DETAILED ACTION
Notice to Applicant
This communication is in response to the amendment submitted February 17, 2025. The present application claims priority benefit of U.S. Provisional Patent Application No. 63/363,939, filed April 29, 2022. Claims 1 – 22 are cancelled. Claims 45 – 47 are new. Claims 23 – 47 are presented for examination.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 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 23 – 47 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.
Step One
Claims 23 – 47 are drawn to a system, which is/are statutory categories of invention (Step 1: YES).
Step 2A Prong One
Independent claim 23 recites a system for meal detection, comprising: generating glucose data for a set of input features, the set of input features having glucoregulatory, insulin, and associated time of day features; receive the glucose data; and present to the user, based on the meal detection output, a meal detection indication.
The recited limitations, as drafted, under their broadest reasonable interpretation, cover certain methods of organizing human activity, as reflected in the specification, which states that “This disclosure relates to the use of machine learning for detecting meals and estimating their size” (paragraph 3 of the published specification). If a claim limitation, under its broadest reasonable interpretation, covers managing personal behavior or relationships or interactions between people, then it falls within the “Certain Methods of Organizing Human Activity” grouping of abstract ideas. The present claims cover certain methods of organizing human activity because they address “…. Inaccurate carbohydrate counts that are used for calculation of meal insulin are associated with high prevalence of postprandial hyperglycemia and hypoglycemia leading to suboptimal postprandial glycemic control even with hybrid closed-loop insulin delivery systems.” (paragraph 5 of the published specification). Accordingly, the claims recite an abstract idea(s) (Step 2A Prong One: YES).”
Step 2A Prong Two
This judicial exception is not integrated into a practical application. The claims are abstract but for the inclusion of the additional elements including:
Claim 23: “system”, “one or more wearable glucose sensing and regulating medical devices”, “smart device”, “wirelessly receive from the one or more wearable glucose sensing and regulating medical devices”. “trained machine learning model, the trained machine learning model including multiple connected layers and an output layer providing a meal detection output”, “user interface”
Claim 24: “system”, “a remote server for hosting the trained machine learning model”
Claim 25: “system”, “the trained machine learning model is a trained multioutput machine learning model with the output layer formed from first and second branches, the first branch providing the meal detection output and the second branch providing a carbohydrate estimation output”
Claims 26, 38: “system”, “user interface”
Claims 27, 34, 35: “system”, “trained machine learning model”
Claims 28, 30, 36 – 37, 40, 44: “system”, “smart device”
Claim 29: “system”, “decision support application executed by the smart device”
Claim 31: “system”, “multiple connected layers are fully connected layers”
Claim 32: “system”, “the trained machine learning model is one or more of a neural network, a random forest model, a support vector regression model, and a logistic regression model”
Claim 33: “system”, “the trained machine learning model is trained using one or more of an ordinary differential equation (ODE) in silico model of glucose metabolism and real-world human glucose, insulin, and nutrition data”
Claim 39: “system”, “smart device”, “automatically”
Claim 41: “system”, “the one or more wearable glucose sensing and regulating medical devices include a continuous glucose monitoring (CGM) device”
Claim 42: “system”, “the one or more wearable glucose sensing and regulating medical devices include an insulin pen or an insulin pump”
Claim 43: “system”, “smart device is configured to wirelessly receive, via a wireless personal area network from the one or more wearable glucose sensing and regulating medical devices”
Claims 45 – 47: “system”
These features are additional elements that are recited at a high level of generality such that they amount to no more than mere instruction to apply the exception using generic computer components. See: MPEP 2106.05(f).
The additional elements are merely incidental or token additions to the claim that do not alter or affect how the process steps or functions in the abstract idea are performed. Therefore, the claimed additional elements do not add meaningful limitations to the indicated claims beyond a general linking to a technological environment. See: MPEP 2106.05(h).
The combination of these additional elements is no more than mere instructions to apply the exception using generic computer components. Accordingly, even in combination, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea.
Hence, the additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. Accordingly, the claims are directed to an abstract idea (Step 2A Prong Two: NO).
Step 2B
The claims 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, using the additional elements to perform the abstract idea amounts to no more than mere instructions to apply the exception using generic components. Mere instructions to apply an exception using a generic components cannot provide an inventive concept. See MPEP 2106.05(f).
Further, the claimed additional elements, identified above, are not sufficient to amount to significantly more than the judicial exception because they are generic components that are not integrated into the claim because they are merely incidental or token additions to the claim that do not alter or affect how the process steps or functions in the abstract idea are performed. Therefore, the claimed additional elements do not add meaningful limitations to the indicated claims beyond a general linking to a technological environment. See: MPEP 2106.05(h).
Further, the claimed additional elements, identified above, are not sufficient to amount to significantly more than the judicial exception because they are generic components that are configured to perform well-understood, routine, and conventional activities previously known to the industry. See: MPEP 2106.05(d). Said additional elements are recited at a high level of generality and provide conventional functions that do not add meaningful limits to practicing the abstract idea. The published specification supports this conclusion as follows:
[0050] Skilled persons will appreciate that trained multioutput neural network model 718 represents (i.e., models) the structure of multioutput neural network 100 after it has been trained such that each neuron is assigned appropriate weights. The term model in trained multioutput neural network model 718, therefore, would be understood to refer to multioutput neural network 100 that is trained and ultimately compiled in the form of a library or other instructions saved to a computer-readable medium and available for execution on a processor (see, e.g., FIG. 10). An example utility for modeling a trained neural network in source code is described in "Keras2c: A library for converting Keras neural networks to real-time compatible C," by Conlin et al. Other utilities may also be used for similar purposes. Accordingly, a trained neural network model may include any type of computer instruction or computer-executable code located within a memory device and configured to perform the function of the corresponding trained neural network. In this context, a model may, for instance, include one or more physical or logical blocks of computer instructions, which may be organized as a routine, program, object, component, data structure, etc., that perform the function of the corresponding trained neural network. Skilled persons will also appreciate that the model may be implemented in hardware or firmware instead of or in addition to a software application.
[0051] For completeness, mobile device software decision support application 708 also shows lower-layer OS components such as a network stack 724 for communication with cloud-based software application 710. For example, cloudbased software application 710 is configured to receive data from mobile device software decision support applications 708 through a secure internet connection 726. The data may then be stored in data storage 728 used to generate a data visualization 730 for display on user interface 714. In some embodiments, algorithms 716 are implemented in cloudbased software application 710, in which case mobile device software decision support application 708 simply provides data (i.e., input features 104) to cloud-based software application 710. In certain embodiments, trained multioutput neural network model 718 may include disparate instructions stored in different locations of a memory device, different memory devices, or different computers, which together implement the described functionality of the model. Indeed, a model may include a single instruction or many instructions, and may be distributed over several different code segments, among different programs, and across several memory devices. Some embodiments may be practiced in a distributed computing environment where tasks are performed by a remote processing device linked through a communications network. In a distributed computing environment, models may be located in local or remote memory storage devices. In addition, data being tied or rendered together in a database record may be resident in the same memory device, or across several memory devices, and may be linked together in fields of a record in a database across a network.
Viewing the limitations as an ordered combination, the claims simply instruct the additional elements to implement the concept described above in the identification of abstract idea with routine, conventional activity specified at a high level of generality in a particular technological environment.
Hence, the claims as a whole, considering the additional elements individually and as an ordered combination, do not amount to significantly more than the abstract idea (Step 2B: NO).
Dependent claim(s) 24 – 47 when analyzed as a whole, considering the additional elements individually and/or as an ordered combination, are held to be patent ineligible under 35 U.S.C. 101 because the additional recited limitation(s) fail(s) to establish that the claim(s) is/are not directed to an abstract idea without significantly more. These claims fail to remedy the deficiencies of their parent claims above, and are therefore rejected for at least the same rationale as applied to their parent claims above, and incorporated herein.
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.
Claim(s) 23, 25 – 42, 44, and 46 is/are rejected under 35 U.S.C. 103 as being unpatentable over Daniels et al., herein after Daniels (Daniels J, Herrero P, Georgiou P. A Deep Learning Framework for Automatic Meal Detection and Estimation in Artificial Pancreas Systems. Sensors. 2022; 22(2):466. https://doi.org/10.3390/s22020466) in view of Hei et al., herein after Hei (U.S. Publication Number 2021/0145370 A1).
Claim 23 (Original). Daniels teaches a system for meal detection (Abstract - we propose an algorithm based on a deep learning framework that performs multitask quantile regression, for both meal detection and carbohydrate estimation), comprising:
one or more wearable glucose sensing and regulating medical devices coupled to a user and configured to generate glucose data for a set of input features, the set of input features having glucoregulatory, insulin, and associated time of day features (Fig. 1; page 1, 3, 4 and 5 - The artificial pancreas is a system that comprises a continuous glucose monitor (CGM), Insulin pump, and an algorithm working in tandem to maintain blood glucose levels in an acceptable range (70 – 180 mg/dL); encoder LSTM input sequence - xenc = (xtl, ..., xt4) - comprises the glucose concentration levels from the CGM, insulin delivered; The estimated glucose level closely follows the reference glucose level from the CGM except during unannounced meals. In the instance of an unannounced meal—such as dinner just before 9 p.m.—a significant and persistent deviation In glucose trajectory from estimated trajectory occurs which leads to a meal detection and estimation);
a smart device configured to receive from the one or more wearable glucose sensing and regulating medical devices the glucose data and to provide the set of input features to a trained machine learning model, the trained machine leaning model including multiple connected layers (Fig. 1; page 1, 3, 4 and 5 - The artificial pancreas is a system that comprises a continuous glucose monitor (CGM), insulin pump, and an algorithm working In tandem to maintain blood glucose levels in an acceptable range (70 – 180 mg/dL); encoder LSTM input sequence - xenc = (xtl, ..., xt4) - comprises the glucose concentration levels from the CGM, insulin delivered; The estimated glucose level closely follows the reference glucose level from the CGM except during unannounced meals. In the Instance of an unannounced meal—such as dinner Just before 9 p.m.—a significant and persistent deviation In glucose trajectory from estimated trajectory occurs which leads to a meal detection and estimation); and
an output layer providing a meal detection output (page 7 - The first stage of the model involves the detection of a meal. First, the last 20 min of the glucose profile Is estimated using the sequence to sequence model. As explained In the previous section, the model outputs multiple quantiles for a 95% prediction interval coverage that arises from the errors arising from the input variables. Consequently, in a scenario where the glucose trajectory persists outside the prediction Interval, this Implies a significant deviation beyond noise In the input variables that can be inferred as a missing Input. In this setting, this missing Input Is attributed to an unannounced ingested meal and therefore a meal detection flag can be activated); and
a user interface configured to present to the user, based on the meal detection output, a meal detection indication (page 5 and 8- this missing input Is attributed to an unannounced ingested meal and therefore a meal detection flag can be activated; In the eventual use case of a meal detection algorithm, the detection of an unannounced meal would prompt a bolus to be delivered either indirectly by notifying the user with an alert, or directly In a sensor augmented pump.).
Daniels fails to explicitly teach the following limitations met by Hei as cited:
wirelessly receiving glucose data. (paragraphs 56 – 57 - correction to protect the glucose sensor and pressure sensor In the wireless Insulin pump system; (3) provision of physiological data-based life-threatening case like hypoglycemia checking).
It would have been obvious to one of ordinary skill before the effective filing date of the claimed invention to expand the method of Daniels to further include the glucose monitoring method and system which protects an artificial pancreas’s sensor, infusion system, and alert system from EMI/wireless attacks using a medical software or application, close the gap between sensor and blood glucose, and build a non-invasive hypoglycemia and hyperglycemia false alarm detection scheme with the help of a wristband to provide accurate blood glucose predictions as disclosed by Hei.
One of ordinary skill in the art, before the effective filing date of the claimed invention, would have been motivated to expand the method of Daniels in this way to detect/predict the real hypoglycemia and hyperglycemia cases by monitoring the patient's physiological data such as heart rate, galvanic skin response, food eaten, and exercise steps to help detect the sensor glucose alert manipulating attacks without missing life-threatening cases. Preprocessing the CGM readings with Kalman smoothing for sensor error correction improves the robustness of the BG prediction (Hei: paragraph 5).
Claim 25 (Original). Daniels and Hei teach the system of claim 23. Daniels teaches the system in which the trained machine learning model is a trained multioutput machine learning model with the output layer formed from first and second branches, the first branch providing the meal detection output and the second branch providing a carbohydrate estimation output (page 3, 5-6 - the deep neural network is based on a multitask sequence-to-sequence model. The sequence-to-sequence (seq2seq) model is primarily a model that is used to map one set of input sequences to an associated set of output sequences and has been used In glucose prediction tasks; The first stage of the model involves the detection of a meal; Carbohydrate Estimation).
Claim 26 (Original). Daniels and Hei teach the system of claim 25. Daniels teaches a system in which the user interface is configured to present to the user, based on the carbohydrate estimation output, a meal size estimation (page 6 and 7- estimating the meal size; output meal size).
Claim 27 (Original). Daniels and Hei teach the system of claim 26. Daniels teaches a system in which the trained machine learning model provides a probability estimate of the meal size estimation (page 3 - develop a probabilistic approach that compares the expected signal and observed signal to detect a meal).
Claim 28 (Original). Daniels and Hei teach the system of claim 26. Daniels teaches a system to determine, based on one or both the meal detection output and the carbohydrate estimation output, an amount of Insulin to dose to a person requiring exogenous insulin delivery (page 6 - the meal detection and estimation module provides an estimate of the carbohydrate size to the bolus calculator In the BIAP controller to determine the meal Insulin bolus.).
Daniels fails to explicitly teach the following limitations met by Hei as cited:
smart device (Figure 1; paragraph 29 discloses an insulin pump, a computer, or a smartphone as main components of an artificial pancreas; paragraph 54 discloses smartphones offer the ability to integrate an applicable app into the system and method).
The motivation to combine Daniels and Hei is discussed in the rejection of claim 23, and incorporated herein.
Claim 29 (Original). Daniels and Hei teach the system of claim 26. Daniels teaches a system in which the meal size estimation is used within a decision support application executed by the smart device to estimate whether a meal was consumed prior to meal insulin dosing (page 3, 5-8).
Claim 30 (Original). Daniels and Hei teach the system of claim 26. Daniels teaches a system in which the smart device is configured to provide the meal size estimation to a weight-loss coaching application (page 6 and 7- estimating the meal size; output meal size). Although Daniels does not explicitly disclose weight loss coaching, it Is well known in the art the use of meal data for weight loss programs.
Claim 31 (Original). Daniels and Hei teach the system of claim 23. Daniels teaches a system in which the multiple connected layers are fully connected layers (para [0124], [0133]- fully connected layers In the network).
Claim 32 (Original). Daniels and Hei teach the system of claim 23. Daniels teaches a system in which the trained machine learning model is one or more of a neural network, a random forest model, a support vector regression model, and a logistic regression model (page 2 - the neural network architecture and proposed framework for detecting and estimating unannounced meals).
Claim 33 (Original). Daniels and Hei teach the system of claim 23. Daniels teaches a system in which the trained machine learning model is trained using one or more of an ordinary differential equation (ODE) in silico model of glucose metabolism and real-world human glucose, insulin, and nutrition data (page 5 and 8- we first train a generalized model using aggregated data from each individual. The individualized model for each participant Is then obtained by finetuning the generalized model on the individual data; T1D Simulator Is used to generate a challenging scenario for training, validating and testing the models).
Claim 34 (Original). Daniels and Hei teach the system of claim 23. Daniels teaches a system in which the trained machine learning model is configured to predict categories of meal sizes or an actual meal amount (page 6 - precise meal size estimate).
Claim 35 (Original). Daniels and Hei teach the system of claim 23. Daniels teaches a system in which the trained machine learning model provides a probability estimate of a likelihood of a meal having occurred (page 3 - develop a probabilistic approach that compares the expected signal and observed signal to detect a meal).
Claim 36 (Original). Daniels and Hei teach the system of claim 23. Daniels teaches a system configured to:
receive a series of periodic glucose measurement samples and deriving one or more glucoregulatory features based on the series of glucose measurement samples (page 3 -sequence-to-sequence (seq2seq) model is primarily a model that is used to map one set of Input sequences to an associated set of output sequences and has been used in glucose prediction tasks. In this task, the objective Is to estimate the last 20 min of the individual's glucose trajectory using historical CGM measures, meals, and insulin).
Daniels fails to explicitly teach the following limitations met by Hei as cited:
smart device (Figure 1; paragraph 29 discloses an insulin pump, a computer, or a smartphone as main components of an artificial pancreas; paragraph 54 discloses smartphones offer the ability to integrate an applicable app into the system and method).
The motivation to combine Daniels and Hei is discussed in the rejection of claim 23, and incorporated herein.
Claim 37 (Original). Daniels and Hei teach the system of claim 23. Daniels teaches a system configured to:
receive insulin bolus data and calculate one or more insulin features based on a weighted sum of amounts in the insulin bolus data over a predetermined time (page 3 and 6 - decoder generates the output sequence— ydec = (yt3, ..., yt)—and estimated quantiles of the glucose trajectory are generated at the output. The decoder input, xdec = (xt3, ..., xt), comprises the insulin delivered and announced/estimated meals and is initialized with the final encoder state; the bolus calculator In the BIAP controller to determine the meal Insulin bolus. However, because of the increase in glucose levels due to the meal ingestion, the controller begins to deliver Insulin to cover the meal. This insulin Is delivered as the deviation remains below the upper bound and the meal flag is not raised. In addition the meal size estimate may be overestimated).
Daniels fails to explicitly teach the following limitations met by Hei as cited:
smart device (Figure 1; paragraph 29 discloses an insulin pump, a computer, or a smartphone as main components of an artificial pancreas; paragraph 54 discloses smartphones offer the ability to integrate an applicable app into the system and method).
The motivation to combine Daniels and Hei is discussed in the rejection of claim 23, and incorporated herein.
Claim 38 (Original). Daniels and Hei teach the system of claim 23. Daniels teaches a system in which the user interface is configured to: notify the user of a meal detection event and receive a user confirmation of the meal detection event (page 8 - the detection of an unannounced meal would prompt a bolus to be delivered either indirectly by notifying the user with an alert).
Claim 39 (Original). Daniels and Hei teach the system of claim 38.
Daniels fails to explicitly teach the following limitations met by Hei as cited:
the smart device (Figure 1; paragraph 29 discloses an insulin pump, a computer, or a smartphone as main components of an artificial pancreas; paragraph 54 discloses smartphones offer the ability to integrate an applicable app into the system and method) is configured to initiate delivery of a fraction of a requisite amount of meal insulin to the user automatically or in response to reception of the user confirmation (paragraphs 100 – 101 and 135 disclose along with other features including carbohydrates from the meal, bolus Insulin, and cumulative step counts In a fixed time interval. Consider the previous 50 minutes of history for calculating the cumulative step numbers. For every five minutes interval, it is denoted with Rinsulin=5*0.014. Calculate the effective Insulin on the body leff at any particular time-index ts with the equation as follows: Ieff(ts)=Ibolus - (ts - tbolus)Rinsulin).
The motivation to combine Daniels and Hei is discussed in the rejection of claim 23, and incorporated herein.
Claim 40 (Original). Daniels and Hei teach the system of claim 23.
Daniels fails to explicitly teach the following limitations met by Hei as cited:
the smart device (Figure 1; paragraph 29 discloses an insulin pump, a computer, or a smartphone as main components of an artificial pancreas; paragraph 54 discloses smartphones offer the ability to integrate an applicable app into the system and method) is configured to determine the fraction as a function of time based on a time after the user confirmation (paragraphs 100 – 101 and 135 disclose along with other features including carbohydrates from the meal, bolus Insulin, and cumulative step counts In a fixed time interval. Consider the previous 50 minutes of history for calculating the cumulative step numbers. For every five minutes interval, it is denoted with Rinsulin=5*0.014. Calculate the effective Insulin on the body leff at any particular time-index ts with the equation as follows: Ieff(ts)=Ibolus - (ts - tbolus)Rinsulin).
The motivation to combine Daniels and Hei is discussed in the rejection of claim 23, and incorporated herein.
Claim 41 (Original). Daniels and Hei teach the system of claim 23. Daniels teaches a system in which the one or more wearable glucose sensing and regulating medical devices include a continuous glucose monitoring (CGM) device (page 1 — continuous glucose monitor (CGM), insulin pump).
Claim 42 (Original). Daniels and Hei teach the system of claim 23. Daniels teaches a system in which the one or more wearable glucose sensing and regulating medical devices include an Insulin pen or an insulin pump (page 1 — continuous glucose monitor (CGM), Insulin pump).
Claim 44 (Original). Daniels and Hei teach the system of claim 23. Daniels teaches a system is configured to initiate delivery of Insulin to a user in response to the meal size estimation (page 8).
Daniels fails to explicitly teach the following limitations met by Hei as cited:
the smart device (Figure 1; paragraph 29 discloses an insulin pump, a computer, or a smartphone as main components of an artificial pancreas; paragraph 54 discloses smartphones offer the ability to integrate an applicable app into the system and method).
The motivation to combine Daniels and Hei is discussed in the rejection of claim 23, and incorporated herein.
Claim 46 (New). Daniels and Hei teach the system of claim 23. Daniels discloses a system in which the glucoregulatory feature comprises a rate-of-change of glucose (page 3 discloses a threshold-based detection using the rate of change (ROC) of glucose levels and (ii) outlier detection using model predictions).
Claim(s) 24 and 43 is/are rejected under 35 U.S.C. 103 as being unpatentable over Daniels et al., herein after Daniels (Daniels J, Herrero P, Georgiou P. A Deep Learning Framework for Automatic Meal Detection and Estimation in Artificial Pancreas Systems. Sensors. 2022; 22(2):466. https://doi.org/10.3390/s22020466) in view of Hei et al., herein after Hei (U.S. Publication Number 2021/0145370 A1) further in view of Nelson et al., herein after Nelson (U.S. Publication Number 2020/0163549 A1).
Claim 24. Daniels and Hei teach the system of claim 23.
Daniels and Hei fails to explicitly teach the following limitations met by Nelson as cited:
further comprising a remote server for hosting the trained machine learning model and receiving the set of input features to the trained machine learning model (paragraph 8 discloses a server operatively coupled to the data network; paragraph 31 discloses the server device can be another portable device, such as a personal digital assistant or a notebook computer; paragraph 32 discloses the server device can also communicate with another device including a PDA synching data with a personal computer (PC), a mobile phone communicating over a cellular network with a computer at the other end, or a household appliance communicating with a computer system at a physician's office, indicating remotely).
It would have been obvious to one of ordinary skill before the effective filing date of the claimed invention to expand the method of Daniels and Hei to further include a device and method for determining and reporting glucose readings in wireless personal area networks for diabetics as disclosed by Nelson.
One of ordinary skill in the art, before the effective filing date of the claimed invention, would have been motivated to expand the method of Daniels and Hei in this way to provide a low cost meter and strip glucose monitoring system that has highly convenient features, including wireless communication capabilities (Nelson: paragraph 6).
Claim 43. Daniels and Hei teach the system of claim 23.
Daniels and Hei fails to explicitly teach the following limitations met by Nelson as cited:
the smart device is configured to wirelessly receive, via a wireless personal area network from the one or more wearable glucose sensing and regulating medical devices, the glucose data including at least a portion of the set of input features (paragraphs 13 and 32 - data network may Include a personal area network, where the personal area network is configured for short range wireless communication; sending glucose data from devices 102 and 104 to data storage in device 105).
The motivation to combine the teachings of Daniels, Hei, and Nelson is discussed in the rejection of claim 24, and incorporated herein.
Claim(s) 45 and 47 is/are rejected under 35 U.S.C. 103 as being unpatentable over Daniels et al., herein after Daniels (Daniels J, Herrero P, Georgiou P. A Deep Learning Framework for Automatic Meal Detection and Estimation in Artificial Pancreas Systems. Sensors. 2022; 22(2):466. https://doi.org/10.3390/s22020466) in view of Hei et al., herein after Hei (U.S. Publication Number 2021/0145370 A1) further in view of Hayter et al., herein after Hayter (U.S. Publication Number 2021/0050085 A1).
Claim 45 (New). Daniels and Hei teach the system of claim 23.
Daniels and Hei fails to explicitly teach the following limitations met by Hayter as cited:
in which the insulin feature models insulin decay over time using a predefined decay function (paragraph 305 discloses rapid acting insulin analogs generally achieve peak plasma concentration in approximately 45 minutes, with a profile that decays exponentially afterward, though this time may be shorter if a user is taking ultra-rapid insulin analogs. This time window can fall within the dormant period where the DGA cannot provide guidance due to rising glucose levels. IOB can be estimated from this profile by directly measuring the exponential decay or by estimating a linear decay from the peak insulin concentration to the pre-prandial value. The IOB value can account for the current decline in glucose).
It would have been obvious to one of ordinary skill before the effective filing date of the claimed invention to expand the method of Daniels and Hei to further include a systems, devices, and methods relating to medication dose guidance such as, for example, the determination of an insulin dose for the treatment of elevated glucose levels resulting from diabetes as disclosed by Hayter.
One of ordinary skill in the art, before the effective filing date of the claimed invention, would have been motivated to expand the method of Daniels and Hei in this way to provide systems, devices, or methods that can automatically utilize glucose information collected by an analyte monitoring system to provide medication dose guidance in a readily accessible manner on an as-needed basis (Hayter: paragraph 5).
Claim 47 (New). Daniels and Hei teach the system of claim 23.
Daniels and Hei fails to explicitly teach the following limitations met by Hayter as cited:
in which the meal detection output comprises a binary classification indicating presence or absence of a meal (paragraph 148 discloses the retrospective meal detection module can take the feature matrix, as input, and output binary detection results for each rising region. Such output can include: a binary classification result, and a probability value of each rising region being an analyte (e.g., glucose) excursion in response to a meal event; paragraph 149 discloses the classification operation can take the feature matrix of a patient as input and output a binary classification result for each relevant medication event (e.g., for each insulin injection). For example, the DGA can output binary data '1' signifying a meal dose and 'O' signifying non-meal dose; paragraph 270 discloses the meal detection module can also be configured to output a probability value with the binary detection result).
The motivation to combine the teachings of Daniels, Hei, and Hayter is discussed in the rejection of claim 45 and incorporated herein.
Response to Arguments
Applicant's arguments filed February 17, 2026 have been fully considered but they are not persuasive. The Applicant’s arguments have been addressed in the order in which they were presented.
Response to Rejection under 35 USC § 101
The Applicant argues claim 23 is not directed to Organizing Human Activity under Step 2A, Prong One. The Examiner disagrees. Under its broadest reasonable interpretation, the Applicant’s claims are an abstract idea that falls into the grouping of “Certain Methods of Organizing Human Activity” which covers fundamental economic principles or practices, commercial or legal interactions, or managing personal behavior or relationships or interactions between people. The Examiner respectfully submits that section 2106.04(a) of the MPEP recites that “Certain Methods of Organizing Human Activity” include managing personal behavior or relationships or interactions between people, including social activities, teaching, and following rules or instructions. The present claims recite the abstract idea of generating glucose data for a set of input features, the set of input features having glucoregulatory, insulin, and associated time of day features; receive the glucose data; and present to the user, based on the meal detection output, a meal detection indication. These features describe interactions with people, thus “Certain Methods of Organizing Human Activity” as the data generated is presented to the user. Thus, if a claim limitation, under its broadest reasonable interpretation, covers interactions with people, but for the recitation of generic components, then it is still in the “Certain Methods of Organizing Human Activity” grouping.
The Applicant argues claim 23 integrates any exception into a practical application under Step 2A, Prong Two. The Examiner respectfully disagrees. The additional elements of the present claims fail to integrate the exception into a practical application of the exception. Section 2106.04(d) of the MPEP defines the phrase “integration into a practical application” to require an additional element or a combination of additional elements in the claim to apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception, such that it is more than a drafting effort designed to monopolize the exception. For example, section 2106.04(d) of the MPEP recite limitations that are indicative of integration into a practical application when recited in a claim with a judicial exception include:
Improvements to the functioning of a computer, or to any other technology or technical field, as discussed in MPEP 2106.05(a);
Applying or using a judicial exception to effect a particular treatment or prophylaxis for disease or medical condition – see Vanda Memo
Applying the judicial exception with, or by use of, a particular machine, as discussed in MPEP 2106.05(b);
Effecting a transformation or reduction of a particular article to a different state or thing, as discussed in MPEP 2106.05(c); and
Applying or using the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is more than a drafting effort designed to monopolize the exception, as discussed in MPEP 2106.05(e) and the Vanda Memo issued in June 2018.
The present claims fail to demonstrate an improvement to the functioning of a computer or to any other technology or technical field. Thus, Applicant’s argument is not persuasive, and the rejection is maintained.
The Applicant argues the Office has not established that the claimed ordered combination is well-understood, routine, and conventional. The Berkheimer memo discloses “a citation to an express statement in the specification or to a statement made by applicant during prosecution that demonstrates the well-understood, routine, conventional nature of additional elements.” The Examiner submits the Applicant’s own specification supports this conclusion and recites:
[0050] Skilled persons will appreciate that trained multioutput neural network model 718 represents (i.e., models) the structure of multioutput neural network 100 after it has been trained such that each neuron is assigned appropriate weights. The term model in trained multioutput neural network model 718, therefore, would be understood to refer to multioutput neural network 100 that is trained and ultimately compiled in the form of a library or other instructions saved to a computer-readable medium and available for execution on a processor (see, e.g., FIG. 10). An example utility for modeling a trained neural network in source code is described in "Keras2c: A library for converting Keras neural networks to real-time compatible C," by Conlin et al. Other utilities may also be used for similar purposes. Accordingly, a trained neural network model may include any type of computer instruction or computer-executable code located within a memory device and configured to perform the function of the corresponding trained neural network. In this context, a model may, for instance, include one or more physical or logical blocks of computer instructions, which may be organized as a routine, program, object, component, data structure, etc., that perform the function of the corresponding trained neural network. Skilled persons will also appreciate that the model may be implemented in hardware or firmware instead of or in addition to a software application.
[0051] For completeness, mobile device software decision support application 708 also shows lower-layer OS components such as a network stack 724 for communication with cloud-based software application 710. For example, cloud based software application 710 is configured to receive data from mobile device software decision support applications 708 through a secure internet connection 726. The data may then be stored in data storage 728 used to generate a data visualization 730 for display on user interface 714. In some embodiments, algorithms 716 are implemented in cloud based software application 710, in which case mobile device software decision support application 708 simply provides data (i.e., input features 104) to cloud-based software application 710. In certain embodiments, trained multioutput neural network model 718 may include disparate instructions stored in different locations of a memory device, different memory devices, or different computers, which together implement the described functionality of the model. Indeed, a model may include a single instruction or many instructions, and may be distributed over several different code segments, among different programs, and across several memory devices. Some embodiments may be practiced in a distributed computing environment where tasks are performed by a remote processing device linked through a communications network. In a distributed computing environment, models may be located in local or remote memory storage devices. In addition, data being tied or rendered together in a database record may be resident in the same memory device, or across several memory devices, and may be linked together in fields of a record in a database across a network.
The Examiner submits the Berkheimer memo requires factual support on the record, which is shown in the 101 rejection above, as well as in the response to the Applicant’s 101 argument. The Applicant’s published specification supports the elements in the claims are “well-understood, routine, and conventional”. The Berkheimer memo recites “In a step 2B analysis, an additional element (or combination of elements) is not well-understood, routine or conventional unless the examiner finds, and expressly supports a rejection in writing with, one or more of the following: 1. A citation to an express statement in the specification or to a statement made by an applicant during prosecution that demonstrates the well-understood, routine, conventional nature of the additional element(s). A specification demonstrates the well-understood, routine, conventional nature of additional elements when it describes the additional elements as well-understood or routine or conventional (or an equivalent term), as a commercially available product, or in a manner that indicates that the additional elements are sufficiently well-known that the specification does not need to describe the particulars of such additional elements to satisfy 35 U.S.C. § 112(a). A finding that an element is well-understood, routine, or conventional cannot be based only on the fact that the specification is silent with respect to describing such element.” The Examiner has cited multiple paragraphs in the Applicant’s published specification (see above) which are directed to well-understood, routine or conventional elements. Thus, the Applicant’s argument is not persuasive and the rejection is maintained.
The Applicant argues the additional elements are not mere instructions to apply the exception. The Examiner respectfully disagrees. The Applicant’s specification states “Skilled persons will appreciate that trained multioutput neural network model 718 represents (i.e., models) the structure of multioutput neural network 100 after it has been trained such that each neuron is assigned appropriate weights. The term model in trained multioutput neural network model 718, therefore, would be understood to refer to multioutput neural network 100 that is trained and ultimately compiled in the form of a library or other instructions saved to a computer-readable medium and available for execution on a processor (see, e.g., FIG. 10). An example utility for modeling a trained neural network in source code is described in "Keras2c: A library for converting Keras neural networks to real-time compatible C," by Conlin et al. Other utilities may also be used for similar purposes. Accordingly, a trained neural network model may include any type of computer instruction or computer-executable code located within a memory device and configured to perform the function of the corresponding trained neural network. In this context, a model may, for instance, include one or more physical or logical blocks of computer instructions, which may be organized as a routine, program, object, component, data structure, etc., that perform the function of the corresponding trained neural network. Skilled persons will also appreciate that the model may be implemented in hardware or firmware instead of or in addition to a software application.” (paragraph 50 of the published specification). The generic computer cited by the Applicant is a general link to execute the abstract idea. Mere instructions to apply an exception cannot provide an inventive concept. Thus, Applicant’s argument is not persuasive and the rejection is maintained.
The Applicant argues, viewed as an ordered combination, the claim amounts to significantly more. The Examiner respectfully disagrees. The Applicant’s specification states “Skilled persons will appreciate that trained multioutput neural network model 718 represents (i.e., models) the structure of multioutput neural network 100 after it has been trained such that each neuron is assigned appropriate weights. The term model in trained multioutput neural network model 718, therefore, would be understood to refer to multioutput neural network 100 that is trained and ultimately compiled in the form of a library or other instructions saved to a computer-readable medium and available for execution on a processor (see, e.g., FIG. 10). An example utility for modeling a trained neural network in source code is described in "Keras2c: A library for converting Keras neural networks to real-time compatible C," by Conlin et al. Other utilities may also be used for similar purposes. Accordingly, a trained neural network model may include any type of computer instruction or computer-executable code located within a memory device and configured to perform the function of the corresponding trained neural network. In this context, a model may, for instance, include one or more physical or logical blocks of computer instructions, which may be organized as a routine, program, object, component, data structure, etc., that perform the function of the corresponding trained neural network. Skilled persons will also appreciate that the model may be implemented in hardware or firmware instead of or in addition to a software application.” (paragraph 50 of the published specification). The generic computer cited by the Applicant is a general link to execute the abstract idea. Mere instructions to apply an exception cannot provide an inventive concept. Thus, Applicant’s argument is not persuasive and the rejection is maintained.
Response to Rejection under 35 USC § 103
The Applicant argues the Office has not shown that Daniels or Hei describe an output layer providing a meal detection output. The Examiner respectfully disagrees. The Examiner submits Daniels discloses a model that is used to map one set of input sequences to an associated set of output sequences and has been used in glucose prediction tasks where the objective is to estimate the last 20 min of the individual's glucose trajectory using historical CGM measures, meals, and insulin (page 4). Daniels discloses the first stage of the model involves the detection of a meal where the model outputs multiple quantiles for a 95% prediction interval coverage that arises from the errors arising from the input variables; If the glucose trajectory persists outside the prediction Interval, this Implies a significant deviation beyond noise In the input variables that can be inferred as a missing input where the missing input Is attributed to an unannounced ingested meal and therefore a meal detection flag can be activated (page 7). Thus, Applicant’s argument is not persuasive and the rejection is maintained.
The Applicant argues the Office has not shown Daniels or Hei describe time of day features. The Examiner respectfully disagrees. The Examiner submits Daniels discloses a meal protocol scenario including four meals with the following average carbohydrate size at the associated average meal times: 70 g (7 a.m.), 100 g (1 p.m.), 30 g (5 p.m.), and 80 g (8 p.m.). A meal-time variability (aT = 60 min) and meal size variability (CV= 10%) is introduced in order to generate realistic scenario of inter-day variability in meals (page 12), indicating various times of day for meals. In addition, Table 1 (page 13) discloses meal detection for different meal times. Thus, Applicant’s argument is not persuasive and the rejection is maintained.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KRISTINE K RAPILLO whose telephone number is (571)270-3325. The examiner can normally be reached Monday - Friday 7:30 - 4 pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Fonya Long can be reached at 571-270-5096. 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.
/K.K.R/Examiner, Art Unit 3682 /ROBERT A SOREY/Primary Examiner, Art Unit 3682