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 .
Acknowledgements
This communication is in response to Request for Continued Examination filed on 12/22/2025.
Claims 1, 8, 15, 21-23 are amended.
Claims 5, 12, 18 are canceled.
Claims 1-4, 6-11, 13-17, 19-23 are currently pending and have been examined.
Claims 1-4, 6-11, 13-17, 19-23 have been rejected as follows.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 12/22/2025 has been entered.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 07/28/2023 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
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 1-4, 6-11, 13-17, 19-23 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Claims 1, 8, 15 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The claim recites a method, system, and non-transitory computer-readable medium storing instructions that, when executed by one or more processors of a computer, cause said one or more processors to perform a method for predicting responses to media.
The limitations of receiving a time series data set comprising media channel variables specifying media delivered to recipients via a plurality of media channels at a plurality of times and a dependent variable specifying one or more responses at the plurality of times; receiving, for each of the plurality of media channels, one or more lookback parameters, each lookback parameter specifying a time lag for a corresponding one of said plurality of media channels; generating, based on said one or more lookback parameters, one or more lagged features from the media channel variables, each lagged feature representing a corresponding media channel variable shifted by the time lag specified by a respective one of said one or more lookback parameters […] using said one or more lagged features in combination with the dependent variable, the […] comprising splitting the time series data into subsets based on media channel of the plurality of media channels and corresponding lagged features; and generating response curves using predictions from the […] based on the lagged features, each of the response curves corresponding to a media channel of the plurality of media channels, the response curves adapted to predict responses based on media delivered and media channel., as drafted, is a process that, under the broadest reasonable interpretation, covers certain methods of organizing human activity (i.e., Commercial or legal interactions, Advertising, Marketing, managing personal behavior including following rules or instructions) but for recitation of generic computer components. That is, other than reciting a system implemented by a computer having one or more processors in communication with a memory (computer), the claimed invention amounts to managing personal behavior or interaction between people. For example, but for a computer having one or more processors in communication with a memory, this claim encompasses a person looking at data, generating response curves based on the data, and predicting responses based on media in the manner described in the identified abstract idea, supra. The Examiner notes that certain “method[s] of organizing human activity” includes a person’s interaction with a computer (see MPEP 2106.04(a)(2)(II)). If a claim limitation, under its broadest reasonable interpretation, covers managing personal behavior or interactions between people but for the recitation of generic computer components, then it falls within the “certain methods of organizing human activity” grouping of abstract ideas. Accordingly, the claim recites an abstract idea.
Step 2A2
This judicial exception is not integrated into a practical application. In particular, the claim recites the additional element of (claim 8) a computer having one or more processors in communication with a memory, and (claim 15) A non-transitory computer-readable medium storing instructions that, when executed by one or more processors of a computer, cause said one or more processors to perform a method that implements the identified abstract idea. The processors and memory are not described by the applicant and is recited at a high-level of generality (i.e., a generic computer performing a generic computer functions of computing, determining, and selecting) such that it amounts no more than mere instructions to apply the exception using a generic computer component. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. The claim is directed to an abstract idea.
The claim further recites the additional element of using a trained random forest model to generate response curves. This represents mere instructions to implement the abstract idea on a generic computer. Implementing an abstract idea using a generic computer or components thereof does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. Alternatively or in addition, the implementation of the trained random forest model to generate response curves merely confines the use of the abstract idea (i.e., the trained model) to a particular technological environment or field of use (random forest) and thus fails to add an inventive concept to the claims.
The claim further recites “training a random forest model”. When given its broadest reasonable interpretation in light of the disclosure, the training of a machine learning model by splitting time series data represents the creation of mathematical interrelationships between data. As such, the training of the machine learning model represents a mathematical concept that is interpreted to be part of the identified abstract idea, supra. The types of identified abstract ideas are considered together as a single abstract idea for analysis purposes.
Step 2B
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional element of using a processor and memory to perform the noted steps amounts to no more than mere instructions to apply the exception using a generic computer component. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept (“significantly more”).
As discussed above with respect to integration of the abstract idea into a practical application, the additional element of using the trained random forest model to generate response curves was found to represent mere instructions to implement the abstract idea on a generic computer and confine the use of the abstract idea (i.e., the trained model) to a particular technological environment or field of use (random forest). This has been re-evaluated under the “significantly more” analysis and determined to be insufficient to provide significantly more. MPEP 2106.05(I) indicates that mere instructions to implement the abstract idea on a generic computer and/or confining the use of the abstract idea to a particular technological environment or field of use cannot provide significantly more. Accordingly, even in combination, this additional element does not provide significantly more. As such the claim is not patent eligible.
Dependent Claims
Claims 2-4, 6-7, 9-11, 13-14, 16-17, 19-23 are similarly rejected because they either further define/narrow the abstract idea and/or do not further limit the claim to a practical application or provide as inventive concept such that the claims are subject matter eligible even when considered individually or as an ordered combination. Claims 2, 9 merely describe the recipients. Claims 4, 11, 17 merely describe the responses. Claims 6, 13, 19 merely describe the media channels. Claims 7, 14, 20 merely describe a lookback parameter for the responses.
Claims 3, 10, 16, 21-23 also includes the additional element of “a random forest model” which is analyzed the same as in the independent claim and does not provide a practical application or significantly more for the same reasons. Claims 3, 10, 16 merely describe the response curves and a component of the model. Claims 21-23 merely describes a first and second lookback parameter as a component of the model.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1, 4, 7, 8, 11, 14, 15, 17, 20-23 are rejected under 35 U.S.C. 103 as being unpatentable over Arunachalam (US 20130166351) in view of Fan (US 20210209486)
CLAIM 1, 8, 15
Arunachalam teaches a method of generating a media mix model, the method comprising: (Arunachalam para 19 teaches system and method for generating a marketing mix solution. Para 56 teaches non-volatile memory. Para 58 teaches implementation of method through computer readable instructions)
receiving a time series data set comprising media channel variables specifying media delivered to recipients via a plurality of media channels at a plurality of times and a dependent variable specifying one or more responses at the plurality of times; (Arunachalam para 23 teaches time series data and media channel variables delivered over time and the sales data over time)
receiving, for each of the plurality of media channels, one or more lookback parameters, each lookback parameter specifying a time lag for a corresponding one of said plurality of media channels; (Arunachalam para 32 teaches marketing-mix variables may include an ad-stock variable, an event variable, a lead variable and a lag variable)
generating, based on said one or more lookback parameters, one or more lagged features from the media channel variables, each lagged feature representing a corresponding media channel variable shifted by the time lag specified by a respective one of said one or more lookback parameters; (Arunachalam para 34 teaches a lag variable to capture lag effects coming from media activities)
[…] model using said one or more lagged features in combination with the dependent variable, […] comprising splitting the time series data into subsets based on media channel of the plurality of media channels and corresponding lagged features; (Arunachalam para 35 teaches generating the response model using linear regression, non-linear regression, or mixed models where dependent variable is sales or revenue and independent variables are marketing-mix variables. Para 32 teaches marketing-mix variables may include an ad-stock variable, an event variable, a lead variable and a lag variable. Fig 6, Fig. 9 show subset of timeseries data for media channels)
generating response curves using predictions from the […] model based on the lagged features, each of the response curves corresponding to a media channel of the plurality of media channels, the response curves adapted to predict responses based on media delivered and media channel. (Arunachalam para 38 teaches generating plots to compare individual contributions for each of the marketing-mix variables, para 32 teaches marketing-mix variable includes lag, for each of the media channels. Para 39 teaches forecasting sales/revenue based on marketing mix of channels.)
Arunachalam does not teach
training a random forest model using said one or more lagged features in combination with the dependent variable, the training comprising splitting the time series data into subsets based on media channel of the plurality of media channels and corresponding lagged features;
generating response curves using predictions from the trained random forest model based on the lagged features, each of the response curves corresponding to a media channel of the plurality of media channels, the response curves adapted to predict responses based on media delivered and media channel.
Fan does teach
training a random forest model using said one or more lagged features in combination with the dependent variable, the training comprising splitting the time series data into subsets based on media channel of the plurality of media channels and corresponding lagged features; (Fan para 14 teaches splitting time series data set into a training and test dataset. Para 39 teaches a Random Forest as a machine learning model)
generating response curves using predictions from the trained random forest model based on the lagged features, each of the response curves corresponding to a media channel of the plurality of media channels, the response curves adapted to predict responses based on media delivered and media channel. (Fan para 14 teaches splitting time series data set into a training and test dataset. Para 39 teaches a Random Forest as a machine learning model. Para 47 teaches predicting outcomes using a random forest)
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the model (which may be linear regression, non-linear regression, or mixed models) as taught by Arunachalam with the trained random forest as taught by Fan. It would be beneficial for the model used to forecast time series data as taught by Fan para 38.
CLAIM 4, 11, 17
Arunachalam teaches wherein said one or more responses comprise at least one of: sales values […] (Arunachalam para 19 teach generating response for sales and or revenue. Additional limitation interpreted as optional due to claim language “at least one of …”)
CLAIM 7, 14, 20
Arunachalam teaches further comprising specifying a lookback parameter defining a lag in the one or more responses. (Arunachalam para 32 teaches marketing-mix variables may include an ad-stock variable, an event variable, a lead variable and a lag variable. )
CLAIM 21, 22, 23
Arunachalam teaches wherein said […] comprises specifying, for each of the plurality of media channels, a first lookback parameter defining a first time lag […], the first […] lookback parameters being components of the […] model. (Arunachalam para 32 teaches marketing-mix variables may include an ad-stock variable, an event variable, a lead variable and a lag variable for each channel. para 35 teaches generating the response model using linear regression, non-linear regression, or mixed models where dependent variable is sales or revenue and independent variables are marketing-mix variables.)
Arunachalam does not teach wherein said training comprises specifying, for each of the plurality of media channels, a first lookback parameter defining a first time lag and a second lookback parameter defining a second time lag, the first and second lookback parameters being components of the random forest model.
Fan does teach wherein said training comprises specifying, for each of the plurality of media channels, a first lookback parameter defining a first time lag and a second lookback parameter defining a second time lag, the first and second lookback parameters being components of the random forest model. (Fan para 14 teaches splitting time series data set into a training and test dataset. Para 39 teaches a Random Forest as a machine learning model. Para 47 teaches predicting outcomes using a random forest. Para 35 teaches lagged values with different lags.)
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the model (which may be linear regression, non-linear regression, or mixed models) and lagged parameter as taught by Arunachalam with the trained random forest and second lagged parameter as taught by Fan. It would be beneficial for the model used to forecast time series data as taught by Fan para 38.
Claims 2, 9 are rejected under 35 U.S.C. 103 as being unpatentable over Arunachalam (US 20130166351) in view of Fan (US 20210209486) in view of Dockery (20060129447)
CLAIM 2, 9
Arunachalam teaches wherein the recipients […] (Arunachalam para 23 teaches sales data captured over time and media channel variables delivered over time and the sales data over time)
Arunachalam does not teach wherein the recipients comprise health care providers in a plurality of defined specialties
Dockery does teach wherein the recipients comprise health care providers in a plurality of defined specialties (Dockery para 23 teaches sales effort being based on calling on a target market population of physicians who are likely to use or prescribe a pharmaceutical product. Para 24 teaches physician categories may be based on specialty.)
It would have been obvious to one or ordinary skill in the art, before the effective filing date of the claimed invention, to modify the recipients as taught by Arunachalam with the recipient comprising health providers in a plurality of defined specialties as taught by Dockery. It would be beneficial for to provide sales insight for pharmaceutical companies trying to target specific physicians to increase sales as taught by Dockery para 4-6.
Claims 3, 10, 16 are rejected under 35 U.S.C. 103 as being unpatentable over Arunachalam (US 20130166351) in view of Fan (US 20210209486) in view of Dockery (20060129447) in view of Dakic (US 20220270129)
CLAIM 3, 10, 16
Arunachalam teaches wherein said […] further comprises specifying, for each of the plurality of media channels, […] the […] model […], and the response curves are generated to be specific […]. (Arunachalam para 35 teaches generating the response model using linear regression, non-linear regression, or mixed models where dependent variable is sales or revenue and independent variables are marketing-mix variables. Para 32 teaches marketing-mix variables may include an ad-stock variable, an event variable, a lead variable and a lag variable. Fig 6, Fig. 9 show subset of timeseries data for media channels)
Arunachalam does not teach wherein said training further comprises specifying, for each of the plurality of media channels, a component of the random forest model […]
Fan does teach wherein said training further comprises specifying, for each of the plurality of media channels, a component of the random forest model […] (Fan para 14 teaches splitting time series data set into a training and test dataset. Para 39 teaches a Random Forest as a machine learning model)
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the model (which may be linear regression, non-linear regression, or mixed models) as taught by Arunachalam with the trained random forest as taught by Fan. It would be beneficial for the model used to forecast time series data as taught by Fan para 38.
Arunachalam does not teach:
wherein said training further comprises specifying, for each of the plurality of media channels, a component of the random forest model […] health care provider specialty
wherein the response curves are generated to be specific to health care provider specialty.
Dockery does teach
wherein said training further comprises specifying, for each of the plurality of media channels, a component of the random forest model […] health care provider specialty
wherein the response curves are generated to be specific to health care provider specialty. (Dockery para 23 teaches sales effort being based on calling on a target market population of physicians who are like to use or prescribe pharmaceutical product. Para 24 teaches physician categories may be based on specialty. Para 33 teaches generating response curves for defined target segments. )
It would have been obvious to one or ordinary skill in the art, before the effective filing date of the claimed invention, to modify the recipients as taught by Arunachalam with the recipient comprising health providers in a plurality of defined specialties as taught by Dockery. It would be beneficial for to provide sales insight for pharmaceutical companies trying to target specific physician specialties to increase sales as taught by Dockery para 4-6.
Arunachalam in view of Dockery does not teach wherein said training further comprises specifying, for each of the plurality of media channels, a component of the random forest model encoding health care provider specialty
Dakic does teach wherein said training further comprises specifying, for each of the plurality of media channels, a component of the random forest model encoding health care provider specialty (Dakic para 68, 82 teach a random forest model using data 130, 181, 233 teach data may include specialty or practice area data)
It would have been obvious to one or ordinary skill in the art, before the effective filing date of the claimed invention, to modify the random forest model as taught by Arunachalam with the encoding of provider specialty as a component of the random forest model as taught by Dakic. It would be beneficial to enable greater accuracy and optimization in reached the desired Health Care Providers (HCP) as taught by Dakic para 136.
Claims 6, 13, 19 are rejected under 35 U.S.C. 103 as being unpatentable over Arunachalam (US 20130166351) in view of Fan (US 20210209486) in view of Beharie (US 20230419345)
CLAIM 6, 13, 19
Arunachalam teaches wherein the plurality of media channels comprises […], and digital engagements; (Arunachalam Fig. 2 teaches google impression)
Arunachalam does not teach wherein the plurality of media channels comprises at least two of: emails, phone calls, and digital engagements.
Beharie does teach wherein the plurality of media channels comprises at least two of: emails, phone calls, and digital engagements. (Beharie para 219 teaches data may include data from a range of communication channels, including a telephone, email, live chat, marketing materials and social media materials.)
It would have been obvious to one or ordinary skill in the art, before the effective filing date of the claimed invention, to modify the media channels as taught by Arunachalam with the media channels including at least two of emails or phone calls as taught by Beharie. It would be beneficial to compile range of different communication channels to target and cater to audiences and drive sales growth as taught by Beharie para 4.
Prior Art Made of Record and Not Relied Upon
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
US 20180046926 Achin
[0333] In some embodiments, a predictive modeling system 100 includes a time series model that can predict the values of a target X at time t and optionally t+1, . . . , t+i, given observations of X at times before t and optionally observations of other predictor variables P at times before t. In some embodiments, the predictive modeling system 100 partitions past observations to train a supervised learning model, measure its performance, and improve accuracy. In some embodiments, the time series model provides useful time-related predictive features, for example, predicting previous values of the target at different lags. In some embodiments, the predictive modeling system 100 refreshes the time series model as time moves forward and new observations arrive, taking into account the amount of new information in such observations and the cost of refitting the model.
US 20240152775 ALAGAPPAN
[0027] Feature engineering is a process of merging customer digital intent data 108 with customer demand data 111 at weekly levels (e.g., 2 weeks, 4 weeks, etc.). To determine the appropriate relationship between customer digital intent data 108 and the customer demand data 111, the system creates multiple combinations of lagged datasets at weekly levels. The multiple combinations include multiple lagged variables from HVAs (e.g., four different variables from each lagged HVAs). In one example, for each individual HVA, 4 additional data points are created. Thus, assuming the existence of 30 HVA variables, 120 different variables may be generated. By generating additional variables, the system is better able to determine data-points that are indicative of the projected lag. The weekly lag can impact the model (e.g., a four week sum can impact a model more than a three week sum or a two week sum can impact a model more than a 4 week sum).
US 20240281831 Shukla
[0073] For example, in some instances, the promotional forecasting computing system 104 may determine (e.g., calculate) lagging data such as lagged average sales. For instance, the promotional forecasting computing system 104 may determine lagged average sales by shifting the sales by a week and/or determining an average of the lagged sales. For instance, a lag of one week may indicate for the promotional forecasting computing system 104 to shift the week 2 sales (e.g., 20 units) to week 3. Additionally, and/or alternatively, the promotional forecasting computing system 104 may determine lagged minimum or maximum saves, direct lag sales features, and/or other lagging data. The promotional forecasting computing system 104 may use the lagging data to train the promotional forecasting ML-AI models, which will be described in further detail below.
Brownlee, “Basic Feature Engineering With Time Series Data in Python”, September 15, 2019
Section 1 teaches lag features are values at prior time steps.
Statworx, “Time Series Forecasting With Random Forest”, 25. September 2019
Section 1 teaches using lag features as a part of feature engineering for a random forest model to be used on time series data.
Benjaminson, “Using ML models for time series forecasting”, 17 February, 2023
Art teaches time series forecasting in supervised learning including a lag method or sliding window method to use random forest to generate time series forecasts.
Tyralis, “Variable Selection in Time Series Forecasting Using Random Forests”, 4 October 2017
Section 1.2 teaches lagged variables in a random forest.
Response to Arguments Regarding U.S.C. 101 Rejection
Applicant argues pg. 9:
The Examiner's assertion that the claims are directed to an abstract idea because they purportedly cover a "method of organizing human activity" related to advertising, marketing, and managing interactions (and that such activities are implemented in the claims using generic computer components) is based on an overly broad interpretation that fails to properly consider the technological context and specifics of the claimed subject matter. As discussed in the Response filed August 4, 2025, which is hereby incorporated herein by reference, the claims are directed to a specific technological process for predicting channel responses using machine learning, which addresses longstanding technical challenges in channel specific predictive modeling based on time series data.
Examiner responds
Examiner disagrees. MPEP 2106. 04(a)(2)(II) states that a claimed invention is directed to certain methods of organizing human activity if the identified claim elements contain limitations that encompass fundamental economic principles or practices (including hedging, insurance, mitigating risk); commercial or legal interactions (including agreements in the form of contracts, legal obligations, advertising, marketing or sales activities or behaviors, and business relations); and managing personal behavior or relationships or interactions between people, (including social activities, teaching, and following rules or instructions). The Examiner submits that the identified claim elements represent a series of rules or instructions that a person or persons, with or without the aid of a computer, would follow to predicting responses to media. Furthermore, the Examiner submits that healthcare itself is inherently represents the organization of human activity. Applicant has not pointed to anything in the claims that fall outside of this characterization. Implementing an abstract idea using a generic computer or components thereof does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea.
Further, Applicant has not discussed what technical challenge is being referred to in “longstanding technical challenges in channel specific predictive modeling based on time series data.” No technological problem is discussed nor a technical solution.
Applicant argues pg. 9:
The Claims Do Not Recite a Mathematical Concept or Relationship. For reasons discussed in the Response filed August 4, 2025, it is deemed that the claims as a whole do not recite an abstract idea in the form of methods of organizing human activity or a mathematical concept or relationship. Claims 1, 8, and 15 are therefore considered to be patent eligible.
Examiner responds
Examiner disagrees. See at least Final Office action 8/25/2025 page 18. The independent claim further recites “training a random forest model”. When given its broadest reasonable interpretation in light of the disclosure, the training of a machine learning model by splitting time series data represents the creation of mathematical interrelationships between data. As such, the training of the machine learning model represents a mathematical concept that is interpreted to be part of the identified abstract idea, supra. The types of identified abstract ideas are considered together as a single abstract idea for analysis purposes.
Applicant argues pg. 10:
[…] Even if claim 1 were deemed to recite a judicial exception (e.g., a mathematical concept), the claim as a whole integrates any such exception into a practical application by reciting specific additional elements that improve the functioning of a computer-implemented predictive modeling system and the technical field of channel specific predictive modeling based on time series data.
As amended, claim 7 now even more clearly recites a specific technical training configuration, including:
Per-channel lookback parameters that define specific time lags for each channel
Generation of lagged features based on those lookback parameters, producing discrete, time-shifted variables that preserve historical values.
Integration of lagged features into the models feature space during training, with splitting based on channel variables in combination with corresponding lagged features
These features are not generic "apply an algorithm on a computer" instructions. Rather, they change how the model learns temporal relationships. This yields measurable improvements in model accuracy, scalability, and the ability to capture channel-specific, non-linear carryover effects --- improvements that conventional adstock--based preprocessing (as in Kisilevich) does not achieve.
Examiner responds
The Examiner respectfully disagrees. MPEP 2106.04(d)(1) states that a practical application may be present where the claimed invention improves the functioning of a computer. See also MPEP 2106.05(a)(I). The technological environment of Applicant’s claim is a general-purpose computer. Applicant has not identified nor can the Examiner locate any physical improvement to the functioning of the computer that results from the implementation of Applicant’s claim. There is no indication that the computer is made to run faster, more efficiently, or utilize less power. In fact, the computer may be caused to operate slower and less efficiently through the implementation of Applicant’s claimed invention; we do not know. Because there is no improvement to the function of the computer, a practical application is not present.
Applicant’s argued improvements of accuracy, scalability, and the ability to capture channel-specific, non-linear carryover effects are directed to model prediction and do not constitute an improvement to computer functionality, another technology or technical field. Applicant’s claim to improve the model using feature engineering features would be an improvement to the abstract idea of the mathematical relationships of training a random forest model and not computer functionality, another technology or technical field.
Applicant argues pg. 10:
The precedential decision in Ex parte Desardins., Appeal 2024-000567., Application
16/319,040 (PTAB Apps. Rev. Panel Sept. 26, 2025) is directly relevant. In Desjardins, the
Appeals Review Panel vacated a patent eligibility rejection under 35 U.S.C. §101 because the
claims were found to recite a method of training a machine learning model that improved the
model's operation --- specifically, by enabling it to learn new tasks while retaining prior task
performance. The Panel emphasized that the improvement was reflected in the claim language
itself: and that such a training improvement constituted "an improvement in the functioning of a
computer, or an improvement to other technology or technical field" under MPEP § 2106.0S(a).
Similarly, claim 1 is a method of generating a model in which the improvement - using
per-channel lookback parameters to generate lagged features for training - is recited in the claim
itself. This is a specific modeling strategy that changes the model's feature space and splitting
behavior during training, thereby improving its ability to model temporal dependencies and
channel interactions. Like the claims in Desjardins, this is an improvement to the operation of
the model itself and to the technical field in which it is applied, i.e., the technical field of
channel specific predictive modeling based on time series data.
Examiner responds
Examiner disagrees. MPEP 2106.04(d), subsection III “In Ex Parte Desjardins, Appeal No. 2024-000567 (PTAB September 26, 2025, Appeals Review Panel Decision) (precedential), the claimed invention was a method of training a machine learning model on a series of tasks. The Appeals Review Panel (ARP) overall credited benefits including reduced storage, reduced system complexity and streamlining,
and preservation of performance attributes associated with earlier tasks during subsequent
computational tasks as technological improvements that were disclosed in the patent
application specification. Specifically, the ARP upheld the Step 2A Prong One finding that
the claims recited an abstract idea (i.e., mathematical concept). In Step 2A Prong Two, the
ARP then determined that the specification identified improvements as to how the machine
learning model itself operates, including training a machine learning model to learn new
tasks while protecting knowledge about previous tasks to overcome the problem of
“catastrophic forgetting” encountered in continual learning systems. Importantly, the ARP
evaluated the claims as a whole in discerning at least the limitation “adjust the first values of
the plurality of parameters to optimize performance of the machine learning model on the
second machine learning task while protecting performance of the machine learning model
on the first machine learning task” reflected the improvement disclosed in the specification.
Accordingly, the claims as a whole integrated what would otherwise be a judicial exception
instead into a practical application at Step 2A Prong Two, and therefore the claims were
deemed to be outside any specific, enumerated judicial exception (Step 2A: NO).”
Applicant’s claimed invention changes the data given to the model, not how the machine learning model itself operates. Applicant’s limitations of per-channel lookback parameters to generate lagged features for training is feature engineering and data preparation because it only changes the data given to the model and does not change how the model itself operates.
Examiner disagrees that the splitting behavior during training is changed. Applicant claims “splitting the time series data into subsets based on media channel of the plurality of media channels and corresponding lagged features.” A random forest splits data based on feature data and therefore would split data into subsets based on channel and corresponding lagged features therefore the random forest splitting behavior is not changed.
Applicant argues pg. 13:
Thus, it is apparent that amended claim 1 recites a technical solution to a technical problem, as well as an improvement in technical approach, thereby providing a practical application of any supposed judicial exception. Specifically, amended claim 1 recites details of how the solution to the problem is accomplished, as opposed to only the idea of a solution or outcome, and describes how the result is accomplished and the mechanism for accomplishing the result, as opposed to merely attempting to cover any solution to the identified problem. Thus, amended claim 1 recites "a particular solution to a problem or a particular way to achieve a desired outcome" and therefore integrates any supposed judicial exception into a practical application, as opposed to reciting "mere instructions to apply an exception." M.P.E.P.
Examiner responds
The Examiner respectfully disagrees. MPEP 2106.04(d)(1) and MPEP 2106.05(a) indicates that a practical application may be present where the claimed invention provides a technical solution to a technical problem. See, e.g., DDR Holdings, LLC. v. Hotels.com, L.P., 773 F.3d 1245, 1259 (Fed. Cir. 2014) (finding that claiming a website that retained the “look and feel” of a host webpage provided a technological solution to the problem of retention of website visitors by utilizing a website descriptor that emulated the “look and feel” of the host webpage, where the problem arose out of the internet and was thus a technical problem). Here, the Examiner cannot find, nor has the Applicant identified, any technological problem that was caused by the technological environment to which the claims are confined.
Applicant argues pg. 13:
Notwithstanding the above, if a claim is deemed to be directed to a judicial exception, then analysis must continue with Step 2B to determine whether the claim recites additional elements that amount to "significantly more" than the alleged judicial exception(s ). This is the second part of the Alice/Mayo test, which is often referred to as a search for an inventive concept. M.P.E.P. § 2106.05(1) (citing Alice Corp. Pty. Ltd v. CLS Bank Int'l, 573 U.S. 208, 217, 110 USPQ2d 1976, 1981 (2014), emphasis added). An "inventive concept" is provided by an element, or combination of elements, recited in a claim - in addition to a judicial exception - sufficient to ensure that the claim as a whole amounts to significantly more than the judicial exception itself Alice Corp. Pty. Ltd v. CLS Bank Int'l, 573 U.S. at 217-218, 110 USPQ2 d at 1981 (citing Mayo Collaborative Servs. v. Prometheus Labs, 566 U.S. at 72-73, 101 USPQ2d at 1966). As discussed above, the claims resolve specific technical challenges rooted in channel-specific modeling based on time series data by introducing a novel combination of features within a machine learning framework. This includes the claimed use of a lookback parameter and the model's capacity for learning channel-specific response contributions. These elements materially improve how predictive models handle temporal carryover effects, nonlinear interdependencies, and segmentation of data across media channels.
Examiner responds
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional element of using a processor and memory to perform the noted steps amounts to no more than mere instructions to apply the exception using a generic computer component. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept (“significantly more”).
As discussed above with respect to integration of the abstract idea into a practical application, the additional element of using the trained random forest model to generate response curves was found to represent mere instructions to implement the abstract idea on a generic computer and confine the use of the abstract idea (i.e., the trained model) to a particular technological environment or field of use (random forest). This has been re-evaluated under the “significantly more” analysis and determined to be insufficient to provide significantly more. MPEP 2106.05(I) indicates that mere instructions to implement the abstract idea on a generic computer and/or confining the use of the abstract idea to a particular technological environment or field of use cannot provide significantly more. Accordingly, even in combination, this additional element does not provide significantly more. As such the claim is not patent eligible.
Applicant argues pg. 13:
The use of the lookback parameter in training the random forest model represents a meaningful advancement in data feature engineering within a machine learning context. It creates lagged variables that incorporate temporal relationships directly as features of the random forest model, enabling it to learn carryover effects without requiring manually configured or optimized adstock decay parameters. This is dramatically different from conventional approaches, such as Kisilevich' s framework, which relies heavily on manual preprocessing to approximate temporal dependencies. The use of the lookback parameter in training the random forest model reduces complexity, eliminates potential errors stemming from manual configuration, and enables more accurate temporal modeling that is scalable for real world datasets. Accordingly, it is deemed that claims 1, 8, and 15 (and the claims depending therefrom) recite additional elements that amount to significantly more that the alleged judicial exception(s) and, therefore, are patent eligible.
Examiner responds
The Examiner disagrees. MPEP 2106.05(d) states: “Another consideration when determining whether a claim recites significantly more than a judicial exception is whether the additional element(s) are well-understood, routine, conventional activities previously known to the industry (emphasis added).” Further, MPEP 2106.05(I) states: “As made clear by the courts, the novelty of any element or steps in a process, or even of the process itself, is of no relevance in determining whether the subject matter of a claim falls within the § 101 categories of possibly patentable subject matter (internal quotations omitted, emphasis original).” As such, it is only the additional elements identified by the Examiner to not be part of the abstract idea that are analyzed to determine whether they represent well-understood, routine, conventional activities in the field of the invention.
In that regard, MPEP 2106.05(d)(I) indicates that in determining whether the additional elements represent are well-understood, routine, conventional activities, the Examiner should consider whether the additional elements (1) provide an improvement to the technological environment to which the claim is confined, (2) whether the additional elements are mere instructions to apply the judicial exception, or (3) whether the additional elements represent insignificant extra-solution activity. The additional elements of the claims do not provide significantly more based on this inquiry.
Taking these in turn, whether the additional elements of the claim provide an improvement was analyzed/addressed in the 2A2 analysis as either no improvement was present. The technological environment to which the claims are confined (a general-purpose computer performing generic computer functions is recited at a high level of generality and has been found by the courts to be insufficient to provide a practical application (see MPEP 2106.05(d)(II); Alice Corp.). None of the additional elements of the claim were found to represent extra-solution activity and thus no well-understood, routine, conventional analysis is required.
Response to Arguments Regarding U.S.C. 102/103 Rejection
Applicant argues pg. 15
Independent claim 1 has been amended to recite that each lookback parameter is specific to a corresponding one of the plurality of media channels and that each lagged feature is generated by shifting the corresponding media channel variable by the time lag specified by its respective lookback parameter. These amendments make explicit that different media channels can have different lookback parameters and therefore different lagged features and that the lagged features are generated based on the per channel lookback parameters. Amended claim 1 recites, in part: […] By virtue of these claimed features, each channel has its own specifically associated lookback parameter(s)and that the lagged features are generated by shifting the actual channel variable values by the channel-specific lag.
Kisilevich uses manually configured adstock transformations to model temporal carryover effects. These transformations require geometric decay functions applied during preprocessing, followed by optimization of decay rates for each media channel. Kisilevich relies on external preprocessing for temporal dynamics and does not disclose or suggest the direct generation of lagged features or their integration into the training process of a random forest model. Kisilevich neither discloses nor suggests the specific training process of the random forest model with the integrated use of channel-specific lookback parameters to define lagged features.
The Examiner has analogized Kisilevich' s use of cross-validation and training set size to the claimed lookback parameters. However, Kisilevich' s training set size merely determines the amount of data used for validation; it does not result in the generation of lagged features or the integration of time-lagged responses as part of the model's feature space.
In contrast to Kisilevich' s approach, the claimed lookback parameters result in lagged channel values which are incorporated into the feature space of the random forest training process. This allows the model to learn temporal carryover effects directly, without needing manual decay approximations. The claimed approach eliminates the manual configuration and computational inefficiencies associated with Kisilevich's adstock-based preprocessing. Amended claim 1 further recites "training a random forest model using said one or more lagged features in combination with the dependent variable, the training comprising splitting the time series data into subsets based on media channel ... and corresponding lagged features." Kisilevich' s random forest training, on the other hand, is performed on adstock—transformed values alone. Splitting based on adstock values is significantly different than splitting based on dependent variables in combination with discrete lagged feature variables tied to specific channels and lags, in the manner claimed.
Examiner responds
Examiner has applied new art in light of Applicant’s amendment to explicitly cite a reference with a lag variable for each media channel in a model that predicts revenue. Examiner uses Fan to teach explicitly a model may be a random forest model that explicitly uses lag variables for time series data.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ANDREW KYLE TAPIA whose telephone number is (703)756-1662. The examiner can normally be reached 830 - 530.
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, Mamon Obeid can be reached at (571) 270-1813. 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.
/A.K.T./Examiner, Art Unit 3687
/MAMON OBEID/Supervisory Patent Examiner, Art Unit 3687