027] and [0039]) and in which retrained models are stored in a version-controlled registry ([0041] and [0086]) that per [0022] are accessible in a cloud computing environment for subsequent analysis (Examiner notes that the version-based storage of updated/retrained models inherently entails accessibility for subsequent use).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Guim Bernat’s teaching of storing retrained models in a version-controlled registry accessible in a cloud computing environment such that the updated model version may be subsequently utilized to the method taught by Kuper ‘010 as modified by Stroman, Singh, Kopic, and Park, such that in combination the method includes storing the retrained model in a version-controlled model registry accessible within the cloud computing environment, wherein subsequent analyses utilize the updated model version.
The motivation would have been to provide systematic, version-specific storage of retrained models to provide users with more optimized access to particular model versions in terms of training/retraining as suggested by Guim Bernat.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MATTHEW W BACA whose telephone number is (571)272-2507. The examiner can normally be reached Monday - Friday 8:00 am - 5:30 pm.
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, Catherine T Rastovski can be reached at (571) 270-0349. 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.
/MATTHEW W. BACA/Examiner, Art Unit 2863
/Catherine T. Rastovski/Supervisory Primary Examiner, Art Unit 2863
Prosecution Insights