DEVELOPING ADAPTABLE PREDICTIVE ANALYTICS FOR SUBJECTS IN MEDICAL FACILITIES

§101 §103 §112

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 . DETAILED ACTION 1. The amendment filed 08/18/2026 has been received and considered. Claims 1-20 are presented for examination. 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. 2. Claims 1-12 are rejected under 35 U.S.C. 101 because the claimed invention recites a judicial exception, is directed to that judicial exception, an abstract idea, as it has not been integrated into practical application and the claims further do not recite significantly more than the judicial exception. (Step 1) The claim 1-11 is directed methods and fall within the statutory category of processes; and, the claim 12-18 is directed to an apparatus and falls within the statutory category of machines. The claim 19-20 is directed to a non-transitory computer readable medium which is one of the statutory categories of invention. (Step 2A – Prong One) For the sake of identifying the abstract ideas, a copy of the claim is provided below. Abstract ideas are bolded. Claim 1, 12, and 19 recite: “training a predictive algorithm for predicting an adverse medical event at a desired notification time before occurrence of the adverse medical event using initial annotations indicating times for diagnosis of the adverse medical event (insignificant extra-solution activity – data gathering and/or field of use); determining real risk scores over time using the predictive algorithm applied to the subjects in the medical facility, and identifying a corresponding trend of the real risk scores (under its broadest reasonable interpretation, a mathematical concept and a mental process that convers performance in the human mind or with the aid of pencil and paper including an observation, evaluation, judgment or opinion), wherein the real risk scores indicate actual probabilities of the adverse medical event occurring at predetermined times before the occurrence of the adverse medical event (insignificant extra-solution activity – data gathering and/or field of use); creating required risk scores over time based on the real risk score trend (under its broadest reasonable interpretation, a mathematical concept and a mental process that convers performance in the human mind or with the aid of pencil and paper including an observation, evaluation, judgment or opinion), wherein the required risk scores indicate modified probabilities of the adverse medical event occurring at the predetermined times before the occurrence of the adverse medical event, thereby eliminating probabilities of the adverse medical event occurring before the desired notification time (insignificant extra-solution activity – data gathering and/or field of use); mapping the initial annotations to a time-series of new annotations that minimizes differences between the required risk scores and the real risk scores (under its broadest reasonable interpretation, a mathematical concept and a mental process that convers performance in the human mind or with the aid of pencil and paper including an observation, evaluation, judgment or opinion); fine-tuning the predictive algorithm for predicting the adverse medical event at the desired notification time using the time-series of new annotations (under its broadest reasonable interpretation, a mathematical concept and a mental process that convers performance in the human mind or with the aid of pencil and paper including an observation, evaluation, judgment or opinion); and monitoring at least one of the subjects in the medical facility by applying the fine-tuned predictive algorithm to indicate the desired notification time (insignificant extra-solution activity - “apply it” and/or field of use). Therefore, the limitations, under the broadest reasonable interpretation, have been identified to recite judicial exceptions, an abstract idea. (Step 2A – Prong Two: integration into practical application) This judicial exception is not integrated into a practical application. In particular, the claims recite the following additional elements of “A system… an interface… a processor… a memory” (Claim 12) and “a display” (Claim 16) which is recited at high level generality and recited so generally that they represent more than mere instruction to apply the judicial exception on a computer (see MPEP 2106.05(f)). The limitation can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of a computer (see MPEP 2106.05(d)). Further the additional limitation “the predictive algorithm” and “recurrent neural network (RNN)-based model” is an insignificant extra-solution activity which is generally linking the use of a judicial exception to a particular technological environment or field of use. Further Claims recite the limitation which is an insignificant extra-solution activity because it is a mere nominal or tangential addition to the claim, amounts to mere data gathering (see MPEP 2106.05(g)): (Claim 1, 12 and 19) “training a predictive algorithm for predicting an adverse medical event at a desired notification time before occurrence of the adverse medical event using initial annotations indicating times for diagnosis of the adverse medical event” (insignificant extra-solution activity – data gathering and/or field of use); wherein the real risk scores indicate actual probabilities of the adverse medical event occurring at predetermined times before the occurrence of the adverse medical event (insignificant extra-solution activity – data gathering and/or field of use); (Claim 12) “receiving initial annotations indicating times for diagnosis of an adverse medical event in subjects of the medical facility” (insignificant extra-solution activity – data gathering and/or field of use)”. Further the claims recite the limitation which insignificant extra-solution activity for the act of outputting itself , is equivalent to “apply it”, and/or generally linking the use of a judicial exception to a particular technological environment or field of use (see MPEP § 2106.05(h)): (Claim 1, 12, and 19) “wherein the required risk scores indicate modified probabilities of the adverse medical event occurring at the predetermined times before the occurrence of the adverse medical event, thereby eliminating probabilities of the adverse medical event occurring before the desired notification time (insignificant extra-solution activity – data gathering and/or field of use); monitoring at least one of the subjects in the medical facility by applying the fine-tuned predictive algorithm to indicate the desired notification time (insignificant extra-solution activity - “apply it” and/or field of use)”. Even when viewed in combination, these additional elements do not integrate the recited judicial exception into a practical application and the claim is directed to the judicial exception. (Step 2B - inventive concept) The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements of “A system… an interface… a processor… a memory” (Claim 12) and “a display” (Claim 16) which is recited at high level generality and recited so generally that they represent more than mere instruction to apply the judicial exception on a computer (see MPEP 2106.05(f)). The limitation can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of a computer (see MPEP 2106.05(d)). Further the additional limitation “the predictive algorithm” and “recurrent neural network (RNN)-based model” is an insignificant extra-solution activity which is generally linking the use of a judicial exception to a particular technological environment or field of use (see MPEP § 2106.05(h)). Further Claims recite the limitation which is an insignificant extra-solution activity because it is a mere nominal or tangential addition to the claim, amounts to mere data gathering (see MPEP 2106.05(g)) which is the element that the courts have recognized as well-understood, routine, conventional activity (see MPEP 2106.05(d) II. i. Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); TLI Communications LLC v. AV Auto. LLC, 823 F.3d 607, 610, 118 USPQ2d 1744, 1745 (Fed. Cir. 2016) (using a telephone for image transmission); OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network); but see DDR Holdings, LLC v. Hotels.com, L.P., 773 F.3d 1245, 1258, 113 USPQ2d 1097, 1106 (Fed. Cir. 2014) (“Unlike the claims in Ultramercial, the claims at issue here specify how interactions with the Internet are manipulated to yield a desired result‐‐a result that overrides the routine and conventional sequence of events ordinarily triggered by the click of a hyperlink.” (emphasis added)); iv. Storing and retrieving information in memory, Versata Dev. Group, Inc. v. SAP Am., Inc., 793 F.3d 1306, 1334, 115 USPQ2d 1681, 1701 (Fed. Cir. 2015); OIP Techs., 788 F.3d at 1363, 115 USPQ2d at 1092-93): (Claim 1, 12 and 19) “training a predictive algorithm for predicting an adverse medical event at a desired notification time before occurrence of the adverse medical event using initial annotations indicating times for diagnosis of the adverse medical event” (insignificant extra-solution activity – data gathering and/or field of use); (Claim 12) “receiving initial annotations indicating times for diagnosis of an adverse medical event in subjects of the medical facility” (insignificant extra-solution activity – data gathering and/or field of use). Also the claims recite the limitation which is insignificant extra-solution activity for the act of outputting itself , is equivalent to “apply it”, and/or generally linking the use of a judicial exception to a particular technological environment or field of use (see MPEP § 2106.05(h)): (Claim 1, 12, and 19) “monitoring at least one of the subjects in the medical facility by applying the fine-tuned predictive algorithm to indicate the desired notification time (insignificant extra-solution activity - “apply it” and/or field of use)”. Further dependent claims 2-11, 13-18, and 20 recite: 2. The method of claim 1, wherein mapping the initial annotations to the time-series of new annotations comprises weighting each of the real risk scores by the required risk scores, respectively (a mathematical concept and a mental process). 3. The method of claim 1, wherein a trend of the required risk scores over time has a curve that is the same as a curve of a portion of the real risk score trend (insignificant extra-solution activity – data gathering). 4. The method of claim 1, further comprising: weighting the new annotations and determining a cut-off of the required risk scores in order to prevent early prediction of the adverse medical event before the desired notification time (a mathematical concept and a mental process). 5. The method of claim 1, further comprising adjusting patient management based on the fine-tuned predictive algorithm and/or the monitoring (insignificant extra-solution activity - “apply it” and/or field of use). 6. The method of claim 5, wherein adjusting the patient management comprises adjusting staff and/or resource availability to compensate for the desired notification time being greater or less than an initial notification time (a mental process). 7. The method of claim 5, wherein adjusting the patient management comprises adjusting staff to compensate for a number of patients the desired notification time being greater or less than an initial notification time (a mental process). 8. The method of claim 1, further comprising adjusting a policy of the medical facility based on the fine-tuned predictive algorithm and/or the monitoring (a mental process). 9. The method of claim 1, wherein the predictive algorithm comprises a recurrent neural network (RNN)-based model (insignificant extra-solution activity – generally linking the use of a judicial exception to a particular technological environment or field of use). 10. The method of claim 1, wherein the adverse medical event comprises one of hemodynamics instability (HI), atrial fibrillation, acute kidney injury (AKI), pressure injury (PI), respiratory distress, acute lung injury or risk of infection (insignificant extra-solution activity – data gathering and/or field of use). 11. The method of claim 1, further comprising: providing a real cut-off indicating a predetermined value of the real risk scores (insignificant extra-solution activity – data gathering and/or field of use); optimizing the real cut-off to provide a required cut-off indicating a value of the required risk scores (a mathematical concept and a mental process); and preventing an alert or notification of the adverse medical event at values of the required risk scores below the required cut-off (insignificant extra-solution activity – data outputting). 13. The system of claim 12, wherein mapping the initial annotations to the time-series of new annotations comprises weighting each of the real risk scores by the required risk scores, respectively (a mathematical concept and a mental process). 14. The system of claim 12, wherein a trend of the required risk scores over time has a curve that is the same as a curve of a portion of the real risk score trend (insignificant extra-solution activity – data gathering). 15. The system of claim 12, wherein the instructions further cause the processor to perform weighting of the new annotations and determining a cut-off of the required risk scores in order to prevent early prediction of the adverse medical event before the desired notification time (a mathematical concept and a mental process). 16. The system of claim 12, further comprising: a display for displaying the desired notification time and/or a graph for tracking the desired notification time with respect to the subject (insignificant extra-solution activity – data outputting). 17. The system of claim 16, wherein patient management is adjusted based on at least one of the displayed graph, the fine-tuned predictive algorithm or the monitoring (insignificant extra-solution activity - “apply it” and/or field of use). 18. The system of claim 12, wherein the predictive algorithm comprises a recurrent neural network (RNN)-based model (insignificant extra-solution activity – generally linking the use of a judicial exception to a particular technological environment or field of use). 20. The non-transitory computer readable medium of claim 19, wherein mapping the initial annotations to the time-series of new annotations comprises weighting each of the real risk scores by the required risk scores, respectively (a mathematical concept and a mental process). Considering the claim both individually and in combination, there is no element or combination of elements recited contains any “inventive concept” or adds “significantly more” to transform the abstract concept into a patent-eligible application. Claim Rejections - 35 USC § 103 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. This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention. 3. Claims 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Tomasev (US 11302446 B2) in view of Tomasev 2019 (“A clinically applicable approach to continuous prediction of future acute kidney injury”) and further in view of Hinton (“Distilling the Knowledge in a Neural Network”). As per Claim 1, 12, and 19, Tomasev '446 teaches a method/system/non transitory medium of developing adaptable predictive analytics for subjects in a medical facility (Tomasev '446: Fig. 3, col. 1 lines 33-49 “This specification describes a system that makes predictions that characterize the likelihood that a specific adverse health event will occur to a patient in the future”), the method comprising: (Claim 12) an interface for receiving initial annotations indicating times for diagnosis of an adverse medical event in subjects of the medical facility; a processor in communication with the interface; and a memory that stores instructions that, when executed by the processor, causes the processor (Tomasev '446: col. 2, lines 45–60: “there is provided a system comprising one or more computers and one or more storage devices storing instructions that, when executed by the one or more computers, cause the one or more computers to perform operations comprising receiving electronic health record data for a patient”; col. 9, lines18-24: “provide the user interface presentation for display on the user computer.”). to perform a method comprising: (Claim 1, 12, and 19) training a predictive algorithm for predicting an adverse medical event at a desired notification time before occurrence of the adverse medical event using initial annotations indicating times for diagnosis of the adverse medical event (Tomasev '446: col. 1 lines 33-49 “This specification describes a system that makes predictions that characterize the likelihood that a specific adverse health event will occur to a patient in the future. The predictions are made based on electronic health record data for the patient. In particular, the system generates, from the electronic health record data, an input sequence that includes a respective feature representation at each of a plurality of time window time steps and processes the input sequence using a neural network to generate a neural network output.”; col. 12 lines 33-67: the neural network is trained, against ground-truth targets for each window, to output at each step a separate score for each of a set of selected lead-time windows (24, 48 and 72 hours), and an alert is transmitted when the score for a window exceeds the threshold set for that window). Tomasev '446 fails to teach explicitly determining real risk scores over time using the predictive algorithm applied to the subjects in the medical facility, and identifying a corresponding trend of the real risk scores, wherein the real risk scores indicate actual probabilities of the adverse medical event occurring at predetermined times before the occurrence of the adverse medical event; creating required risk scores over time based on the real risk score trend, wherein the required risk scores indicate modified probabilities of the adverse medical event occurring at the predetermined times before the occurrence of the adverse medical event, thereby eliminating probabilities of the adverse medical event occurring before the desired notification time; mapping the initial annotations to a time-series of new annotations that minimizes differences between the required risk scores and the real risk scores; fine-tuning the predictive algorithm for predicting the adverse medical event at the desired notification time using the time-series of new annotations; and monitoring at least one of the subjects in the medical facility by applying the fine-tuned predictive algorithm to indicate the desired notification time. Tomasev 2019 teaches determining real risk scores over time using the predictive algorithm applied to the subjects in the medical facility, and identifying a corresponding trend of the real risk scores, wherein the real risk scores indicate actual probabilities of the adverse medical event occurring at predetermined times before the occurrence of the adverse medical event (Tomasev 2019 p. 116 right column: “the model outputs a probability of AKI occurring at any stage of severity within the next 48 h (although our approach can be extended to other time windows or severities of AKI; see Extended Data Table 1)… At every point throughout an admission, the model provides updated estimates of future AKI risk along with an associated degree of uncertainty “; p. 121 left column, Methods: “ PNG media_image1.png 182 613 media_image1.png Greyscale ”: the probabilities output at every time point for each of the eight fixed prediction windows form a continuous risk trajectory, plotted for the 48-h window in Fig. 1b against time since admission and rising 48 h before the observed event, i.e., a “trend of the real risk scores” whose values are probabilities of the event at “predetermined times before the occurrence” as claimed); … a time-series of new annotations that minimizes differences between the required risk scores and the real risk scores (p. 121 right column, “ PNG media_image2.png 94 613 media_image2.png Greyscale : sparse diagnosis annotations converted to per-timestep labels via forward-fill”; p. 122 left column, “Each of the resulting eight outputs provides a binary prediction for AKI severity at a specific time window and is compared to the ground-truth label using the cross-entropy loss function (Bernoulli log-likelihood).”: cross-entropy loss between predicted and target distributions is the divergence-minimization operation); thereby eliminating probabilities of the adverse medical event occurring before the desired notification time (Methods, 'Ground-truth labels using KDIGO', left col. “The prediction target at each point in time is a binary variable that is positive if the AKI category of interest … occurs within a chosen future time horizon. If no AKI state was recorded within the chosen horizon, this was interpreted as a negative. We used 8 future time horizons (6-h, 12-h, 18-h, 24-h, 36-h, 48-h, 60-h and 72-h ahead) that were all available at each time point.”; p. 117, right col. “Of the false-positive alerts made by our model, 24.9% were positive predictions that were made even earlier than the 48-h window in patients who subsequently developed AKI (Extended Data Fig. 4).”: the correct label at each time step, for the chosen alerting horizon, is positive only inside that horizon and is a negative, i.e., a zero probability, at every earlier time step, and a positive prediction made earlier than the chosen 48-h window is counted as a false positive under the reference's evaluation rather than as a detection); monitoring at least one of the subjects in the medical facility by applying the fine-tuned predictive algorithm to indicate the desired notification time (Figure 1, p. 116 right column, “With our approach, 55.8% of inpatient AKI events of any severity were predicted early, within a window of up to 48 h in advance and with a ratio of 2 false predictions for every true positive… Identifying an increased risk of future AKI sufficiently far in advance is critical, as longer lead times may enable preventative action to be taken”: the described model is applied to each inpatient at each time point throughout the admission and its alert indicates that the event is predicted within the chosen 48-h notification window, i.e., the “desired notification time” as claimed). In particular, Tomasev 2019 teaches that the RNN output is read as a continuous time-varying risk-score trajectory across multiple predetermined lead-time windows which converting sparse initial diagnosis annotations into a per-timestep time-series of ground-truth labels for training. Further Hinton teaches creating required risk scores over time based on the real risk score trend, wherein the required risk scores indicate modified probabilities of the adverse medical event occurring at the predetermined times before the occurrence of the adverse medical event (section Introduction “An obvious way to transfer the generalization ability of the cumbersome model to a small model is to use the class probabilities produced by the cumbersome model as 'soft targets' for training the small model”; section Distillation “Using a higher value for T produces a softer probability distribution over classes.”: the probabilities output by the trained model are softened and taken as “soft targets”, a modified probability distribution produced from the trained model's output probabilities, for training the predictor; in the framework of Tomasev '446 as modified by Tomasev 2019 the trained model's outputs are the per-time-step real risk scores for the chosen notification horizon, that horizon being one of the eight predetermined horizons, so the soft-target series, as modified by the correct labels as set out next, is a per-time-step profile of modified probabilities of the event at the predetermined times, i.e., the “required risk scores” as claimed.) ; mapping the initial annotations… (section Distillation : “In the simplest form of distillation, knowledge is transferred to the distilled model by training it on a transfer set and using a soft target distribution for each case in the transfer set that is produced by using the cumbersome model with a high temperature in its softmax”, “The first objective function is the cross entropy with the soft targets and this cross entropy is computed using the same high temperature in the softmax of the distilled model as was used for generating the soft targets from the cumbersome model”: the correct labels, i.e., the initial annotations, are used to modify the soft targets, one of the two ways Hinton describes for training toward both, and the soft target distribution for each case of the transfer set replaces that case's original label as the target on which the predictor is trained, so that each initial annotation is mapped to a new annotation whose value is the required risk score at that time step, and the predictor being trained minimizes the cross entropy between its output probabilities, i.e., the real risk scores it produces, and those targets, i.e., the required risk scores; in the framework of Tomasev 2019 the labels and the outputs are defined per time step, so the new annotations form a time-series); fine-tuning the predictive algorithm for predicting the adverse medical event at the desired notification time using the time-series of new annotations (section Distillation: “we could use the original training set. We have found that using the original training set works well, especially if we add a small term to the objective function that encourages the small model to predict the true targets as well as matching the soft targets provided by the cumbersome model”: the predictor is trained on the original training set, each case of which now carries its soft-target new annotation). In particular, Hinton teaches taking a trained model’s output distribution, applying temperature smoothing, and using the resulting modified probability distribution as a training target and fine-tuning the predictor on the time-series of soft-target labels by re-running training against the modified distribution Tomasev '446, Tomasev 2019 and Hinton are analogous art because they are both from the same field of endeavor, neural networks to model a system for predicting medical event (In re Bigio, 381 F.3d 1320, 1325 (Fed. Cir. 2004); In re Klein, 647 F.3d 1343, 1348 (Fed. Cir. 2011)). It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to combine the teaching of Tomasev 2019 and Hinton int the method for predicting future adverse health events using neural networks of Tomasev '446 in order to provide an improved model with confidence assessments for offering opportunities for identifying patients at risk within a time window that enables early treatment (Tomasev 2019: pg 116 left column) and to provide soft targets that provide much more information per training case than hard targets and much less variance in the gradient between training cases, so that a predictor trained on the original training set generalizes in the same way as the model that produced the targets therefore to significantly improve the model which learn to distinguish fine-grained classes (Hinton: Abstract, § 1 Introduction, p. 2). As per Claim 2, 13, and 20, Tomasev '446 fails to teach explicitly wherein mapping the initial annotations to the time-series of new annotations comprises weighting each of the real risk scores by the required risk scores, respectively. Hinton teaches wherein mapping the initial annotations to the time-series of new annotations comprises weighting each of the real risk scores by the required risk scores, respectively (section Distillation “The first objective function is the cross entropy with the soft targets and this cross entropy is computed using the same high temperature in the softmax of the distilled model as was used for generating the soft targets from the cumbersome model.”: Cross-entropy of predicted distribution against soft targets is mathematically a weighting of each real probability by the corresponding required-target value). As per Claim 3 and 14, Tomasev '446 fails to teach explicitly wherein a trend of the required risk scores over time has a curve that the same as a curve of a portion of the real risk score trend. Hinton teaches wherein a trend of the required risk scores over time has a curve that is substantially the same as a curve of the real risk score trend (section Introduction “When we are distilling the knowledge from a large model into a small one, however, we can train the small model to generalize in the same way as the large model”; section Soft Targets as Regularizers “the same model trained with soft targets is able to recover almost all the information in the full training set”: soft-target-trained model recovers the full information of the source distribution). As per Claim 4 and 15, Tomasev '446 fails to teach explicitly further comprising: weighting the new annotations and determining a cut-off of the required risk scores in order to prevent early prediction of the adverse medical event before the desired notification time. Tomasev 2019 teaches determining a cut-off of the required risk scores in order to prevent early prediction of the adverse medical event before the desired notification time (pg 116 right column, “When the predicted probability exceeds a specified operating-point threshold, the prediction is considered positive.”: a cut-off on the output probability determines when a positive prediction is made, and the operating point is chosen according to clinical priority to reduce the percentage of false predictions, among which the reference counts predictions made earlier than the 48-h window). Hinton teaches weighting the new annotations (p 2 “we add a small term to the objective function that encourages the small model to predict the true targets as well as matching the soft targets provided by the cumbersome model”: the loss term computed from the soft targets, i.e., from the new annotations, is multiplied by the square of the temperature, i.e., the new annotations are weighted). As per Claim 5, Tomasev '446 fails to teach explicitly further comprising adjusting patient management based on the fine-tuned predictive algorithm and/or the monitoring. Tomasev 2019 teaches further comprising adjusting patient management based on the fine-tuned predictive algorithm and/or the monitoring (pg 116 left column, “For predictive alerts to be effective, they must empower clinicians to act before a major clinical decline has occurred”: the alerts of the described model prompt clinicians to act and to take preventative action for the flagged patient before clinical decline, i.e., patient management is adjusted on the basis of the monitoring). As per Claim 6, Tomasev '446 fails to teach explicitly wherein adjusting the patient management comprises adjusting staff and/or resource availability to compensate for the desired notification time being greater or less than an initial notification time. Tomasev 2019 teaches wherein adjusting the patient management comprises adjusting staff and/or resource availability to compensate for the desired notification time being greater or less than an initial notification time (p. 117 left column “To respond to these alerts on a daily basis, clinicians would need to attend to approximately 0.8% of in-hospital patients (Extended Data Table 2).”: the number of in-hospital patients that clinicians must attend to in response to the alerts, approximately 0.8% for stage 3 alerts at the 48-h horizon, and hence the staff and resources required, follows from the alerting configuration, which is set according to clinical priority and which the reference states can be extended to other time windows). As per Claim 7, Tomasev '446 fails to teach explicitly wherein adjusting the patient management comprises adjusting staff to compensate for a number of patients the desired notification time being greater or less than an initial notification time. Tomasev 2019 teaches wherein adjusting the patient management comprises adjusting staff to compensate for a number of patients at the desired notification time being greater or less than an initial notification time (p. 117 left column “To respond to these alerts on a daily basis, clinicians would need to attend to approximately 0.8% of in-hospital patients (Extended Data Table 2).”: the number of patients flagged at the chosen 48-h notification horizon, approximately 0.8% of in-hospital patients per day for stage 3 alerts, is the number of patients the clinical staff must attend to, and that number changes when the alerting configuration is set for a different horizon or operating point according to clinical priority). As per Claim 8, Tomasev '446 fails to teach explicitly further comprising adjusting a policy of the medical facility based on the fine-tuned predictive algorithm and/or the monitoring. Tomasev 2019 teaches further comprising adjusting a policy of the medical facility based on the fine-tuned predictive algorithm and/or the monitoring (p. 117 left column “An operating point can be chosen to further increase the proportion of AKI that is predicted early or to reduce the percentage of false predictions at each step, according to clinical priority”; p. 121 left column “We use a stacked multiple-layer recurrent network with highway connections… We use the simple recurrent unit network as the RNN architecture, with tanh activations.”: choosing the operating point of the described model according to clinical priority, trading earlier prediction against fewer false predictions, is a facility-level alerting policy set on the basis of the model's monitoring output). As per Claim 9 and 18, Tomasev '446 teaches wherein the predictive algorithm comprises a recurrent neural network (RNN)-based model (col. 10 lines 47-56: the predictive neural network is built on recurrent neural network layers). As per Claim 10, Tomasev '446 fails to teach explicitly wherein the adverse medical event comprises one of hemodynamics instability (HI), atrial fibrillation, acute kidney injury (AKI), pressure injury (PI), respiratory distress, acute lung injury or risk of infection. Tomasev 2019 teaches wherein the adverse medical event comprises one of hemodynamics instability (HI), atrial fibrillation, acute kidney injury (AKI), pressure injury (PI), respiratory distress, acute lung injury or risk of infection (Title “acute kidney injury”: acute kidney injury is one member of the recited group of adverse medical events). As per Claim 11, Tomasev '446 fails to teach explicitly further comprising: providing a real cut-off indicating a predetermined value of the real risk scores; optimizing the real cut-off to provide a required cut-off indicating a value of the required risk scores; and preventing an alert or notification of the adverse medical event at values of the required risk scores below the required cut-off. Tomasev 2019 teaches providing a real cut-off indicating a predetermined value of the real risk scores (pg 116 right column, “When the predicted probability exceeds a specified operating-point threshold, the prediction is considered positive.”; p. 117 left column “An operating point can be chosen to further increase the proportion of AKI that is predicted early or to reduce the percentage of false predictions at each step, according to clinical priority”: the chosen operating threshold on the trained model's output probabilities is a predetermined cut-off value of the real risk scores); preventing an alert or notification of the adverse medical event at values of the required risk scores below the required cut-off (pg 116 right column, “When the predicted probability exceeds a specified operating-point threshold, the prediction is considered positive.”; p. 117 left column “An operating point can be chosen to further increase the proportion of AKI that is predicted early or to reduce the percentage of false predictions at each step, according to clinical priority”: a positive prediction, and hence an alert, is made only when the output probability exceeds the chosen threshold, so that no alert is triggered at values below the cut-off). Hinton teaches optimizing the real cut-off to provide a required cut-off indicating a value of the required risk scores (p 2 “we add a small term to the objective function that encourages the small model to predict the true targets as well as matching the soft targets provided by the cumbersome model”). As per Claim 16, Tomasev '446 fails to teach explicitly further comprising: a display for displaying the desired notification time and/or a graph for tracking the desired notification time with respect to the subject. Tomasev 2019 teaches further comprising: a display for displaying the desired notification time and/or a graph for tracking the desired notification time with respect to the subject (Fig. 1: Fig. 1b is a graph of a patient's predicted risk of AKI within 48 h plotted against time since admission). As per Claim 17, Tomasev '446 fails to teach explicitly wherein patient management is adjusted based on at least one of the displayed graph, the fine-tuned predictive algorithm or the monitoring. Tomasev 2019 teaches wherein patient management is adjusted based on at least one of the displayed graph, the fine-tuned predictive algorithm or the monitoring (p 116 “predictive alerts… must empower clinicians to act before a major clinical decline has occurred”, “Identifying an increased risk of future AKI sufficiently far in advance is critical, as longer lead times may enable preventative action to be taken”, Fig. 1: the alerts of the described model prompt clinicians to act and to take preventative action for the flagged patient before clinical decline, i.e., patient management is adjusted on the basis of the monitoring). Response to Arguments 4. Applicant's arguments filed 08/18/2026 have been fully considered but they are not persuasive. Examiner respectfully withdraws Claim Objections in view of the amendment and/or applicant’s arguments. Examiner respectfully withdraws Claim Rejections - 35 USC § 112 in view of the amendment and/or applicant’s arguments. As per 101 rejection: Step 2A, Prong One, (Remarks, pp. 9-10): applicants have argued that PNG media_image3.png 367 669 media_image3.png Greyscale The argument treats the mathematical-concept grouping as limited to claims that recite a formula or an equation. Under MPEP 2106.04(a)(2)(I)(C), a claim recites a mathematical calculation when the calculation is expressed in words, and the claim need not use the word "calculate" or recite a formula. The limitations identified in the rejection express calculations in words : “determining real risk scores over time using the predictive algorithm applied to the subjects in the medical facility, and identifying a corresponding trend of the real risk scores”, “creating required risk scores over time based on the real risk score trend”, “mapping the initial annotations to a time-series of new annotations”, and “fine-tuning the predictive algorithm for predicting the adverse medical event at the desired notification time using the time-series of new annotations” which recite computing a series of probabilities and how they change with time, deriving a second series of probabilities from the first by changing its values, and minimization (see the Spec. paragraph [0042], [0045]-[0047], Table 1). Therefore, the limitations, under the broadest reasonable interpretation, have been identified to recite judicial exceptions, an abstract idea. Applicants have argued that: PNG media_image4.png 302 675 media_image4.png Greyscale Unlike Example 38 where the audio-mixer claim recited initializing a circuit model and generating random values with a pseudo-random number generator, and did not recite any relationship among numerical quantities, the claim 1 recites the relationship between the real and required scores and the relationship between the new annotations and those scores as explained previously. Therefore, the limitations, under the broadest reasonable interpretation, have been identified to recite judicial exceptions, an abstract idea. Applicants have argued that: PNG media_image5.png 487 686 media_image5.png Greyscale Applicant's argument is more specific than the claim language. Claim 1 recites "a predictive algorithm" and does not recite artificial intelligence, machine learning, or a neural network; paragraph [0020] itself says only that the predictive algorithm "may be an AI model, such as a recurrent neural network (RNN) based model, for example." Under the broadest reasonable interpretation, "a predictive algorithm" encompasses any rule that assigns a risk score to a patient from observed data, including the additive early-warning scores that clinicians compute by hand, and the specification's own illustration of the claimed steps is a table of hourly values in which the annotations are zeroed before the notification time and weighted afterward (paras. [0042], [0045]-[0047], Table 1), an operation that can be carried out with pencil and paper. A limitation that can practically be performed in the human mind recites a mental process even when the claim requires a computer or a generic tool (MPEP 2106.04(a)(2)(III)(C)). reciting that the scores are obtained "using the predictive algorithm" therefore does not take identifying the trend, creating the modified scores, or mapping the annotations out of the mental-process grouping. Reciting that the scores are obtained "using the predictive algorithm" therefore does not take identifying the trend, creating the modified scores, or mapping the annotations out of the mental-process grouping. The specification confirms the point: the required scores are created from "feedback ... collected by the medical facility," and that feedback "may be provided manually by the clinicians" (paras. [0021], [0056]), which is human evaluation and judgment. Therefore, the limitations, under the broadest reasonable interpretation, have been identified to recite judicial exceptions, an abstract idea. Step 2A, Prong Two (Remarks, pp. 11-14): applicants have argued that PNG media_image6.png 454 677 media_image6.png Greyscale First, it is noted that the rejection's Prong Two analysis does not rest on whether any element is well-understood, routine, or conventional; it evaluates the additional elements under MPEP 2106.05(f), (g), and (h), as MPEP 2106.04(d) directs. The well-understood, routine, conventional discussion appears only in Step 2B, where MPEP 2106.05(d) places it. Second, most of the limitations Applicant relies on for integration are the judicial exception itself. The limitation creating required risk scores that eliminate probabilities before the desired notification time is the modification of the score series identified under Prong One, and the "thereby" clause states the result of that calculation. It is not an additional element. As explained above, the claims recite additional elements: “A system… an interface… a processor… a memory” (Claim 12) and “a display” (Claim 16) which is recited at high level generality and recited so generally that they represent more than mere instruction to apply the judicial exception on a computer (see MPEP 2106.05(f)) and which can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of a computer (see MPEP 2106.05(d)). “the predictive algorithm” and “recurrent neural network (RNN)-based model” which is an insignificant extra-solution activity and generally linking the use of a judicial exception to a particular technological environment or field of use. (see MPEP 2106.05(h)) (Claim 1, 12 and 19) “training a predictive algorithm for predicting an adverse medical event at a desired notification time before occurrence of the adverse medical event using initial annotations indicating times for diagnosis of the adverse medical event” (insignificant extra-solution activity – data gathering and/or field of use); wherein the real risk scores indicate actual probabilities of the adverse medical event occurring at predetermined times before the occurrence of the adverse medical event (insignificant extra-solution activity – data gathering and/or field of use); (Claim 12) “receiving initial annotations indicating times for diagnosis of an adverse medical event in subjects of the medical facility” (insignificant extra-solution activity – data gathering and/or field of use)” - which is an insignificant extra-solution activity because it is a mere nominal or tangential addition to the claim, amounts to mere data gathering (see MPEP 2106.05(g)) (Claim 1, 12, and 19) “monitoring at least one of the subjects in the medical facility by applying the fine-tuned predictive algorithm to indicate the desired notification time (insignificant extra-solution activity - “apply it” and/or field of use)” - which is insignificant extra-solution activity for the act of outputting itself , is equivalent to “apply it”, and/or generally linking the use of a judicial exception to a particular technological environment or field of use (see MPEP § 2106.05(h)). Further applicant have argued that: PNG media_image7.png 294 681 media_image7.png Greyscale … PNG media_image8.png 184 677 media_image8.png Greyscale The asserted improvement is not an improvement to technology. Most of the limitations Applicant relies on Claim 1 recites abstract idea; an improvement in the abstract idea itself, or in a clinical or administrative practice, does not qualify ((MPEP 2106.05(a)(II): " However, it is important to keep in mind that an improvement in the abstract idea itself (e.g. a recited fundamental economic concept) is not an improvement in technology"). Further it does not change how any computer, sensor, or monitoring device operates. Applicant's own characterization of the improvement, accurately predicting probabilities "sufficient to meet clinical requirements," describes a better-timed output of the mathematical model, and an improvement in the mathematical analysis is not a technological improvement. "Monitoring" is recited only as applying the algorithm to a subject "to indicate the desired notification time," which is insignificant extra-solution activity for the act of outputting itself , is equivalent to “apply it”, and/or generally linking the use of a judicial exception to a particular technological environment or field of use (see MPEP § 2106.05(h)). The specification frames the problem as a clinical and administrative one: an early prediction "may be considered a false positive under the clinical requirements or policies," "may result in a premature response," and may conflict with "insufficient resources" or cause "overstaffing" (para. [0002]); the notification time is chosen "based on the particular needs and policies of the medical facility," including "staffing and other resource availability" (paras. [0038]-[0039]); and the stated benefit is that the method "improves patient management by the medical facility ... in a cost effective manner" (para. [0021]). Shifting the time at which a risk score is reported so that it matches a facility's staffing policy changes when a prediction is communicated; it does not change how any computer, sensor, or monitoring device operates. Instead here the computer and the predictive algorithm are unchanged and serve only to carry out the score adjustments. Unlike Enfish where the improvement was to the structure of the database itself, a change in computer functionality that the claims reflected, the computer and the predictive algorithm of the instant invention are unchanged and serve only to carry out the score adjustments. In Ex parte Desjardins, the claims recited a particular technique for training the machine-learning model that the specification described as improving the model's own functioning. However, the present claims recite no particular training or fine-tuning technique; they recite fine-tuning "using the time-series of new annotations," which is applying an unspecified learning procedure to relabeled data; the additional elements of computer component are recited so generally that they represent more than mere instruction to apply the judicial exception on a computer (see MPEP 2106.05(f)). The limitation can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of a computer (see MPEP 2106.05(d)). Considered individually and as an ordered combination, the additional elements do not integrate the exception into a practical application. Step 2B (Remarks, pp. 14-16): Applicants have argued that the training, required-score, fine-tuning, and monitoring limitations, individually and as an ordered combination, amount to significantly more, citing BASCOM and the same asserted improvement to automated patient monitoring. For the reasons given under Prong Two, the limitation “training a predictive algorithm for predicting an adverse medical event at a desired notification time before occurrence of the adverse medical event using initial annotations indicating times for diagnosis of the adverse medical event” (insignificant extra-solution activity – data gathering and/or field of use)” is an insignificant extra-solution activity because it is a mere nominal or tangential addition to the claim, amounts to mere data gathering (see MPEP 2106.05(g)) which is the element that the courts have recognized as well-understood, routine, conventional activity (see MPEP 2106.05(d) II.) while monitoring limitation is insignificant extra-solution activity for the act of outputting itself , is equivalent to “apply it”, and/or generally linking the use of a judicial exception to a particular technological environment or field of use (see MPEP § 2106.05(h)). The limitation “creating required risk scores over time based on the real risk score trend…” and “fine-tuning the predictive algorithm …”, under the broadest reasonable interpretation, have been identified to recite judicial exceptions, an abstract idea. For the reasons given under Prong Two, the training and fine-tuning limitations are the use of a generic predictive algorithm and generic computer components as tools to implement the exception (see MPEP 2106.05(f)); and, monitoring limitation is insignificant extra-solution activity for the act of outputting itself , is equivalent to “apply it”, and/or generally linking the use of a judicial exception to a particular technological environment or field of use (see MPEP § 2106.05(h)). Further, applicant relies on Claim 1 recites abstract idea; an improvement in the abstract idea itself, or in a clinical or administrative practice, does not qualify ((MPEP 2106.05(a)(II): " However, it is important to keep in mind that an improvement in the abstract idea itself (e.g. a recited fundamental economic concept) is not an improvement in technology"). Further it does not change how any computer, sensor, or monitoring device operates. In Bascom, the combination of additional elements, and specifically "the installation of a filtering tool at a specific location, remote from the end‐users, with customizable filtering features specific to each end user" where the filtering tool at the ISP was able to "identify individual accounts that communicate with the ISP server, and to associate a request for Internet content with a specific individual account," were held to be meaningful limitations because they confined the abstract idea of content filtering to a particular, practical application of the abstract idea. 827 F.3d at 1350-51, 119 USPQ2d at 1243. PNG media_image9.png 18 19 media_image9.png Greyscale Unlike the claims in Bascom, the present claims is the mathematical concept itself as explained above. Neither the problem nor the solution is rooted in computer technology. There is nothing in the claim to suggest that, once generating/providing a linear model is performed, the computer functionalities are improved. Likewise, the remaining dependent Claims recite no additional claim limitations that would amount to significantly more than the abstract idea(s) defined in their respective independent claims, when taken either alone or in combination with the limitations of the claims upon which they depend, because they are further directed to mathematical calculation of linear modeling. Thus 101 rejection maintains. As per 103 rejection: Regarding Tomasev '446 (Remarks, pp. 17-18): Applicants argue that: PNG media_image10.png 547 684 media_image10.png Greyscale Applicants' argument is more specific than the claim language. The claims recite “training a predictive algorithm for predicting an adverse medical event at a desired notification time before occurrence of the adverse medical event….”, not training a predictive algorithm for predicting an adverse medical event at “a desired notification time preceding the predicted time of an adverse medical event”. In response to applicant's argument that the references fail to show certain features of the invention, it is noted that the features upon which applicant relies (i.e., “training a predictive algorithm …at a desired notification time preceding the predicted time of an adverse medical event) are not recited in the rejected claim(s). Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993). As rejected above, it is Examiner’s position that Tomasev '446 to teach the claimed limitation. In particular, See Tomasev '446 of col. 1 lines 33-49 and col. 12 lines 33-67 where describes that the neural network is trained, against ground-truth targets for each window, to output at each step a separate score for each of a set of selected lead-time windows (24, 48 and 72 hours), and an alert is transmitted when the score for a window exceeds the threshold set for that window, so that the adverse health event is predicted well in advance of its occurrence. A lead-time window whose score is thresholded for alerting, e.g., the 48-hour window, is the “desired notification time” as claimed. Further the claim does not recite how the desired notification time is selected, that it is entered by a user, or any relationship between the desired notification time and the predicted event other than that the notification time precedes the occurrence of the event. Regarding Tomasev 2019 (Remarks, p. 18): Applicants argue that: PNG media_image11.png 392 678 media_image11.png Greyscale Examiner disagrees. As rejected above, Tomasev 2019 states in its Methods that eight future time horizons, from 6 h to 72 h ahead, were all available at each time point, and that the prediction layer makes predictions over all eight windows, and states at p. 116 that the model provides updated estimates of future AKI risk at every point throughout an admission; Fig. 1b plots the resulting continuous risk trajectory for the 48-h window against time since admission and shows the predicted risk rising 48 h before the observed AKI. Probabilities computed at every time point for a fixed set of horizons ahead of that time point are probabilities of the event at predetermined times before its occurrence, and the sequence of those probabilities over the admission is the trend of the real risk scores. Moreover, the probabilities output at every time point for each of the eight fixed prediction windows form a continuous risk trajectory, plotted for the 48-h window in Fig. 1b against time since admission and rising 48 h before the observed event, i.e., a “trend of the real risk scores” whose values are probabilities of the event at “predetermined times before the occurrence” as claimed. Regarding Hinton (Remarks, pp. 18-19): Applicants argue that: PNG media_image12.png 476 689 media_image12.png Greyscale In response to applicant's arguments against the references individually, one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986). Hinton is relied upon only for deriving a modified set of per-case target probabilities from a trained model's output probabilities, using the correct labels to modify those targets, and training the predictor on the original training set against them as its new annotations for teaching the limitation: “creating required risk scores over time based on the real risk score trend, wherein the required risk scores indicate modified probabilities of the adverse medical event occurring at the predetermined times before the occurrence of the adverse medical event”, “mapping the initial annotations…” and “fine-tuning the predictive algorithm for predicting the adverse medical event at the desired notification time using the time-series of new annotations” while Tomasev 2019 is relied upon for a teaching of "thereby eliminating probabilities of the adverse medical event occurring before the desired notification time” where whose correct label at each time point for the chosen horizon is positive only if the event occurs within that horizon and is "interpreted as a negative" otherwise, and which scores predictions made earlier than the 48-h window as false positives. Further when, per Hinton, those correct labels are used to modify the soft targets, it is the examiner's position that the resulting required-score profile carries no probability of the event at any time earlier than the chosen notification horizon, the modified targets being the required risk scores and, as the targets on which the predictor is trained, the new annotations of the claim. As to the purpose for which Hinton derives its soft targets, a reference is available for all that it teaches to one of ordinary skill in the art, including the disclosed alternative of using the correct labels to modify the soft targets, and the motivation for the combination need not be the same as that of the applicant. See MPEP 2123 and MPEP 2144(IV). The rationale for the combination is set forth in the rejection above. Regarding claims 12 and 19 (Remarks, p. 19), Applicant relies on the same arguments presented for claim 1, which are not persuasive for the reasons given above. Regarding claims 2-11, 13-18, and 20 (Remarks, p. 19), Applicant relies on dependence from claims 1, 12, and 19 and on "their additional recitations" without identifying any recitation asserted to be absent from the applied references; those claims remain rejected for the reasons set forth in the rejection above. Thus 103 rejection maintains. Conclusion 5. The prior art made of record and not relied upon is considered pertinent to applicant's disclosure: Hyland, S. L. et al. ("Early prediction of circulatory failure in the intensive care unit using machine learning") teaches a trend of real risk scores indicating probabilities of an adverse medical event at a predetermined time before its occurrence. Morrill, J. et al. ("The Signature-based Model for Early Detection of Sepsis from Electronic Health Records in the Intensive Care Unit") teaches required risk scores that eliminate probabilities of an adverse medical event before a desired notification time and mapping initial annotations to a time-series of new annotations. Winslow (US 2020/0176115 A1) teaches required risk scores that eliminate probabilities of an adverse medical event before a desired notification time. 6. Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a). A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action. 7. Any inquiry concerning this communication or earlier communications from the examiner should be directed to EUNHEE KIM whose telephone number is (571)272-2164. The examiner can normally be reached Monday-Friday 9am-5pm ET. 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, Ryan Pitaro can be reached at (571)272-4071. 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. EUNHEE KIM Primary Examiner Art Unit 2188 /EUNHEE KIM/ Primary Examiner, Art Unit 2188