Notice of Pre-AIA or AIA Status
1. 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
2. This Final Office Action is responsive to Applicants’ amendments and arguments, as received 10/16/25. Claims 1-6, 8, 10, 12, 14, 16, 23, 25, 27-28, 30, 32, 34, 36, and 38 were pending. Claim 16 has now been cancelled by way of the recent reply. Hence, claims 1-6, 8, 10, 12, 14, 23, 25, 27-28, 30, 32, 34, 36, and 38 are now presently pending, of which claims 1 and 14 are independent.
3. The claim rejections under 35 U.S.C. 101 and 35 U.S.C. 112(b), previously presented in the Non-Final Office Action dated 6/16/25, are withdrawn, in view of Applicants’ amendments and arguments.
Claim Rejections - 35 USC § 103
4. 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.
5. 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.
6. 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.
7. Claims 1-6, 8, 10, 12, 14, 23, 25, 27-28, 30, 32, 36, and 38 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Patent Application Publication No. 2019/0188587 (“Gupta”) in view of U.S. Patent Application Publication No. 2015/0250429 (“Hampapuram”) and further in view of Non-Patent Literature “Gradual DropIn of Layers to Train Very Deep Neural Networks” (“Smith”).
Regarding claim 1, GUPTA teaches a method of predicting an analyte level in a biological liquid of a subject (Gupta’s [0002]-[0004]: predicting glucose level in blood, and [0021] establishing that “The disclosed approach is applied to the closed loop blood glucose control system. Using this disclosed approach the blood glucose level prediction accuracy is improved 100 times with respect to simply using the Bergman Minimal model”, and FIG. 1’s architecture is described per [0027] to teach “... an overall architecture for model guided deep learning using a predictive physiological model as a guide, as applied to the prediction of blood glucose levels in T1DM patients”), the method comprising:
receiving a time-ordered series of levels of the analyte, monitored over a time-period (Gupta’s [0042] discussing the collection of blood glucose data from patients, e.g., specifying a collection period and a collection frequency);
feeding a trained neural network procedure with said monitored levels (Gupta’s FIG. 1 element 130, receiving the raw input data as illustrated, the input data as discussed just above per [0042]); and
predicting, based on an output received from said procedure, a predicted level of the analyte ..., where said predicted level of the analyte is an estimation of said level in a future time (Gupta’s [0027]-[0028]: “... an overall architecture for model guided deep learning using a predictive physiological model as a guide, as applied to the prediction of blood glucose levels in T1DM patients. .... The system 100 may be utilized to predict future blood glucose levels. In such embodiments, the raw input data 110 may include data that represents a patient's measured plasma glucose. Plasma insulin and interstitial insulin in the blood may also be included in the raw data input 110.”, and where the prediction as mentioned is attributable to processing by the neural network as discussed per [0029]); and
wherein said procedure comprises a plurality of layers ... (the neural network per Gupta’s FIG. 1 and as discussed just above, which is understood to be comprised of various layers per [0022]-[0023] and for example more generally stated per [0009] and [0012]).
As discussed above, Gupta teaches the prediction of a predicted level of analyte level in a future time. While Gupta does teach computer elements having display features, see e.g., [0059]: element 680, Gupta does not teach specifically that they are used to display the prediction as generated, e.g. per Applicants’ further limitation for displaying ... the predicted level of the analyte in the future time and responsively to the predicted level, then treating the subject by a treatment specifically selected for reducing or increasing the level of the analyte in the body liquid. Rather, the Examiner relies upon HAMPAPURAM to teach what Gupta otherwise lacks:
Hampapuram’s [0037] discussing “... display devices 14-20 and the like ... and generate reports providing high-level information, such as statistics, regarding the measured analyte over a certain time frame”, which the Examiner believes reads on the limitation for displaying...
Hampapuram’s [0009] discussing “In an embodiment of the second aspect, providing output comprises transmitting a message to an insulin delivery device including instructions associated with at least one of: a) suspending insulin delivery, b) initiating a hypoglycemia and/or hyperglycemia minimizer algorithm, c) controlling insulin delivery or d) information associated with the hypoglycemia indicator.”, which the Examiner believes reads on the limitation for treating....
The Examiner believes Gupta also teaches a similar aspect of regulating dosage and/or levels in conjunction with the results of the deep learning via an artificial pancreas. [0029], for example: “... the second estimated future blood glucose levels output from the deep neural network 130 may be utilized to determine input to a closed loop control system, such as an artificial pancreas to determine parameters for controlling dosage of a patient's insulin or other substances”
Both Gupta and Hampapuram relate to glucose monitoring in a patient’s blood and principally involve monitoring and processing to deliver prediction and prediction-related benefits/advantages. Hence, they are similarly directed and therefore analogous. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Hampapuram’s explicit and overt display aspect into a framework such as Gupta’s, per Gupta’s contemplated device/environment teachings which are amenable to display/presentation aspects, with a reasonable expectation of success, such that the results of a prediction can be communicated, presented, and otherwise displayed for the benefit and advantage of patients, caregivers, and any other invested persons responsible for patient care in the instant cases.
Applicants’ claims additional recite the further limitation wherein for at least one pair of layers, a number of inter-layer connections within said pair is higher for later monitored levels than for earlier monitored levels. Gupta does not teach this, and hence the Examiner relies upon SMITH to teach what Gupta otherwise lacks, see e.g., Smith’s pages 1-2, in the Abstract, Introduction, and Developing new architecture sections, teaching a gradual training framework to implement a neural network where some layers are slowly and organically added to a developing network under training, such that the inclusion of the new layer includes an increase in connections for the layer with its adjacent layers, e.g., to make them more expressive through learning by way of these new connections.
Both Gupta and Smith relate to neural networks and their training and their use. Hence, they are sufficiently related and therefore analogous. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to train a neural network as used per Gupta in the manner taught per Smith, with a reasonable expectation of success, such that expressivity through learning/training is organically developed based on the training data provided.
Regarding claim 2, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 1, as discussed above. The aforementioned references further teach the additional limitations wherein the analyte comprises glucose (Gupta’s [0027]: “FIG. 1 represents an overall architecture for model guided deep learning ... as applied to the prediction of blood glucose levels.”). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 3, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 1, as discussed above. The aforementioned references further teach the additional limitations comprising administering a drug to the biological liquid during said time-period, and feeding said procedure with a time-ordered series of dose levels of said drug (Gupta’s [0028]: “The system 100 may be utilized to predict future blood glucose levels. In such embodiments, the raw input data 110 may include data that represents a patient's measured plasma glucose. Plasma insulin and interstitial insulin in the blood may also be included in the raw data input 110”, and feasibly these features as taught are involved in implementing and otherwise controlling [0029]’s artificial pancreas such that the patient’s dosage of insulin can be controlled, and see also Gupta’s [0029] and Hampapuram’s [0009] as specifically cited and discussed in relation to claim 1’s treating... limitation). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 4, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 1, as discussed above. The aforementioned references further teach the additional limitations wherein said time-period is selected such that said dose levels include basal dose levels but not bolus dose levels (Hampapuram’s [0132] and [0168] providing configurable capability to suspend basal or bolus delivery of glucose as part of an insulin delivery framework, and where the configurability of the feature as taught permits a basal dose but not bolus dose). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 5, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 3, as discussed above. The aforementioned references further teach the additional limitations wherein said future time is before administration of a bolus dose level of said drug (Hampapuram’s [0132] and [0168] providing configurable capability to suspend basal or bolus delivery of glucose as part of an insulin delivery framework, and where the configurability of the feature as taught permits a flexibility of operation before a basal dose for example). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 6, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 3, as discussed above. The aforementioned references further teach the additional limitations wherein the analyte comprises glucose and the drug comprises insulin (glucose as measured and monitored in the blood per Gupta’s [0027]-[0028], where Gupta’s implementation can be used with an artificial pancreas per [0029] to provide patients with a regulated dose of insulin; and see also insulin as delivered medically per Hampapuram’s [0127]). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 8, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 1, as discussed above. The aforementioned references further teach the additional limitations wherein said time-ordered series is characterized by a frequency of at least 6 analyte levels per hour (Gupta’s [0042] discussing measuring glucose every five minutes, which would exceed 6 times an hour as recited). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 10, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 1, as discussed above. The aforementioned references further teach the additional limitations wherein said time-ordered series is characterized by a frequency of less than four analyte levels per hour, and the method comprises interpolating said time-ordered series to provide a plurality of interpolated analyte levels, and updating said time-ordered series using said interpolated analyte levels (Hampapuram’s [0072] teaching a trend graph which can be used to discern levels across various time ranges, i.e., interpolating beyond any frequency/interval that Gupta alone might implement). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 12, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 1, as discussed above. The aforementioned references further teach the additional limitations for A computer software product, comprising a computer-readable medium in which program instructions are stored, which instructions, when read by a data processor, cause the data processor to receive a time-ordered monitored levels of the analyte over a time-period and to execute the method according to claim 1 (Hampapuram’s FIG. 2, teaching a computing system capable of implementing the comparable-to-Gupta framework, the computing system having program memory and a processor module which can be leveraged to implement Gupta’s modified framework). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 14, the claim includes the same or similar limitations as claim 1 discussed above. The system of the present claim includes a monitoring device (Gupta’s framework can be read to encompass medical/glucose monitoring, e.g., per [0009]-[0010] and [0029]’s mention of an artificial pancreas; however, Hampapuram is even more explicit, see e.g., its FIG. 1 including a host equipped with elements 4 and 10), a data processor (Hampapuram’s FIG. 2: element 214), and a communication device (Gupta’s [0068] discussing a framing for its framework in terms of a client and server that communicate over networks, thereby implying the need for communication devices so that it could function as intended; however, more explicitly see Hampapuram’s FIG. 2: element 238), which are further taught.
Regarding claim 23, the claim includes the same or limitations as claim 10 discussed above, and is therefore rejected under the same rationale.
Regarding claim 25, Gupta in view of Hampapuram and further in view of Smith teaches the system according to claim 14, as discussed above. The aforementioned references further teach the additional limitations wherein said communication device communicates wirelessly with said data processor (Gupta’s [0069] teaching an implementation assuming a client-server configuration for its framework, where a remote situation is explicit and the communication therefor would be understood to be wireless or feasibly wireless, such as leveraging a communication device as shown per Hampapuram’s FIG. 2 element 238 for example). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 27, Gupta in view of Hampapuram and further in view of Smith teaches the system according to claim 25, as discussed above. The aforementioned references further teach the additional limitations wherein said data processor is a component of a server computer, and is configured to transmit said predicted level of the analyte to a mobile device having a display for displaying said predicted level of the analyte on said display (Gupta’s [0069] teaching an implementation assuming a client-server configuration for its framework, and Gupta is clear that its client device would have display capabilities, e.g., per [0059] and [0064], and could be simply modified to accommodate the display step that Hampapuram teaches for a same/similar result (see Hampapuram’s [0037] discussing “... display devices 14-20 and the like ... and generate reports providing high-level information, such as statistics, regarding the measured analyte over a certain time frame.”)). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 28, the claim includes the same or limitations as claim 27 discussed above, and is therefore rejected under the same rationale.
Regarding claim 30, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 1, as discussed above. The aforementioned references further teach the additional limitations wherein said time-period is from about 1 hour to about 6 hours (Gupta teaching specific time ranges per [0042] for collection, which can be subject to modification in view of Hampapuram’s [0072] (teaching time ranges: “Although FIG. 4A illustrates a 1-hour trend graph (e.g., depicted with a time range 188 of 1-hour), a variety of time ranges can be represented on the screen 226, for example, 3-hour, 9-hour, 1-day, and the like.”)). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 32, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 1, as discussed above. The aforementioned references further teach the additional limitations wherein said future time is at least 10 minutes after an end of said time-period (Gupta’s [0003]: “ The current blood glucose level is used by the controller to predict the blood glucose level 30 minutes in the future.”; and see also Hampapuram’s [0077]: “In some embodiments, the processor module 214 may provide a predictive alert on a smartphone 18 display or user interface 222 when a severe hypoglycemic event is predicted to occur in the near future. For example, the predictive alert may be shown when a severe hypoglycemic event is predicted to occur within 5 minutes, 10 minutes, 15 minutes, 20 minutes, 30 minutes, etc.”). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 36, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 1, as discussed above. The aforementioned references further teach the additional limitations wherein at least one layer of said procedure, other than said at least one pair of layers, is a fully connected layer (Gupta’s [0006]: “After being weighted and transformed by a function, the activations (i.e., outputs) of the neurons of one “layer” are then passed on to other neurons of another “layer.””; and Hampapuram’s [0115]: “Neurons may be totally connected and feedforward”). The motivation for combining the references is as discussed above in relation to claim 1.
Regarding claim 38, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 1, as discussed above. The aforementioned references further teach the additional limitations wherein at least one layer of said pair is a hidden layer (Gupta’s [0009]-[0012, [0023], and [0025]]). The motivation for combining the references is as discussed above in relation to claim 1.
8. Claim 34 is rejected under 35 U.S.C. 103 as being unpatentable over Gupta in view of Hampapuram and Smith and further in view of U.S. Patent No. 5649066 (“Lacher”).
Regarding claim 34, Gupta in view of Hampapuram and further in view of Smith teaches the method according to claim 1, as discussed above. The aforementioned references teach the additional limitations wherein said inter-layer connections are defined in terms of having weight (Gupta’s [0005]) but not defined in terms of weight via a triangular weight matrix. Rather, the Examiner relies upon LACHER to teach what Gupta etc. otherwise lack, see e.g., Lacher’s column 23 lines 25-36 discussing a weight arrangement for a neural network as further recited (“... Symmetry is not an obviously natural property of rule bases but preliminary ideas indicate possible applications of "artificial" symmetry: begin with an acyclic net with a triangular weight matrix W; apply MFT to the symmetric matrix W+W.sup.t, maintaining symmetry; then apply the upper triangle back to the original net. Other possibilities can be explored. MFT is potentially significant since it displays better memory stability properties than many other neural net learning algorithms.”).
Like Gupta and Smith, Lacher is explicitly directed to machine learning and neural networks specifically, and their training. Hence, the aforementioned references are similarly directed and therefore analogous. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Lacher’s weight matrix as taught into Gupta’s modified framework, with a reasonable expectation of success, to promote improved memory stability as Lacher mentions per the cited portion.
Response to Arguments
9. Applicants’ arguments, as received 10/16/25, have been fully considered but are respectfully not persuasive:
On page 8 of Applicants’ Reply, Applicants argue there is no support for Gupta’s input data including “a time-ordered series of levels”, and specifically there is no correlation between the intermittent collection of data based on intervals, as Gupta teaches, with that data being “fed into the deep neural network as a time-ordered series of levels.”
At the very outset, Applicants’ argument is not commensurate to the scope of the claims and their broadest reasonable interpretation.
Applicants’ claims recite “receiving a time-ordered series of levels of the analyte, monitored over a time-period.” In this limitation, there is no specificity as to whom or what is doing this receiving. Hence, Gupta’s teaching, as cited by the Examiner, and as admitted by Applicants, reads on this, because intermittently collecting the data as described on a prescribed/regulated interval as Gupta teaches is a “receiving” that is ordered in accordance with time.
The claims further recite “feeding a trained neural network procedure with said monitored levels.” There is no mention of time here, frankly. Just “levels.” Why couldn’t this limitation be fairly read to merely provide the values of the collected/monitored readings into a deep neural network or its equivalent?
To put a sharp point on this, Applicants have argued that data collected at time intervals is not necessarily time-ordered; however, this limitation, with respect to the argued scope, is lacking in exactly the same regard.
The Examiner is not persuaded that Gupta is actually deficient in the way Applicants have characterized it.
Gupta’s inputs, e.g. “raw input data” as show per FIG. 1 element 110, is an input that is shown to be provided to both a modified Bergman Minimal Model and a DNN. The most reasonable reading, unless Gupta says otherwise, is that the raw input data as shown is the same for both.
As an additional consideration, the Examiner has attached Bergman’s paper “Quantitative estimation of insulin sensitivity”, noted in the present 892 form and listed below in section 10 of this Action. The Examiner posits that the teachings in this paper inform an understanding of the model as relied upon in Gupta’s FIG. 1, and by extension informs what Gupta’s version of this model takes as inputs.
In this paper, the inputs to the Bergman model are described as a “time course”, see e.g., its Abstract on page E667 and also its FIG. 1 on page E669, which the Examiner believes fairly reads on the notion of time-series data, which the Examiner reasonably likens to the general understanding of what constitutes time-series data, and which Gupta with its determined schedule of input collection is clearly equipped to generate.
Bergman’s FIG. 3 on its page E671 further supports the notion that inputs to a Bergman model would be inputs that are time-dependent and hence time-ordered.
Gupta makes no discussion about departing from the nature of the data as is established in the state of the art that it is seeking to borrow from and improve upon. To assert this without making a proper showing is unsupported speculation on Applicants’ part. While Gupta frames its Bergman model as a modified one, Gupta does not teach a modification that varies the nature of its inputs from what is shown in Bergman’s paper.
Moreover, insulin sensitivity and metabolism are inherently time-based, as generally understood. Given that, it is not persuasive, to then assert that the data at hand, e.g. inputs that are used in what is essentially an artificial pancreas directed to monitor/regulate/improve these bodily processes, are not necessarily time-ordered.
On page 8 of Applicants’ Reply, Applicants further argue with respect to Smith, e.g. that Smith’s drop-in feature is merely a temporary training mechanism that ceases to exist once training is complete. The Examiner notes that no support is provided by Applicants in arguing any of this. No excerpts or citations are provided detailing where Smith teaches this. The Examiner has reread Smith and does not find support for these arguments. If Applicants can, with detail, point to teachings in Smith that support these conclusions with clarity, then the Examiner will reconsider. However, just on the Examiner’s rereading of the reference in light of Applicants’ arguments, the Examiner has not understood it in the same manner that Applicants are arguing.
For these reasons given here, the Examiner does not find Applicants’ arguments to be persuasive at this time.
Conclusion
10. The prior art made of record and not relied upon is considered pertinent to Applicants’ disclosure:
WO 2017/129634 A1
Non-Patent Literature “Quantitative estimation of insulin sensitivity” (BERGMAN)
US 2021/0212606 TRAN: [0341]
US 10561788 ROY: col. 2 lines 1-13, col. 6 lines 40-58, and col. 22 lines 29-43
11. THIS ACTION IS MADE FINAL. Applicants are 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.
12. Any inquiry concerning this communication or earlier communications from the examiner should be directed to SHOURJO DASGUPTA whose telephone number is (571)272-7207. The examiner can normally be reached M-F 8am-5pm CST.
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, Tamara Kyle can be reached at 571 272 4241. 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.
/SHOURJO DASGUPTA/Primary Examiner, Art Unit 2144