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 .
Acknowledgments
This communication is in response to Amendment filed on 12/29/2025.
Claims 1-2, 7-8, 11, 13-15, 17 are amended.
Claims 6, 12, 18, 22 are cancelled.
Claims 24 is new.
Claims 1-5, 7-11, 13-17, 19-21, 23-24 are currently pending and have been examined.
Claims 1-5, 7-11, 13-17, 19-21, 23-24 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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 12/29/2025 has been entered.
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 1-5, 7-11, 13-17, 19-21, 23-24 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, 7, 13 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 system, method, and non-transitory computer-readable storage media for determining the probability of an entity being associated with a prediction feature.
The limitations of receiving, […], user input comprising a request that includes prediction criteria data comprising (i) outer cohort definition data that defines a target population within a population, (ii) inner cohort definition data that identifies a prediction feature of the target population, and (iii) a confidence threshold that controls the number of one or more entities output […]; in response to the request, generating, […] for a permutation of a plurality of potential permutations within the population, the permutation being associated with the prediction feature for the target population by: determining, using a knowledge graph data object and prior to […], one or more inner cohort features correlated to the inner cohort definition data based at least in part on a subpopulation, within the target population, that is associated with the prediction feature, wherein (i) the knowledge graph data object comprises (a) a node associated with a set of features and an outer cohort entity, and (b) an edge defining a relationship between a first feature and a second feature of the set of features, and (ii) determining the one or more inner cohort features is, based at least in part on one or more different sets of edges associated with a set of nodes of the knowledge graph data object, the one or more inner cohort features being indicated as related to the prediction feature, generating, without user input and prior to […], a training dataset specific to the request based at least in part on the one or more inner cohort features and the target population by determining a ground truth for an entity within the target population based at least in part on a correlation between the one or more inner cohort features and a set of input features of the entity, and generating the […] based at least in part on the training dataset, wherein training is limited to the training dataset and the training dataset is excluded from being stored […]; and outputting, […] and based on the confidence threshold, one or more entities within the target population in response to the request, as drafted, is a process that, under the broadest reasonable interpretation, covers performance of the limitation in the mind but for recitation of generic computer components. That is, other than reciting a processor (Claim 1, 7, 13) or a memory (Claim 7) or a non-transitory computer readable storage media (Claim 13) nothing in the claim precludes the step from practically being performed in the mind. For example, but for the computer, memory and storage media this claim encompasses a person observing patient data, thinking about relationships between data and judging that an entity is associated with a prediction feature in the manner described in the identified abstract idea, supra. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then it falls within the “Mental Processes” grouping of abstract ideas. Accordingly, the claim recites an abstract idea.
The limitation of […] wherein (i) the knowledge graph data object comprises (a) a first node associated with a set of features and an outer cohort entity, and (b) a first edge defining a relationship between a first feature and a second feature of the set of features, and (ii) determining the one or more inner cohort features comprises determining, based at least in part on one or more different sets of edges associated with a set of nodes of the knowledge graph data object, the one or more inner cohort features indicated as being related to the prediction feature […] as drafted, represents a mathematical concept. The abstract idea is considered to be one abstract idea for analysis purposes.
Step 2A2
This judicial exception is not integrated into a practical application. In particular, the claim recites the additional element of (claim 1, 7, 13) a processor and a memory (Claim 7) and a non-transitory computer readable storage media (Claim 13) that implements the identified abstract idea. The processor, memory and computer readable storage media 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 an interactive user interface. The interactive user interface merely generally links the abstract idea to a particular technological environment or field of use. MPEP 2106.04(d)(I) indicates that generally linking an abstract idea to a particular technological environment or field of use cannot provide a practical application. Utilization of the interactive user equates to saying “apply it.” MPEP 2106.04(d)(I) indicates that merely saying “apply it” or equivalent to the abstract idea cannot provide a practical application. Accordingly, even in combination, this additional element does not integrate the abstract idea into a practical application.
The claim further recites the additional element of using the trained machine learning model to identify a probability that an entity is associated with a prediction feature. 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.
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, memory and computer readable storage media 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”).
Also, as discussed above with respect to integration of the abstract idea into a practical application, the additional element of an interactive user interface was determined to generally link the abstract idea to a particular technological environment or field of use. This has been re-evaluated under the “significantly more” analysis and has also been found insufficient to provide significantly more. MPEP 2106.05(A) indicates that generally linking an 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.
As discussed above with respect to integration of the abstract idea into a practical application, the additional element of using the trained machine learning model to identify a probability that an entity is associated with a prediction feature was found to represent mere instructions to implement the abstract idea on a generic computer. 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 cannot provide significantly more.
Dependent Claims
Claims 2-5, 8-11, and 14-17, 19-21, 23-24 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, 8, 14 merely describe the dataset. Claims 4, 10, 16 merely describe the features associated with the dataset. Claims 5, 11, 15 merely describe the correlation values. Claims 6, 12, 18 merely describe the training data. Claim 24 merely describes a predicted diagnosis.
Claims 3, 9, 15 also includes the additional element of “a machine learning model” which is analyzed the same as the “a machine learning model” and does not provide a practical application or significantly more for the same reasons. Claims 3, 9, 15 merely describes types of machine learning model. Claim 20 merely describes the model training. Claim 21 merely describes a risk score. Claim 22 merely describes discarding the model. Claim 23 merely describes the interface and 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 text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
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, 6-10, 12-16, 18-19 are rejected under 35 U.S.C. 103 as being unpatentable over Birnbaum (US 20190258950) in view of Zhang (US 20190057316) in view of Guetz (US 20140278472)
Claims 1, 7, 13 :
Birnbaum teaches A computer-implemented method comprising: (Birnbaum para 7 teaches method and non-transitory computer readable storage media may store program instructions, which are executed by at least one processing device and perform any of the methods described herein. )
receiving, by one or more processors […] a request that includes prediction criteria data that comprising (i) outer cohort definition data that defines a target population within a population, (Birnbaum para 19, 22, Fig 1. Para 19 teaches a two-step cohort selection filter. Fig 1 shows sorting patients into a first group of “Maybe in cohort” and then sorting into a subgroup “in group”. “Maybe in cohort” is treated synonymously with the outer cohort and group is treated synonymously with population. “In cohort” is treated synonymously as the inner cohort. Para 22 teaches that the first sorting into “maybe in cohort” is based on rules that identify characteristics of patients (rules that identify patients as “maybe in cohort” are treated synonymously as outer cohort definition data. Para 18 teaches analyzing patient data and select cohorts. )
(ii) inner cohort definition data that identifies a prediction feature for the target population, and (Birnbaum para 19, 22, 26, 35 Fig 1. Para 19 teaches a two-step cohort selection filter. Fig 1 shows sorting patients into a first group of “Maybe in cohort” and then sorting into a subgroup “in group”. “Maybe in cohort” is treated synonymously with the outer cohort. “In cohort” is treated synonymously as the inner cohort. Para 26 teaches labels indicate a patient as suitable for a cohort (Labels that indicate the patient as suitable for a cohort are treated as synonymous with inner cohort definition data))
(iii) a confidence threshold that controls a number of one or more entities output by a trained model; (Birnbaum para 52 teaches a confidence level or score and comparison to a threshold. Para 52 teaches using the threshold to determine if a candidate may be for the cohort and therefore how if it is output by the model)
In response to the request, generating, by the one or more processors, the trained machine learning model for a permutation of a plurality of potential permutations within the population, the permutation being associated with the prediction feature of the target population by: (Birnbaum Para 29 teaches using a model to take patient data and determine a score (i.e., risk score) that indicates a likelihood that a patient is a viable candidate for a cohort (i.e., population). Para 28 teaches a machine learning neural network model relating feature vector objects to group probability labels. Examiner notes a permutation of a plurality of potential permutations is interpreted using the broadest reasonable interpretation to include structured data as taught in para 36 because structured data is a specific permutation of a plurality of potential permutations.)
determining, using a knowledge graph data object and prior to training the trained machine learning model, one or more inner cohort features correlated to the inner cohort definition data based at least in part on a subpopulation, within the target population, that is associated with the prediction feature, (Birnbaum para 26, 27, 28, 31. Para 26 teaches labels are used to indicate a patient is a candidate for a cohort (labels are treated as being synonymous with inner cohort definition). Para 27 then teaches extracting features from the labeled records. Para 28 teaches a machine learning neural network model relating feature vector objects to group probability labels (Examiner notes that the correlations/edges between the knowledge graph data objects are determined via a neural network consistent with the instant specification in ¶ 72, 73, 81. Para 27, 28, 32, 33 teaches creating training data based on extracted features based on correlations. )
wherein (i) the knowledge graph data object […] (ii) determining the one or more inner cohort features is, based at least in part on one or more different sets of edges associated with a set of nodes of the knowledge graph data object, the one or more inner cohort features indicated as being related to the prediction feature, (Birnbaum para 26, 27, 28, 31. Para 31 teaches extracting features such as key words or key phrases from labeled records (i.e., inner cohort) and scoring those features for a level of relevance to inclusion in a cohort (a cohort is treated as synonymous to a prediction feature). Para 37 further teaches determining top features correlated with inclusion or exclusion from the cohort. Para 28 teaches a neural network relating feature vectors to probable labels where labels are the cohort (Examiner notes that the structure of a neural network is related nodes and edges and outputs of a neural network are determined based on different edge associated with nodes of the network))
generating, without the user input and prior to training the trained machine learning model, a training dataset specific to the request based at least in part on the one or more inner cohort features and the target population by determining a ground truth for the entity within the target population based at least in part on a correlation between the one or more inner cohort features and the set of input features of the entity, and (Birnbaum para 27, 28, 32, 33 teaches creating training data based on extracted features. Para 27 teaches extracting features from labeled records to determine correlations. Para 28 teaches generating models based on determined correlations. Para 52 teaches supervised learning based on a set of data labels and generating results consistent with that set of labels (i.e., ground truth). Para 55 teaches the training dataset may be a labeled data set which a desired outcome is known such as a reference standard to create the model understood as not requiring user input because the data set is a reference standard and not created by the user. )
generating the trained machine learning model by training the trained machine learning model based at least in part on the training dataset; (Birnbaum para 32 teaches training the model on the training data)
wherein training is limited to the training dataset and the training dataset is excluded from being stored in memory and (Birnbaum para 53 teaches the model may be logistic regression which Examiner notes uses data to train, but does not have to store training data because only the weights are stored from training and then used for predictions and not the actual training data used to train the model. )
outputting, by the one or more processors, using the trained machine learning model and based on the confidence threshold, one or more entities within the target population in response to the request. (Birnbaum Para 29 teaches using models to take patient data and determine a score that indicates a likelihood that a patient is a viable candidate for a cohort (scores are treated as synonymous to risk score and cohort is treated as synonymous with the entity))
Birnbaum does not explicitly teach wherein (i) the knowledge graph data object comprises (a) a first node associated with a set of features and an outer cohort entity, and (b) a first edge defining a relationship between a first feature and a second feature of the set of features,
Zhang does teach wherein (i) the knowledge graph data object comprises (a) a first node associated with a set of features and an outer cohort entity, and (b) a first edge defining a relationship between a first feature and a second feature of the set of features, (Zhang para 123 explicitly teaches a knowledge graph comprising a node that refers to a category (a category is treated as synonymous with an outer cohort entity) and is associated with features where edges define relationships between the category and other features)
It would have been prima facie obvious to one of ordinary skill in the art at the time the invention was made to combine the noted features of Birnbaum with teaching of Zhang since the combination of the two references is merely combining prior art elements according to known methods to yield predictable results (KSR rational A); see MPEP 2143(I)(A)). It can be seen that each element claimed is present in either Birnbaum or Zhang. The knowledge graph comprising a first node with a set of features and an outer cohort entity and an edge defining a relationship between a first feature and second feature of the set of features does not change or affect the normal predicting of an entity being associated with a prediction feature. Predicting of an entity being associated with a prediction feature would be performed the same way even with the addition of the knowledge graph comprising a first node with a set of features and an outer cohort entity and an edge defining a relationship between a first feature and second feature of the set of features. Since the functionalities of the elements in Birnbaum and Zhang do not interfere with each other, the results of the combination would be predictable.
Birnbaum in view of Zhang does not teach receiving, by one or more processors and via a user interface, a request that includes prediction criteria data comprising
Guetz does teach
receiving, by one or more processors and via a user interface, a request that includes prediction criteria data comprising (Guetz para 28 teaches a requestor computer configured to receive from a user a prediction request. Para 36 teaches an interface handling unit configured to receive a prediction request )
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 system as taught by Birnbaum in view of Zhang with the user interface and user input comprising a request comprising prediction criteria as taught by Guetz. It would be beneficial for a user interface to take input for a prediction request as taught by Guetz because interactivity with a prediction system is important to a user for repeatability, convenience, and user friendly interaction to determine how small changes in population characteristics may potentially impact health care outcome and risk factors.
Claims 2, 8, 14:
Birnbaum teaches wherein the knowledge graph data object comprises a graph-structured model that stores interlinked descriptions of entities (Birnbaum para 3, 45. 49. Para 3 teaches clinic notes, radiology reports, pathology reports, doctor or nurse observations, structured and unstructured data, and any other type of information that may be included in a patient's medical record. Para 28 teaches a neural network relating feature vectors to probable labels where labels are the cohort (Examiner notes that the structure of a neural network is related nodes and edges and outputs of a neural network are determined based on different edge associated with nodes of the network))
Claims 4, 10, 16:
Birnbaum teaches wherein the knowledge graph data object includes one or more of: a diagnosis code, a medication code, demographics data, medical history data or a procedure code. (Birnbaum para 45, 46. Birnbaum teaches medical records include structured and unstructured data such tumor progression and response, medication orders (i.e., medication code), age, race, gender (i.e., demographics data) and lines of therapy. Para 28 teaches a neural network relating feature vectors to probable labels where labels are the cohort (Examiner notes that the structure of a neural network is related nodes and edges and outputs of a neural network are determined based on different edge associated with nodes of the network))
Birnbaum may not explicitly teach diagnosis or procedure code. However, the limitation claims information that do not result in a manipulative difference between the information of the prior art and the functionally of the claimed method. The function taught by the prior art would be performed the same regardless of whether the information was substituted with nothing. Because Birnbaum teaches that data containing information is stored, substituting the information of the claimed invention for the information of the prior art would be an obvious substitution of one known element for another, producing predictable results. Therefore would have been prima facie obvious to one of ordinary skill in the art at the time of filing to have substituted the information applied to the stored data of the prior art with any other information because the results would have been predictable.
Claim 19:
Birnbaum teaches wherein the prediction feature comprises one of: a diagnosis code, a procedure code, a treatment code, a health outcome, a medication code, an age, or a gender. (Birnbaum para 49. Birnbaum teaches a selected cohort may be a group of individuals that share demographic or clinical characteristics (demographic or clinical characteristics are treated as synonymous with prediction feature. Clinical characteristics are treated as synonymous to health outcome. Demographic characteristics are treated as synonymous to age and gender))
Birnbaum may not explicitly teach diagnosis code, procedure code, treatment code, or medication code. However, the limitation claims information that do not result in a manipulative difference between the information of the prior art and the functionally of the claimed method. The function taught by the prior art would be performed the same regardless of whether the information was substituted with nothing. Because Birnbaum teaches that data containing information is stored, substituting the information of the claimed invention for the information of the prior art would be an obvious substitution of one known element for another, producing predictable results. Therefore would have been prima facie obvious to one of ordinary skill in the art at the time of filing to have substituted the information applied to the stored data of the prior art with any other information because the results would have been predictable.
Claim 20:
Birnbaum teaches wherein the trained machine learning model is trained (i) without having to store the training data corresponding to the permutation and (ii) without user input outside of the prediction criteria data. (Birnbaum para 53 teaches the model may be logistic regression which uses data to train, but does not have to store training data because only the weights are stored from training and then used for predictions and not the actual training data used to train the model. Para 32 teaches training which does not use user input. )
Claim 21:
Birnbaum teaches generating, using the trained machine learning model, a risk score identifying a probability that the entity within the target population is associated with the prediction feature based on the set of input features for the entity. (Birnbaum Para 29 teaches using models to take patient data and determine a score that indicates a likelihood that a patient is a viable candidate for a cohort (scores are treated as synonymous to risk score and cohort is treated as synonymous with the entity))
Claim 23:
Birnbaum teaches wherein the user interface is rendered by a user device and the trained machine learning model is provided for storage on the user device. (Birnbaum para 43 teaches a user interface and memory medium. Para 28 teaches generation of models. Examiner notes “for storage on the user device” is intended use. Para 42 teaches a system including server systems, databases and computing systems and a network may facilitate cloud sharing, storage, and computing. Examiner notes a model is data which is stored.)
Claim 24:
Birnbaum teaches generating a predicted medical diagnosis for the one or more entities within the target population in response to the request. (Birnbaum para 79 teaches the model may be configured to estimate the probability of metastatic cancer based on the feature vectors. )
Claims 3, 9, 15 are rejected under 35 U.S.C. 103 as being unpatentable over Birnbaum (US 20190258950) in view of Zhang (US 20190057316) in view of Guetz (US 20140278472) in view of Shan (US 20180060728)
Claims 3, 9, 15:
Birnbaum teaches wherein the trained machine learning model […]. (Birnbaum para 22, 53. Para 22 teaches neural network and para 52 teaches logistic regression.
Birnbaum does not teach wherein the trained machine learning model comprises a gradient boosted decision tree.
Shan does teach wherein the trained machine learning model comprises a gradient boosted decision tree. (Shan para 33 teaches a gradient boosted decision tree)
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 model as taught by Birnbaum in view of Zhang in view of Guetz with the gradient boosted decision tree as taught by Shan. It would be beneficial for the model to be a gradient boosted decision tree because of improved accuracy and runtime speed as taught by Shan para 33 and inexpensive runtime of forest based models as taught by Shan para 2.
Claims 5, 11, 17 are rejected under 35 U.S.C. 103 as being unpatentable over Birnbaum (US 20190258950) in view of Zhang (US 20190057316) in view of Guetz (US 20140278472) in view of Kenedy (US 20100063835)
Claim 5, 11, 17:
Birnbaum teaches determining a correlation value between the set of input features and the one or more inner cohort features, (Birnbaum Para 27 teaches correlations being determined between sets of features for patients labeled a first way and second way. para 28 teaches model based on determined correlations.)
wherein the correlation values comprise […] of the prediction feature being present for a given entity in the subpopulation compared to another given entity. (Birnbaum para 23, 27. Birnbaum teach correlations being determined for patients in a subgroup “in a cohort” and “not in cohort”)
Birnbaum does not teach wherein the correlation values comprise a ratio of odds of the prediction feature being present for a given entity in the subpopulation compared to another given entity
Kenedy does teach wherein the correlation values comprise a ratio of odds of the prediction feature being present for a given entity in the subpopulation compared to another given entity. (Kenedy para 41 teaches a ratio of odds indicates the strength of correlations as well as the statistical significance of correlations)
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 correlation values as taught by Birnbaum with the odds ratio as taught by Kenedy. It would be beneficial to have a statistical measure to indicate the strength of correlation as taught by Kenedy, "[0041] Various statistical measures can be used to provide results which indicate the strength of the correlations as well as the statistical significance (confidence) of the correlations. Examples of statistical measures that provide values indicating the strength of correlations include probability, likelihood (odds), likelihood ratio (odds ratio), absolute risk and relative risk."
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 20200320400 David
[0036] Because embodiments of the invention can mimic and modify a target model without accessing the original training dataset, only the target model itself, but not the original training dataset, needs be stored. Thus, after a target model is generated, the original training dataset may be deleted, resulting in a significant reduction in memory space (e.g., a reduction of gigabytes, terabytes, or in applications such as video analysis, petabytes, used to store a training dataset).
US 20220019850 TAKESHIMA
[0030] By implementing the discard function 114, the processing circuitry 11 disposes of the derived model. More specifically, the processing circuitry 11, after using the derived model by the inference function 113, saves a dump of the derived model in the storage 15 and discards (erases) the derived model. The discard function 114 is one example of a discard unit.
Hardesty, Data diversity: Preserving variety in subsets of unmanageably large data sets should aid machine learning. December 16, 2016
Para 1 teaches “When data sets get too big, sometimes the only way to do anything useful with them is to extract much smaller subsets and analyze those instead. Those subsets have to preserve certain properties of the full sets, however, and one property that’s useful in a wide range of applications is diversity. If, for instance, you’re using your data to train a machine-learning system, you want to make sure that the subset you select represents the full range of cases that the system will have to confront.”
Response to Arguments Regarding U.S.C. 103
Applicant argues pg. 10-11:
The cited references fail to disclose, teach, or suggest at least […] as recited by claims 1, 7, and 13.
First, the Office Action rejects the confidence threshold recited by the claims as suggested by Birnbaum's description of a "confidence level or score and comparison to a threshold." Office Action, p. 9 (citing Birnbaum ,i 52). Birnbaum describes a predetermined threshold level that may be adjustable based on desired levels of efficiency and performance. Birnbaum , [0080]. To do so, the threshold is adjusted using one or more loss functions. Id Birnbaum does not contemplate using a threshold to control a number of entities output by a trained machine learned model. Therefore, Birnbaum does not teach or suggest "receiving ... a confidence threshold that controls a number of one or more entities output by a trained machine learned model" as recited by the claims.
Examiner responds:
Examiner disagrees. Birnbaum para 52 teaches a confidence level or score and comparison to a threshold. Para 52 teaches using the threshold to determine if a candidate may be for the cohort. Therefore the threshold does control the number of entities output by the trained machine learned model as the member is either part of the cohort or not and the confidence threshold controls how many would be determined to be part of the cohort or not where higher thresholds limit how many are predicted as part of the cohort. Examiner submits this is the exact same use by Applicant as shown in specification para 87, “For example, a user can specify a confidence threshold in order to control the number of individuals the predictive machine learning model returns that have a high risk of developing specified conditions (e.g., based at least in part on inner cohort definition data)”.
Applicant argues pg. 11
Second, the Office Action states that a "reference standard to create a model [is] understood as not requiring user input because the data set is a reference standard not created by the user." Office Action, p. 11 (citing Birnbaum ,i [0055]). Birnbaum explicitly states that "training of the model may involve the use of a labeled data set for which a desired outcome is already known. Such data may be referred to as 'reference standard' and may be generated, for example, through an abstraction process in which all of the individual of a particular population are screened relative to one or more cohorts, and each individual is assigned to an appropriate cohort." Birnbaum ,I [0055]. Thus, the "reference standard" described by Birnbaum, simply refers to the manually labeled data set for which a desired output is already known. For example, Birnbaum explains that the labels of the labeled data set (i.e., the "reference standard") are "added by a medical professional." Id. ,i [0026]. Birnbaum does teach or suggest "generating, without user input and prior to training the trained machine learning model, a training dataset specific to the request based at least in part on the one or more inner cohort features and the target population" as recited by the claims.
Further, Birnbaum explains that, once labeled, the labeled records are input to a "training or abstraction process to determine correlations between one set of features shared amongst patients labeled a first way and a second set of features shared amongst patients labeled a second way." Id ,i [0027]. These correlations are embodied by trained models. See id ,i [0028]. Thus, Birnbaum describes determining correlations between one or more features within a training dataset after or during the training of a model. Birnbaum does not teach or suggest "generating, ... prior to training the trained machine learning model, a training dataset ... based at least in part on a correlation between the one or more inner cohort features and a set of input features of the entity" as recited by the claims
Examiner responds:
Birnbaum para 55 teaches “[0055] Training of the model may involve the use of a labeled data set for which a desired outcome is already known. Such data may be referred to as “reference standard” and may be generated, for example, through an abstraction process in which all of the individuals of a particular population are screened relative to one or more cohorts, and each individual is assigned to an appropriate cohort.” Thus the reference standard is generated and not merely referring to the manually labeled data set for which a desired output is known. Birnbaum Para 26 teaches “As depicted in FIG. 2A, framework 200 may accept, as input, labeled records 210. For example, records 210 may include data associated with a plurality of patients such that each patient is associated with one or more medical records and is associated with a label. In such an example, the label may have been added by a medical professional.” Thus, Birnbaum para 26 teaches labeled records may have been added by a medical professional, but does not require labels to be added by medical professional and data may be associated with a label another way not requiring human input. As such human input is not required for the dataset is created prior to training and without user input. Further, Para 27 teaches “[0027] As further depicted in FIG. 2A, framework 200 may input labeled records 210 to a training or abstraction process 220. Process 220 may extract one or more features (e.g., feature vectors or the like) from labeled records 210 to determine correlations between one set of features shared amongst patients labeled a first way and a second set of features shared amongst patients labeled a second way.” Para 28 teaches “Process 220 may therefore generate one or more models 230 based on the determined correlations”. Therefore the training dataset is created prior to training without user input and based on correlations.
Applicant argues pg. 11
In fact, the Office Action's own interpretation of Birnbaum supports this conclusion. The Office Action states that "models are generated through training therefore training data is based on the determined correlations." Office Action, p. 11 (citing Birnbaum ,i [0055]). This statement acknowledges that the determined correlations are, in fact, generated during training - not before training. Therefore, by the Office Action's own admission, Birnbaum does not teach or suggest "generating, ... prior to training the trained machine learning model, a training dataset ... based at least in part on a correlation between the one or more inner cohort features and a set of input features of the entity" as recited by the claims.
Examiner responds:
Para 27 teaches “[0027] As further depicted in FIG. 2A, framework 200 may input labeled records 210 to a training or abstraction process 220. Process 220 may extract one or more features (e.g., feature vectors or the like) from labeled records 210 to determine correlations between one set of features shared amongst patients labeled a first way and a second set of features shared amongst patients labeled a second way.” Para 28 teaches “Process 220 may therefore generate one or more models 230 based on the determined correlations. Examiner submits therefore this occurs before training and the model being trained on training data is therefore training based on correlations.
Response to Arguments Regarding U.S.C. 101
Applicant argues pg. 12-13:
Specifically, in response to Applicant's previous arguments, the Office Action states that "the training of a machine learning model 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." Office Action, p. 25. The assessment is expressly rejected in the precedential decision of Ex Parte Desjardines, where the Appeals Review Panel acknowledged that "improvements in training the machine learning model itself' were sufficient integrate an abstract idea into a practical application. Ex Parte Desjardines, p. 8. Just like the claims in Ex Parte Desjardines, the present claims recite an improved training technique for generating a machine learning model in a manner that used less of their storage capacity. Compare Specification ,-i,i [0018] & [0020] to Ex Parte Desjardines, p. 9. Thus, like the claims in Ex Parte Desjardines, the present claims are direct to patent eligible subject matter under 35 U.S.C. § 101.”
Examiner responds:
Applicant argues the claims are not directed to an abstract idea and then admits their claims are directed to an abstract idea, but integrated into a practical application as in Desjardins.
Examiner disagrees with both assertions and begins with Step 2A1 assertion.
Examiner cannot find and Applicant has not pointed to Desjardins rejecting the assessment of "the training of a machine learning model 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."
Examiner points Applicant to July 2024 Subject Matter Eligibility pg. 6: “Step (c) recites training an ANN using a selected algorithm. The training algorithm is a backpropagation algorithm and a gradient descent algorithm. When given their broadest reasonable interpretation in light of the background, the backpropagation algorithm and gradient descent algorithm are mathematical calculations. The plain meaning of these terms are optimization algorithms, which compute neural network parameters using a series of mathematical calculations. The fourth paragraph of the background supports the plain meaning by stating the “gradient descent begins by initializing the values of parameters and then applying a gradient descent calculation, which uses mathematical calculations to iteratively adjust the values so they minimize a loss function.”
Applicant specification para 81 characterizes the machine learning model as “The predictive machine learning model may be based on at least one of neural network, random forest, logistic regression, and gradient boosting learning techniques.”
Therefore Examiner asserts training methods of these models when given the broadest reasonable interpretation in light of the background are mathematical calculations. The plain meaning of these terms are optimization algorithms, which compute parameters using a series of mathematical calculations. Applicant has not pointed and Examiner cannot find disclosure in the specification which characterizes training of a model that does not use a series of mathematical operations and so Examiner uses the plain meaning of training a model which is through the use of mathematical operations and therefore directed to a mathematical concept as in the case above.
Applicant then presents 101 Step 2A2 assertion that the abstract idea was integrated into an practical application as in the case of Desjardins. Examiner disagrees.
Desjardins pg. 9 finds “We are persuaded that constitutes an improvement to how the machine learning model itself operates, and not, for example, the identified mathematical calculation.” Applicant has not pointed to an improvement in how the machine learning model itself operates. Applicant has pointed to using methods which amount to using less training data and then asserts that using less training data results in reduced storage, training speed and cost. Examiner submits this is an improvement to the abstract idea and not the model itself because the model operates the exact same way where the only difference is the data provided to the model where providing less data to the model results in reduced storage, training speed and cost because there is less training data overall. This amounts to improving the abstract idea of a mental process / mathematical operations by consolidating data into a smaller dataset based on features of the data.
Applicant asserts "improvements in training the machine learning model itself' were sufficient integrate an abstract idea into a practical application. Desjardins Pg. 8 recites “Paragraph 21 of the Specification, which the Appellant cites, identifies improvements in training the machine learning model itself. Of course, such an assertion in the Specification alone is insufficient to support a patent eligibility determination, absent a subsequent determination that the claim itself reflects the disclosed improvement.”
Thus Applicant’s assertion insufficient to support a patent eligibility determination absent a subsequent determination that the claim itself reflects the disclosed improvement. Examiner has not found that the claim itself reflects the disclosed asserted improvement.
Examiner notes Desjardins recites claims that integrate the abstract idea into a practical application. Desjardins Pg. 7 recites, “In particular, the Appellant identifies certain limitations of independent claim 1 and asserts that "the claimed subject matter provides technical improvements over conventional systems by addressing challenges in continual learning and model efficiency by reducing storage requirements and preserving task performance across sequential training," citing paragraph 21 of the Specification for support. Id. at 7-9; see also id. at 8 ("This training strategy allows the model to preserve performance on earlier tasks even as it learns new ones, directly addressing the technical problem of 'catastrophic forgetting' in continual learning systems.").”
Examiner agrees constraints or limitations relating to computer storage capacity or memory usage may constitute a technical problem in some contexts such as in Desjardins. However, Examiner cannot find that a computer storage capacity or memory usage or training speed or cost problem that was caused by the computer or model itself. This is contrast to Desjardins whose problem was a challenge in continual learning due to a technical problem of “catastrophic forgetting” in continual learning systems that causes increased storage and task performance. Examiner characterizes Applicant’s problem as directed to the abstract idea of a mental process in organizing data because Applicant asserts there is a problem of storage and learning speed due to using more data than required which is why Applicant asserts to limit the training dataset which would solve this asserted problem of storage and training speed. However, Examiner asserts that the problem of storage and training/learning speed exists regardless of whether a computer or model is involved because having too much data is a problem that exists/has existed regardless in mental process operations. Therefore, organizing a training dataset to be smaller is an improvement to the abstract idea of a mental process in organizing data and not a technical solution to a technical problem.
Further, there is no nexus between the argued problem and the argued solution because there is no indication that the claimed invention actually solves this problem. The Applicant has identified that there is a technical problem relating to increased storage requirements for storing labeled datasets and decreased training speeds due to labeling constraints. Examiner asserts there is no indication that the claim actually solves this problem. The claim does not define what storage constraints there are or training speeds must or must not be and thus the claim may actually result in more data being stored and longer training speeds. Examiner submits generation of the reduced dataset may take more storage itself by using the knowledge graph data object and take more training time to create the dataset. Examiner further points out the operation of the model itself is not changed in contrast to Desjardins which found an improvement in how the model itself operated. Examiner asserts Applicant’s claimed improvement is to the quality of data and an improvement to the abstract idea of a mental process of organizing data. Because the claim does not explicitly solve a technical problem, a practical application is not present.
Applicant argues pg. 15:
“The MPEP defines the mental processes grouping as "concepts performed in the human mind (including observation, evaluation, judgment, opinion)." MPEP § 2106.04(a)(2)(11). It further notes that" [ c ]laims do not recite a mental process when they do not contain limitations that can ractically be performed in the human mind, for instance when the human mind is not equipped to perform the claim limitation." Id Following this guidance, in an informative decision, the Patent Trial and Appeal Board (PTAB) held that a claim directed to speech recognition systems did not recite a mental process because the claim included steps such as "receiving predicted character probabilities from a trained neural network" that "are not steps that can be practically performed mentally." See e.g., Hannun, No. 2018-003323, p. 9-10. Like the claims of Hannun, claim 1 recites a machine learning technique that cannot as a practical matter be performed in the human mind.
[…] As amended, claim 1 recites a new approach to model creation in which a permutation-specific machine learning model is created in response to receiving a request. The human mind cannot practically (i) receive a request from a user interface, (ii) generate a trained machine learning model in response to the request, or (iii) use the trained machine learning model to output a response to the request. Accordingly, no element of claim 1, as amended, under its broadest reasonable interpretation may be considered a mental process as defined by the MPEP.
Examiner responds
The Examiner respectfully disagrees. MPEP 2106.04(a)(2)(III)(A) states that a claimed invention is directed to a mental process if the identified claim elements contain limitations that the human mind is equipped to perform. Abstract ideas that have been held to be practically performable in the human mind include collection/analysis of data, collection/comparison of data, and identifying/applying hair designs. The Examiner submits that Applicant’s claims fall within the mental process grouping of abstract ideas and the Applicant identified a machine learning technique as cannot be practically performed in the human mind. However, the technique is directed to relating data to reduce the training set size which is not a machine learning technique. It is a technique of data consolidation by observing data, comparing, and judging relationships. The human mind can observe data, compare data, and judge relationships between data to determine predictive features of an inner cohort and label a dataset. Because the identified features of the claim can be practically performed in the human mind, the claims are directed to an abstract idea. A machine learning model is used in the claim and is an additional element that is addressed below under Step 2A2.
The claim further recites the additional element of an interactive user interface. The interactive user interface merely generally links the abstract idea to a particular technological environment or field of use. MPEP 2106.04(d)(I) indicates that generally linking an abstract idea to a particular technological environment or field of use cannot provide a practical application. Utilization of the interactive user equates to saying “apply it.” MPEP 2106.04(d)(I) indicates that merely saying “apply it” or equivalent to the abstract idea cannot provide a practical application. Accordingly, even in combination, this additional element does not integrate the abstract idea into a practical application.
The claim further recites the additional element of generating a trained machine learning model to identify a probability that an entity is associated with a prediction feature. 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.
Applicant argues pg. 15-16:
“The MPEP defines the mathematical concepts grouping as "mathematical relationships, mathematical formulas or equations, mathematical equations." Id.,§ 2106.04(a)(2)(I). The MPEP further states "[w]hen determining whether a claim recites a mathematical concept (i.e., mathematical relationships, mathematical formulas or equations, and mathematical calculations), examiners should consider whether the claim recites a mathematical concept or merely limitations that are based on or involve a mathematical concept." Id.,§ 2106.04(a)(2)(I). As explained above, claim 1 recites a new approach to model creation. While this involves the use of "mathematical relationships" ultimately modeled by the machine learning model, claim 1 does not recite these concepts. Rather, claim 1 is directed to techniques for creating new, permutation-specific machine learning models from minimal user input. It does not recite, nor rely upon specific mathematical concepts, such as backpropagation of errors. Consistent with USPTO subject matter eligibility examples, such as Example 47, the mere recitation of training within the claims - without specifying the selected algorithm - is not a mathematical concept. See e.g., USPTO Memorandum: Reminders on evaluating subject matter eligibility of claims under 35 U.S.C. 101, dated August 4, 2025 ("Even though "training the neural network" involves a broad array of techniques and/or activities that may involve or rely upon mathematical concepts, the limitation does not set forth or describe any mathematical relationships, calculations, formulas, or equations using words or mathematical symbols"). For at least these reasons, Applicant respectfully requests withdrawal of the rejection under 35 U.S.C. § 101 because the claimed invention is not directed to a judicial exception under prong one of Step 2A..”
Examiner responds:
The claim further recites “training the trained machine learning model” When given its broadest reasonable interpretation in light of the disclosure, the training of a machine learning model 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.
Examiner points Applicant to July 2024 Subject Matter Eligibility pg. 6: “Step (c) recites training an ANN using a selected algorithm. The training algorithm is a backpropagation algorithm and a gradient descent algorithm. When given their broadest reasonable interpretation in light of the background, the backpropagation algorithm and gradient descent algorithm are mathematical calculations. The plain meaning of these terms are optimization algorithms, which compute neural network parameters using a series of mathematical calculations. The fourth paragraph of the background supports the plain meaning by stating the “gradient descent begins by initializing the values of parameters and then applying a gradient descent calculation, which uses mathematical calculations to iteratively adjust the values so they minimize a loss function.”
Applicant specification para 81 characterizes the machine learning model as “The predictive machine learning model may be based on at least one of neural network, random forest, logistic regression, and gradient boosting learning techniques.”
Therefore Examiner asserts training methods of these models when given the broadest reasonable interpretation in light of the background are mathematical calculations. The plain meaning of these terms are optimization algorithms, which compute parameters using a series of mathematical calculations. Applicant has not pointed and Examiner cannot find disclosure in the specification which characterizes training of a model that does not use a series of mathematical operations and so Examiner uses the plain meaning of training a model which is through the use of mathematical operations and therefore directed to a mathematical concept as in the case above.
Applicant argues pg. 16-17
The MPEP states that "limitations the courts have found indicative that an additional element ( or combination of elements) may have integrated the exception into a practical application include an improvement in the functioning of a computer, or an improvement to other technology or technical field." MPEP § 2106.04(d). Ex Parte Desjardines found that improvements in training a machine learning model itself is sufficient to integrate an exception into a practical application. Ex Parte Desjardines, p. 8. Claim 1 recites an improved training technique for a machine learning model that improves the computational efficiency, storage-wise efficiency, and speed of training machine learning models. See Specification, ,-i,i [0017] and [0019]-[0020].
[…] Claim 1 recites a model creation approach in which a machine learning model is created from a request with minimal, high level user input. See Specification ,i,i [0019], [0041], [0072]. As claimed, a permutation-specific model is created based only on definition data regarding the specific permutation and without storing training data for the specific permutation. See Specification ,i,i [0020], [0067], and [0084]. By doing so, claim 1 recites an improvement in machine learning training that addresses a technical challenge unique to machine learning -increased storage requirements for storing labeled datasets and decreased training speeds due to labeling constraints.
Thus, by "generating, without user input and prior to training the trained machine learning model, a training dataset specific to a request based at least in part on the one or more inner cohort features and the target population", the techniques recited by claim 1 prevent permutation specific training data from being stored by the system, thereby reducing storage requirements. Moreover, by "generating the trained machine learning model by training a machine learning model based at least in part on the training dataset, wherein training is limited to the training dataset and the training dataset is excluded from being stored in memory", the techniques recited in claim 1 reduce labeling constraints on training, enabling increased training speeds at reduced memory cost.
Examiner responds:
Examiner respectfully disagrees. Examiner’s analysis considers and is consistent with the viewpoints expressed by the Appeals Review Panel in Ex Parte Desjardins. Applicant asserts "improvements in training the machine learning model itself' were sufficient integrate an abstract idea into a practical application. Desjardins Pg. 8 recites “Paragraph 21 of the Specification, which the Appellant cites, identifies improvements in training the machine learning model itself. Of course, such an assertion in the Specification alone is insufficient to support a patent eligibility determination, absent a subsequent determination that the claim itself reflects the disclosed improvement.”
Thus Applicant’s assertion insufficient to support a patent eligibility determination absent a subsequent determination that the claim itself reflects the disclosed improvement. Examiner has not found that the claim itself reflects the disclosed asserted improvement.
Examiner notes Desjardins recites claims that integrate the abstract idea into a practical application. Desjardins Pg. 7 recites, “In particular, the Appellant identifies certain limitations of independent claim 1 and asserts that "the claimed subject matter provides technical improvements over conventional systems by addressing challenges in continual learning and model efficiency by reducing storage requirements and preserving task performance across sequential training," citing paragraph 21 of the Specification for support. Id. at 7-9; see also id. at 8 ("This training strategy allows the model to preserve performance on earlier tasks even as it learns new ones, directly addressing the technical problem of 'catastrophic forgetting' in continual learning systems.").”
Examiner agrees constraints or limitations relating to computer storage capacity or memory usage may constitute a technical problem in some contexts such as in Desjardins. However, Examiner cannot find that a computer storage capacity or memory usage or training speed or cost problem that was caused by the computer or model itself. This is contrast to Desjardins whose problem was a challenge in continual learning due to a technical problem of “catastrophic forgetting” in continual learning systems that causes increased storage and task performance. Examiner characterizes Applicant’s problem as directed to the abstract idea of a mental process in organizing data because Applicant asserts there is a problem of storage and learning speed due to using more data than required which is why Applicant asserts to limit the training dataset which would solve this asserted problem of storage and training speed. However, Examiner asserts that the problem of storage and training/learning speed exists regardless of whether a computer or model is involved because having too much data is a problem that exists/has existed regardless in mental process operations to make conclusions. Therefore, organizing a training dataset to be smaller is an improvement to the abstract idea of a mental process in organizing data and not a technical solution to a technical problem.
Further, there is no nexus between the argued problem and the argued solution because there is no indication that the claimed invention actually solves this problem. The Applicant has identified that there is a technical problem relating to increased storage requirements for storing labeled datasets and decreased training speeds due to labeling constraints. Examiner asserts there is no indication that the claim actually solves this problem. The claim does not define what storage constraints there are or training speeds must or must not be and thus the claim may actually result in more data being stored and longer training speeds. Examiner submits generation of the reduced dataset may take more storage itself by using the knowledge graph data object and take more training time to create the dataset. Examiner further points out the operation of the model itself is not changed in contrast to Desjardins which found an improvement in how the model itself operated. Examiner asserts Applicant’s claimed improvement is to the quality of data and an improvement to the abstract idea of a mental process of organizing data. Because the claim does not explicitly solve a technical problem, a practical application is not present.
Applicant argues pg. 18:
In response to Applicant's previous response, the Office Action states that "the training of a machine learning model 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." Office Action, p. 25. This statement is (1) not supported by any legal support, and (2) directly conflicts with guidance provided by the Appeal Review Panel in Ex Parte Desjardines, where the Panel expressly recognized machine learning training as constituting "an improvement to how the machine learning model itself operations, and not, for example, the identified mathematical calculation." Ex Parte Desjardines, p. 9."
Examiner responds:
Examiner points to July 2024 Subject Matter Eligibility pg. 6: “Step (c) recites training an ANN using a selected algorithm. The training algorithm is a backpropagation algorithm and a gradient descent algorithm. When given their broadest reasonable interpretation in light of the background, the backpropagation algorithm and gradient descent algorithm are mathematical calculations. The plain meaning of these terms are optimization algorithms, which compute neural network parameters using a series of mathematical calculations. The fourth paragraph of the background supports the plain meaning by stating the “gradient descent begins by initializing the values of parameters and then applying a gradient descent calculation, which uses mathematical calculations to iteratively adjust the values so they minimize a loss function.”
Applicant specification para 81 characterizes the machine learning model as “The predictive machine learning model may be based on at least one of neural network, random forest, logistic regression, and gradient boosting learning techniques.”
Therefore Examiner asserts training methods of these models when given the broadest reasonable interpretation in light of the background training are mathematical calculations. The plain meaning of these terms are optimization algorithms, which compute parameters using a series of mathematical calculations. Applicant has not pointed and Examiner cannot find disclosure in the specification which characterizes training of a model that does not use a series of mathematical operations.
Further Examiner disagrees that Desjardins broadly “recognized machine learning training as constituting "an improvement to how the machine learning model itself operations, and not, for example, the identified mathematical calculation." Desjardins pointed to specific technical problems in continual learning due to catastrophic forgetting. Desjardins then pointed to specific claim language in claim 1 that reflects the disclosed improvement by “"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.” In contrast, Applicant has not pointed to a technical problem nor a technical solution nor claim language that reflects a disclosed improvement.
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