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 .
This office action is in response to the request for continued examination filed March 16, 2026.
Claims 1-2 and 4-14 have been amended.
Claims 3 and 15 have been canceled.
Claim 21-22 have been added.
Claims 1-2, 4-14, and 16-22 are pending.
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.
Claims 1-2, 4-8, 11-14, and 16-22 are rejected under 35 U.S.C. 103 as being unpatentable Mikhno et al. (US Publication 20220039755A1) in further view of “Three Approaches to Encoding Time Information as Features for ML Models” by Eryk Lewinson (publicly accessible February 17, 2022); hereinafter referred to as Lewinson.
Regarding claim 1, Mikhno teaches a device for determining a predicted value of glycemia, the device including a data processing unit configured to process time-based data (estimating glucose values of a user ... computing device can include a processor device ... inputs can be received and processed via an estimation model for a user to generate a set of estimated glucose values)([0010]), wherein:
the data processing unit is configured to implement a temporal predictive model trained to learn a glycemia cyclic temporal behaviour, where said glycemia cyclic temporal behaviour is defined by variations of glycemia values in a cyclic way over a given cyclic period of time (a customized machine learning-based estimation/prediction model can be built for each patient during a training process ... so that the predictions are not only personalized based on the individual's diabetes profile and behavior, but also equipped to account for deviations from a patient's regular routine … data for a patient can change over time, the estimation/prediction models can be updated on a regular basis)([0051]; a model learns user’s glycemia values over a period of time),
the temporal predictive model being configured to:
receive, as inputs: said time-based data … temporal data, said … temporal data being … as to represent the given cyclic period of time … (FIG. 10 is a block diagram of a system 1000 for estimating glucose values for a particular user or patient ... The system 1000 includes an estimation model 1040 that is configured to receive various inputs 1010, 1020, 1030 ... the inputs include discrete glucose measurements 1010 from a blood glucose meter (BGM) ... The contextual data 1030 from other sources can include, for example, historical sensor glucose measurement data)([0117] and [0118]; current data and historical data are inputted into the model);
deliver, as output, a predicted value of glycemia (The estimation model 1040 processes at least some of the various inputs to continuously generate estimated glucose values 1080)([0118]).
Mikhno differs from the claim in that Mikhno fails to teach the temporal data is encoded, wherein at least a part of the encoded temporal data is encoded according to an angular distribution over the given cyclic period of time according to a predetermined constant angle interval and wherein the angular distribution defines a continuous trigonometric transformation that preserves mathematical continuity between an end of the given cyclic period and a beginning of a subsequent cyclic period to reduce prediction error at a cycle boundary.
However, Lewinson discloses encoding temporal data (Imagine you have just started a new data science project. The goal is to build a model predicting Y, the target variable ... how to create meaningful features using date-related information ... The easiest way to encode time-related information is to use dummy variables (also known as one-hot encoding))(pages 1, 2, and 4), wherein at least a part of the encoded temporal data is encoded according to an angular distribution over a given cyclic period of time according to a predetermined constant angle interval (functions come to the rescue. We can use the following sine/cosine transformations to encode the cyclical time feature into two features ... In the snippet below, we copy the initial DataFrame, add the column with month numbers, and then encode both the month and day of year columns using the sine/cosine transformations)(pages 6 and 7) and wherein the angular distribution defines a continuous trigonometric transformation that preserves mathematical continuity between an end of the given cyclic period and a beginning of a subsequent cyclic period to reduce prediction error at a cycle boundary (Lewinson - Trigonometric functions come to the rescue. We can use the following sine/cosine transformations to encode the cyclical time feature into two features)(page 6; sine/cosine functions define a continuous trigonometric transformation that preserves mathematical continuity).
The examiner notes Mikhno and Lewinson teach a predicting values. As such, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to modify the device of Mikhno to include the encoding of Lewinson such that the device encodes temporal data, wherein at least a part of the encoded temporal data is encoded according to an angular distribution over a given cyclic period of time according to a predetermined constant angle interval and wherein the angular distribution defines a continuous trigonometric transformation that preserves mathematical continuity between an end of the given cyclic period and a beginning of a subsequent cyclic period to reduce prediction error at a cycle boundary. One would be motivated to make such a combination to provide the advantage of improving model accuracy (page 2; Lewinson).
Regarding claim 2, Mikhno-Lewinson teach the device according to claim 1, wherein at least a part of the encoded temporal data is encoded according to a linear distribution over the given cyclic period of time according to a predetermined constant time step (Lewinson - The easiest way to encode time-related information is to use dummy variables (also known as one-hot encoding) ... Below, you can see the output of our operation ... we used the dummy variable approach to capture the month in which the observation was recorded. However, this same approach could be used to indicate a range of other information from the DatetimeIndex)(pages 4 and 5).
Regarding claim 4, Mikhno-Lewinson teach the device according to claim 1, wherein the time-based data includes at least: one past glycemia value (Mikhno - machine learning to determine which combination of historical sensor glucose measurement data)([0107]).
Regarding claim 5, Mikhno-Lewinson teach the device according to claim 1, wherein the temporal predictive model is further configured to receive as input:
a variable IOB(t) representative of the time variation of the patient's quantity of insulin on board (Mikhno - other contextual data from other sources can include, for example, a third set of inputs from an insulin infusion device comprising insulin dispensed)([0147]); and
a variable COB(t) representative of the time variation of the patient's quantity of carbohydrate on board (Mikhno - fourth set of inputs comprising user entries that describe meal information)([0147]).
Regarding claim 6, Mikhno-Lewinson teach the device according to claim 1, wherein the given cyclic period of time is: a day; a week; and/or a year (Mikhno - analysis of variability of glucose on an hourly, daily, weekly, or monthly basis)([0193]).
Regarding claim 7, Mikhno-Lewinson teach the device according to claim 6, wherein the encoded temporal data is tagged according to: a night-time and/or a day-time when the given cyclic period of time is a day; a particular day of the week when the given cyclic period of time is a week; a working day and/or a week-end day when the given cyclic period of time is a week; and/or a day of the year when the given cyclic period of time is a year (Lewinson - we can easily see that the curves are step-wise when using the months for encoding)(page 7).
Regarding method claim 8, the claim generally corresponds to device claim 1, and recites similar features in method form; therefore, the claim is rejected under similar rational.
Regarding claim 11, Mikhno teaches a medical system for regulating a glycemia of a person including:
a device for determining a predicted value of glycemia, the device including a data processing unit configured to process time-based data (estimating glucose values of a user ... computing device can include a processor device ... inputs can be received and processed via an estimation model for a user to generate a set of estimated glucose values)([0010]), wherein:
the data processing unit is configured to implement a temporal predictive model trained to learn a glycemia cyclic temporal behaviour, where said glycemia cyclic temporal behaviour is defined by variations of glycemia values in a cyclic way over a given cyclic period of time (a customized machine learning-based estimation/prediction model can be built for each patient during a training process ... so that the predictions are not only personalized based on the individual's diabetes profile and behavior, but also equipped to account for deviations from a patient's regular routine … data for a patient can change over time, the estimation/prediction models can be updated on a regular basis)([0051]; a model learns user’s glycemia values over a period of time),
the temporal predictive model being configured to:
receive, as inputs: said time-based data … temporal data, said … temporal data being … as to represent the given cyclic period of time (FIG. 10 is a block diagram of a system 1000 for estimating glucose values for a particular user or patient ... The system 1000 includes an estimation model 1040 that is configured to receive various inputs 1010, 1020, 1030 ... the inputs include discrete glucose measurements 1010 from a blood glucose meter (BGM) ... The contextual data 1030 from other sources can include, for example, historical sensor glucose measurement data)([0117] and [0118]; current data and historical data are inputted into the model);
deliver, as output, a predicted value of glycemia (The estimation model 1040 processes at least some of the various inputs to continuously generate estimated glucose values 1080)([0118]),
said device including a memory storing the trained temporal predictive model trained to learn the glycemia cyclic temporal behaviour, where said glycemia cyclic temporal behaviour is defined by variations of glycemia values in a cyclic way over a given cyclic period of time (FIG. 10 is a block diagram of a system 1000 for estimating glucose values ... The processing system can be implemented at any of the devices described ... above with reference to FIGS. 1 and 8, and more generically at a computer-based or processor-based device 900)([0117]); a processing device includes memory for storing models);
a medical device wearable by a patient (FIG. 1, one exemplary embodiment of an infusion system 100 ... infusion device 102 and the sensing arrangement 104 are secured at desired locations on the body of a user)([0057]), including:
a dispenser of insulin, controllable according to the predicted value of glycemia (a fluid infusion device includes a motor or other actuation arrangement that is operable to linearly displace a plunger ... to deliver a dosage ... the method 1200 can proceed to 1280, where a corrective action can be taken at the insulin infusion device to deliver the patient an amount of insulin required)([0049] and [0153]).
Mikhno differs from the claim in that Mikhno fails to teach the temporal data is encoded. However, encoding temporal data is taught by Lewinson (Imagine you have just started a new data science project. The goal is to build a model predicting Y, the target variable ... how to create meaningful features using date-related information ... The easiest way to encode time-related information is to use dummy variables (also known as one-hot encoding))(pages 1, 2, and 4). The examiner notes Mikhno and Lewinson teach a predicting values. As such, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Mikhno to include the encoding of Lewinson such that the system encodes temporal data. One would be motivated to make such a combination to provide the advantage of improving model accuracy (page 2; Lewinson).
Regarding claim 12, Mikhno-Lewinson teach the medical system according to claim 11, wherein the medical device further includes:
a data acquisition system configured to acquire at least time-based data (Mikhno - exemplary embodiment of an infusion system 100 includes ... a sensing arrangement 104)([0057]);
a communication module configured to repeatedly transmit the acquired time-based data to the data processing unit of the device (Mikhno - includes a transmitter, a receiver, and/or other transceiver electronics that allow for communication ... so that the sensing arrangement 104 may transmit sensor data)([0059]).
Regarding claim 13, Mikhno-Lewinson teach the medical system according to claim 11, wherein the device according to claim 1 is embedded in the medical device (Mikhno - FIG. 10 is a block diagram of a system 1000 for estimating glucose values ... The processing system can be implemented at any of the devices described ... above with reference to FIGS. 1 and 8)(Figures 1-8 – various embodiments of embedding a device in a medical device are shown).
Regarding computer product claim 14, the claim generally corresponds to device claim 1, and recites similar features in computer product form; therefore, the claim is rejected under similar rational.
Regarding claim 16, Mikhno teach a device for determining a predicted value of glycemia, the device comprising:
a data processing unit configured to process time-based data (estimating glucose values of a user ... computing device can include a processor device ... inputs can be received and processed via an estimation model for a user to generate a set of estimated glucose values)([0010]), the data processing unit implementing a temporal predictive model trained to identify a glycemia cyclic temporal behavior, the glycemia cyclic temporal behavior being defined by variations of glycemia values over a given cyclic period of time (a customized machine learning-based estimation/prediction model can be built for each patient during a training process ... so that the predictions are not only personalized based on the individual's diabetes profile and behavior, but also equipped to account for deviations from a patient's regular routine … data for a patient can change over time, the estimation/prediction models can be updated on a regular basis)([0051]; a model learns user’s glycemia values over a period of time),
wherein the temporal predictive model is configured to:
receive, as inputs, the time-based data … temporal data representing the given cyclic period of time … (FIG. 10 is a block diagram of a system 1000 for estimating glucose values for a particular user or patient ... The system 1000 includes an estimation model 1040 that is configured to receive various inputs 1010, 1020, 1030 ... the inputs include discrete glucose measurements 1010 from a blood glucose meter (BGM) ... The contextual data 1030 from other sources can include, for example, historical sensor glucose measurement data)([0117] and [0118]; current data and historical data are inputted into the model); and
deliver, as output, a predicted value of glycemia (The estimation model 1040 processes at least some of the various inputs to continuously generate estimated glucose values 1080)([0118]).
Mikhno differs from the claim in that Mikhno fails to teach the temporal data is encoded according to an angular distribution over a cyclic period having a predetermined constant angle interval. However, encoding temporal data according to an angular distribution over a cyclic period having a predetermined constant angle interval is taught by Lewinson (Imagine you have just started a new data science project. The goal is to build a model predicting Y, the target variable ... how to create meaningful features using date-related information ... The easiest way to encode time-related information is to use dummy variables (also known as one-hot encoding) … Trigonometric functions come to the rescue. We can use the following sine/cosine transformations to encode the cyclical time feature into two features ... In the snippet below, we copy the initial DataFrame, add the column with month numbers, and then encode both the month and day of year columns using the sine/cosine transformations)(pages 1, 2, 4, 6, and 7). The examiner notes Mikhno and Lewinson teach a predicting values. As such, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to modify the device of Mikhno to include the encoding of Lewinson such that the device encodes temporal data according to an angular distribution over a cyclic period having a predetermined constant angle interval. One would be motivated to make such a combination to provide the advantage of improving model accuracy (page 2; Lewinson).
Regarding claim 17, Mikhno-Lewinson teach the device according to claim 16, wherein the angular distribution includes a sine function and a cosine function mapping the time-based data to a continuous geometric representation (Lewinson - Trigonometric functions come to the rescue. We can use the following sine/cosine transformations to encode the cyclical time feature into two features))(page 6).
Regarding claim 18, Mikhno-Lewinson teach the device according to claim 16, wherein the given cyclic period of time is a day, and the predetermined constant angle interval corresponds to a minute or an hour of the day (Mikhno - analysis of variability of glucose on an hourly, daily, weekly, or monthly basis)([0193]) mapped to a unit circle (Lewinson - Trigonometric functions come to the rescue. We can use the following sine/cosine transformations to encode the cyclical time feature into two features))(page 6; sine/cosine functions map data to a circle).
Regarding claim 19, Mikhno-Lewinson teach the device according to claim 16, wherein the temporal predictive model is a recurrent neural network (RNN) or Long Short-Term Memory (LSTM) network (Mikhno - Examples of deep learning algorithms can include, for example ... recurrent neural networks (RNNs), long short term memory (LSTM) networks)([0132]) configured to process the angularly distributed encoded temporal data as a continuous input vector (Lewinson - Trigonometric functions come to the rescue. We can use the following sine/cosine transformations to encode the cyclical time feature into two features))(page 6; sine/cosine functions encode data in an angular continuous manner).
Regarding claim 20, Mikhno-Lewinson teach a medical system including the device of claim 16, further comprising an insulin pump configured to dispense insulin based on the predicted value of glycemia output by the temporal predictive model (Mikhno - a fluid infusion device includes a motor or other actuation arrangement that is operable to linearly displace a plunger ... to deliver a dosage ... the method 1200 can proceed to 1280, where a corrective action can be taken at the insulin infusion device to deliver the patient an amount of insulin required)([0049] and [0153]).
Regarding claim 21, Mikhno-Lewinson teach the device of claim 1, wherein the temporal predictive model is further configured to identify particular times of interest within the given cyclic period of time, the particular times of interest being defined by a recurring pattern of glycemia behaviour linked to the patient's habits (Mikhno - the patient's historical meal data and corresponding measurement data or other contextual data may be analyzed to predict a future time when the next meal is likely to be consumed by the patient … Machine learning facilitates a continuous advancement of computing through exposure to new scenarios, testing and adaptation, while employing pattern and trend detection for improved decisions and subsequent, though not identical, situations)([0106] and [0121]; machine learning is used to identify times of interest (e.g., future meal time) based on pattern detection).
Regarding claim 22, Mikhno-Lewinson teach the device of claim 16, wherein the angular distribution defines a continuous trigonometric transformation that preserves mathematical continuity between an end of the given cyclic period of time and a beginning of a subsequent cyclic period of time, to reduce prediction error at a cycle boundary (Lewinson - Trigonometric functions come to the rescue. We can use the following sine/cosine transformations to encode the cyclical time feature into two features)(page 6; sine/cosine functions define a continuous trigonometric transformation that preserves mathematical continuity).
Claims 9 and 10 are rejected under 35 U.S.C. 103 as being unpatentable Mikhno, Lewinson, in further view of “Project Adam: Building an Efficient and Scalable Deep Learning Training System” by Chilimbi et al. (publicly accessible October 6-8, 2014); hereinafter referred to as Chilimbi.
Regarding claim 9, Mikhno-Lewinson teach the computerized method as applied above, including acquiring a plurality of at least time-based data for a plurality of patients, injecting said plurality of time-based data and the encoded temporal data in the temporal predictive model, training said temporal predictive model, and storing said temporal predictive model (Mikhno – FIG. 14 is a block diagram of a machine learning system 1400 for generating an optimized population model 1470 for estimating glucose values … The processing system can be implemented at any of the devices described ... above with reference to FIGS. 1 and 8, and more generically at a computer-based or processor-based device 900 … Data for each of the inputs can be collected for any number of users for a time period ... the discrete glucose measurement data 1410, contextual data 1420 from the source of user activity data and contextual data 1430 from the other sources can then be processed at a population learning processor 1440 to generate the optimized population model ... For example, the population learning processor 1440 can apply various machine learning models 1450)([0165], [0167], and [0168]; a processing device includes memory for storing models).
Mikhno-Lewinson differ from the claim in that Mikhno-Lewinson fails to explicitly teach training a model until it converges.
However, training a model until it converges is taught by Chilimbi (Typically, training continues for multiple epochs, reprocessing the training data set each time, until the validation set error converges to a desired (low) value)(page 4).
The examiner notes Mikhno, Lewinson, and Chilimbi teach a predicting values. As such, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to modify the method of Mikhno-Lewinson to include the training of Chilimbi such that the method trains until a model converges. One would be motivated to make such a combination to provide the advantage of reducing error (page 4; Chilimbi).
Regarding claim 10, Mikhno-Lewinson teach the computerized method as applied above, including acquiring a plurality of at least time-based data for a patient, injecting said plurality of time-based data and the encoded temporal data in the temporal predictive model, training said temporal predictive model, and storing said temporal predictive model (Mikhno – FIG. 10 is a block diagram of a system 1000 for estimating glucose values ... The processing system can be implemented at any of the devices described ... above with reference to FIGS. 1 and 8, and more generically at a computer-based or processor-based device 900 …the inputs include discrete glucose measurements 1010 from a blood glucose meter ... contextual data 1030 from other sources ... The contextual data 1030 from other sources can include, for example, historical sensor glucose measurement data ... The estimation model 1040 processes at least some of the various inputs to continuously generate estimated glucose values 1080 ... FIG. 11 is a flowchart that illustrates a method 1100 for generating an estimation model ... The method 1100 begins at 1110 where a computing system trains multiple generic estimation models)([0117], [0118], and [0141]; a processing device includes memory for storing models).
Mikhno-Lewinson differ from the claim in that Mikhno-Lewinson fails to explicitly teach training a model until it converges.
However, training a model until it converges is taught by Chilimbi (Typically, training continues for multiple epochs, reprocessing the training data set each time, until the validation set error converges to a desired (low) value)(page 4).
The examiner notes Mikhno, Lewinson, and Chilimbi teach a predicting values. As such, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to modify the method of Mikhno-Lewinson to include the training of Chilimbi such that the method trains until a model converges. One would be motivated to make such a combination to provide the advantage of reducing error (page 4; Chilimbi).
Response to Arguments
Applicant's arguments, see page 8 of applicant's remarks, filed March 16, 2026, with respect to the 35 U.S.C. §112 rejection have been fully considered and are persuasive. The 35 U.S.C. §112 rejection has been withdrawn.
Applicant's arguments, see pages 8-12 of applicant's remarks, filed March 16, 2026, with respect to the 35 U.S.C. §103 rejection have been fully considered but they are not persuasive.
Regarding claims 1-2, 4-14, and 16-22, applicant argues the combination of Mikhno and Lewinson is improper because Lewinson teaches away from Mikhno; the examiner respectfully disagrees.
While Mikhno does recite using tree models “Examples of machine learning algorithms or models that can be implemented at the machine learning model 1060 can include, but are not limited to ... one or more decision tree models (e.g., a random forest model)” ([0128]), Mikhno explicitly states learning models are not limited to just tree models. Specifically, Mikhno discloses any known models including deep learning models (i.e., neural network) may be used “Examples of machine learning algorithms or models that can be implemented at the machine learning model 1060 can include, but are not limited to … one or more deep learning networks (e.g., at least one recurrent neural network, sequence to sequence mapping using deep learning, sequence encoding using deep learning … Deep learning is a specific approach used for building and training neural networks” ([0128] and [0132]).
It is readily apparent to one of ordinary skill in the art that neural networks are capable of encoding data in angular manner. For example, in the post “Encoding Angle Data for Neural Network” states that neural network data is encoded using sine/cosine “I am training a neural network … So the encoding function is: Ɵ ↦(sin(Ɵ),cos(Ɵ))” (page 1). As such, Lewinson does not teach away from Mikhno nor render it unsatisfactory for it intended purpose.
Regarding claims 1-2, 4-14, and 16-22, applicant argues the combination of Mikhno and Lewinson is improper because Lewinson is non-analogous art; the examiner respectfully disagrees.
The examiner notes when determining field of endeavor, the primary focus is on what the reference discloses. Airbus, 41 F.3d at 1380. The examiner must consider the disclosure of each reference “in view of the ‘the reality of the circumstances.’” Airbus, 41 F.3d at 1380 (quoting Bigio, 381 F.3d at 1326, 72 USPQ2d at 1212). These circumstances are to be weighed “from the vantage point of the common sense likely to be exerted by one of ordinary skill in the art in assessing the scope of the endeavor.” Airbus, 41 F.3d at 1380. See also Donner Technology, LLC v. Pro Stage Gear, LLC, 979 F.3d 1353, 2020 USPQ2d 11335 (Fed. Cir. 2020); Sanofi-Aventis, 66 F.4th at 1378; and Netflix, Inc. v. DivX, LLC, 80 F.4th 1352, 1358-59, 2023 USPQ2d 1057 (Fed. Cir. 2023).
It would have been common sense to one of ordinary skill in the art that predicting values is within the scope Mikhno. In particular, Mikhno discloses training a temporal model based on a patient’s diabetic profile and routine “a customized machine learning-based estimation/prediction model can be built for each patient during a training process ... so that the predictions are not only personalized based on the individual's diabetes profile and behavior, but also equipped to account for deviations from a patient's regular routine … data for a patient can change over time, the estimation/prediction models can be updated on a regular basis” ([0051]), receiving current time based data to represent a given cyclic period of time “FIG. 10 is a block diagram of a system 1000 for estimating glucose values for a particular user or patient ... The system 1000 includes an estimation model 1040 that is configured to receive various inputs 1010, 1020, 1030 ... the inputs include discrete glucose measurements 1010 from a blood glucose meter (BGM) ... The contextual data 1030 from other sources can include, for example, historical sensor glucose measurement data” ([0117] and [0118]), and predicting a value of glycemia “The estimation model 1040 processes at least some of the various inputs to continuously generate estimated glucose values 1080” ([0118]).
Similarly, it would have been common sense to one of ordinary skill in the art that predicting values is within the scope of Lewinson. In particular, Lewinson discloses encoding temporal data to predict values “Imagine you have just started a new data science project. The goal is to build a model predicting Y, the target variable ... how to create meaningful features using date-related information ... The easiest way to encode time-related information is to use dummy variables (also known as one-hot encoding)” (pages 1, 2, and 4).
As both, Lewinson and Mikhno are in the same field of inventor’s endeavor (i.e., predicting values), Lewinson is analogous art.
Regarding claims 1-2, 4-14, and 16-22, applicant argues the combination of Mikhno and Lewinson is improper because there is no motivation to combine; the examiner respectfully disagrees.
The examiner recognizes that obviousness may be established by combining or modifying the teachings of the prior art to produce the claimed invention where there is some teaching, suggestion, or motivation to do so found either in the references themselves or in the knowledge generally available to one of ordinary skill in the art. See In re Fine, 837 F.2d 1071, 5 USPQ2d 1596 (Fed. Cir. 1988), In re Jones, 958 F.2d 347, 21 USPQ2d 1941 (Fed. Cir. 1992), and KSR International Co. v. Teleflex, Inc., 550 U.S. 398, 82 USPQ2d 1385 (2007). In this case, Mikhno and Lewinson teach a predicting values. As such, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to modify Mikhno to include the encoding of Lewinson such that temporal data is encoded. One would be motivated to make such a combination to provide the advantage of improving model accuracy (page 2; Lewinson).
Regarding claims 1-2, 4-14, and 16-22, applicant argues the combination of Mikhno and Lewinson is improper because there is no reasonable expectation of success; the examiner respectfully disagrees.
As noted above, Lewinson does not teach away from Mikhno nor render it unsatisfactory for it intended purpose. Additionally, the examiner notes reasonable expectation of success can be implicitly shown via the prior art teachings or as part of the obviousness analysis. See Elekta Ltd. v. ZAP Surgical Sys., Inc., 81 F.4th 1368, 1376-77, 2023 USPQ2d 1100 (Fed. Cir. 2023). As both Mikhno and Lewinson teach a predicting values there would be likelihood of success of combining of Mikhno and Lewinson.
Conclusion
The prior art made of record on form PTO-892 and not relied upon is considered pertinent to applicant's disclosure. Applicant is required under 37 C.F.R. § 1.111(c) to consider the reference fully when responding to this action. The document cited therein and enumerated below teaches a method and apparatus for encoding temporal data.
Encoding Angle Data for Neural Network
Encoding and Decoding Cyclic Data
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Yongjia Pan whose telephone number is (571)270-1177. The examiner can normally be reached Monday - Friday, 9:00 AM - 5:00 PM EST.
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, Scott Baderman can be reached at 571-272-3644. 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.
/YONGJIA PAN/Primary Examiner, Art Unit 2118
