DETAILED ACTION
1. This office action is in response to the Application No. 18154269 filed on 02/21/2026. Claims 1-20 are presented for examination and are currently pending.
Notice of Pre-AIA or AIA Status
2. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Allowable Subject Matter
3. Claims 8 and 18 are objected to as being dependent upon a rejected base
claim, but would be allowable if rewritten in independent form including all of the
limitations of the base claim.
Response to Arguments
4. Upon further review, the Examiner who has now been assigned this case has determined that the Applicant’s argument are persuasive.
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.
5. Claims 1-20 are rejected under 35 U.S.C 101 because the claimed invention is directed towards an abstract idea without significantly more.
Step 1
Independent claim 1 is directed to a method, and falls into one of the four statutory categories.
Step 2A, Prong 1
Claim 1 recites the following abstract ideas:
dynamically selecting a learning model (Mental process directed to selecting a learning model. This can be done by observing the learning model and making a judgement on the selection of the model) for
determining output data based on sensor data from the at least one sensor (Mental process directed to determining output data from the sensor data. This can be done by observing the output data and making a judgement on the data to be determined),
detecting a need for a new learning model for the sensor device based on performance of a currently used learning model in the sensor device (Mental process directed to detecting a need for a new learning model for the sensor device. This is can be done by observing the sensor device and making a judgement on the need a new learning model);
determining at least one feature candidate based on sensor data from the at least one sensor (Mental process directed to determining a feature candidate. This is can be done by observing the sensor data and making the judgement on the feature candidate to be determined),
selecting a new learning model, from a set of candidate learning models, based on the at least one feature candidate and input features of each one of the candidate learning models (Mental process directed to selecting new learning model from a set of candidate learning models based on one feature candidate and input features of each one of the candidate learning models. This can be done by observing the set of candidate learning models and making a judgement on the new learning model to be selected based on the feature candidate and input features of each one of the candidate learning models); and
Step 2A, Prong 2
a sensor device comprising at least one sensor (This limitation is directed to a computer component i.e., sensor device. This is directed to high level recitation of generic computer component and does not integrate the abstract idea into a practical application. See MPEP 2106.05(f)),
the learning model being configured for (This is directed to mere instruction to apply an exception. This does not integrate the abstract idea into a practical application. See (MPEP 2106.05(f))
the method being performed in a model determiner comprising a processor and memory, the method comprising the steps of (This limitation is directed to a computer component i.e., model determiner. This is directed to high level recitation of generic computer component and does not integrate the abstract idea into a practical application. See MPEP 2106.05(f)):
wherein each one of the at least one feature candidate is associated with a different source of sensor data (This limitation is directed to a particular type or source of data, which is field of use. This does not integrate the abstract idea into a practical application. See MPEP 2106.05(h));
using the selected new learning model for operations on the sensor device (This amount to generally linking the use of a judicial exception to a particular technological environment or field of use. This does not integrate the abstract idea into a practical application. See MPEP 2106.05(h)).
Step 2B
for a sensor device comprising at least one sensor (This limitation is directed to a computer component i.e., sensor device. This is directed to high level recitation of generic computer component and does not amount to significantly more than judicial exception See MPEP 2106.05(f)),
the learning model being configured for (This is directed to mere instruction to apply an exception. This does not amount to significantly more than judicial exception See (MPEP 2106.05(f))
the method being performed in a model determiner comprising a processor and memory, the method comprising the steps of (This limitation is directed to a computer component i.e., model determiner. This is directed to high level recitation of generic computer component and does not amount to significantly more than judicial exception. See MPEP 2106.05(f)):
wherein each one of the at least one feature candidate is associated with a different source of sensor data (This limitation is directed to a particular type or source of data, which is field of use. This does not amount to significantly more than judicial exception. See MPEP 2106.05(h));
using the selected new learning model for operations on the sensor device (This amount to generally linking the use of a judicial exception to a particular technological environment or field of use. This does not amount to significantly more than judicial exception. See MPEP 2106.05(h)).
6. Dependent claim 2 is directed to a method, and falls into one of the four statutory categories.
Claim 2 recites the following abstract ideas:
determining a number of correct classifications and number of misclassifications, during a time period, of the currently used learning model (Mental process directed to determining the number of correct classifications and number of misclassifications, during a time period. This can be done by observing the learning model and making a judgement on the number of correct classifications and misclassifications); and
wherein, in the step of detecting a need for a new learning model, the performance is based on the number of correct classifications and the number of misclassifications (Mental process directed to detecting the need for a new learning model. This can be done by observing the performance and making a judgement for detection based on the number of correct classifications and misclassifications).
Claim 2 do not recite any additional elements.
7. Dependent claim 3 is directed to a method, and falls into one of the four statutory categories.
Claim 3 recites the following abstract ideas:
wherein the step of determining further comprises determining a number of no
classifications, during the time period, of the currently used learning model (Mental process directed to determining the number of no classifications in the learning model. This can be done by observing the learning model and making a judgement on the learning model to determine number of no classification); and
wherein, in the step of detecting a need for a new learning model, the performance is based on the number of no classifications (Mental process directed to detecting the need for a new learning model. This can be done by observing the performance and making a judgement for detection based on the number of no classifications) .
Claim 3 do not recite any additional elements.
8. Dependent claim 4 is directed to a method, and falls into one of the four statutory categories.
Claim 4 recites the following abstract ideas:
wherein, in the determining at least one feature candidate step, the sensor data is normalised over time for each sensor prior to determining the at least one feature candidate (Mental process directed to determining the one feature candidate. This can be done by observing the sensor data and making a judgement on the determination of the feature candidate).
Claim 4 do not recite any additional elements.
9. Dependent claim 5 is directed to a method, and falls into one of the four statutory categories.
Claim 5 recites the following abstract ideas:
wherein the step of selecting a new learning model comprises selecting the candidate learning model, from the set of candidate learning models, that has the greatest number of overlapping features with the at least one feature candidate (Mental process directed to selecting a new learning model. This can be done by observing the set of candidate learning models and making a judgement on which learning model has the greatest number of overlapping features with the at least one feature candidate).
Claim 5 do not recite any additional elements.
10. Dependent claim 6 is directed to a method, and falls into one of the four statutory categories.
Claim 6 recites the following abstract ideas:
finding, for each candidate learning model, a set of overlapping features consisting of any feature candidates overlapping the input features (Mental process directed to finding a set of any feature candidates that overlaps with the input features in each candidate learning model. This can be done by observing a set of overlapping features and the input features, then making a judgement if the feature candidate overlaps the input feature);
adding up, for each candidate learning model, the candidate feature weights of the candidate features forming part of the set of overlapping features, yielding a weighted candidate feature score (Mental process directed to adding up the candidate feature weights of the candidate features forming part of the set of overlapping features to yield a weighted candidate feature score, which can be done with a pen and paper); and
selecting the new learning model to be the candidate learning model having the greatest weighted candidate feature score (Mental process directed to selecting the new learning model to be the candidate learning model that has greatest weighted candidate feature score. This can be done by observing the learning models and making a judgement on which learning model has the greatest weighted score).
Claim 6 recite the following additional elements:
obtaining a candidate feature weight for each feature candidate (This limitation is directed to insignificant extra solution activity of data transfer. This limitation does not integrate the abstract idea into a practical application. See MPEP 2106.05(g))
Claim 6 recite the following additional elements:
obtaining a candidate feature weight for each feature candidate (This limitation is directed to insignificant extra solution activity of data transfer and it is well understood routine and conventional. This does not amount to significantly more than judicial exception. See MPEP 2106.05(d)(II), example i)
11. Dependent claim 7 is directed to a method, and falls into one of the four statutory categories.
Claim 7 recites the following abstract ideas:
finding, for each candidate learning model, a set of overlapping features being input features overlapping the at least one feature candidate (Mental process directed to finding a set of input features that overlaps with at least one feature candidate in each candidate learning model. This can be done by observing a set of overlapping candidate features and the input features, then making a judgement if the input feature overlaps the candidate feature);
adding up, for each candidate learning model, the input feature weights for the input features of the overlapping features, yielding a weighted input feature score (Mental process directed to adding up the input feature weights of the overlapping features to yield a weighted input feature score, which can be done with a pen and paper); and
selecting the new learning model to be the candidate learning model having the greatest weighted input feature score (Mental process directed to selecting the new learning model to be the candidate learning model that has greatest weighted input feature score. This can be done by observing the learning models and making a judgement on which learning model has the greatest weighted score).
Claim 7 recites the following additional elements:
obtaining, for each candidate learning model, an input feature weight for each input feature (This limitation is directed to insignificant extra solution activity of data transfer. This limitation does not integrate the abstract idea into a practical application. See MPEP 2106.05(g));
Claim 7 recites the following additional elements:
obtaining, for each candidate learning model, an input feature weight for each input feature (This limitation is directed to insignificant extra solution activity of data transfer and it is well understood routine and conventional. This does not amount to significantly more than judicial exception. See MPEP 2106.05(d)(II), example i);
12. Dependent claim 8 is directed to a method, and falls into one of the four statutory categories.
Claim 8 recites the following abstract ideas:
wherein for at least one subsequent iteration of the method, the step of determining at least one feature candidate is omitted (Mental process directed to determining that at least one feature candidate is omitted. This can be done by observing the feature candidates and making a judgement that at least one feature candidate is omitted),
the step of obtaining an input feature weight is omitted (Mental process directed to omitting the step of obtaining an input feature weight. This can be done by observing the input feature weight and making a judgement about the omission of the input feature weight),
and the step of finding a set of overlapping features is omitted (Mental process directed to finding a set of overlapping features is omitted. This can be done by observing the overlapping features and making a judgement on the overlapping features that is omitted), and
wherein the step of adding up is based on the stored set of input feature weights (Mental process directed to adding up values of the stored input feature weights. This can be done with a pen and paper).
Claim 8 recites the following additional elements:
further comprising the step of: storing the input feature weights associated with respective input features as a set of input feature weights (This limitation is directed to insignificant extra solution activity of data transfer of input feature weights. This limitation does not integrate the abstract idea into a practical application. see
MPEP 2106.05(g));
Claim 8 recites the following additional elements:
further comprising the step of: storing the input feature weights associated with respective input features as a set of input feature weights (This limitation is directed to insignificant extra solution activity of data transfer of input feature weights and it is well understood routine and conventional. This does not amount to significantly more than judicial exception. See MPEP 2106.05(d)(II), example i)
13. Dependent claim 9 is directed to a method, and falls into one of the four statutory categories.
Claim 9 recites the following abstract ideas:
wherein the step of detecting a need for a new learning model for the sensor device is also based on user input (Mental process directed to detecting the need for a new learning model. This can be done by observing the user input and making a judgement for detection based on the user input).
Claim 9 do not recite any additional elements.
14. Dependent claim 10 is directed to a method, and falls into one of the four statutory categories.
Claim 10 do not recite any abstract ideas.
Claim 10 recites the following additional elements:
wherein the learning model is a machine learning, ML, model (This limitation amount to generally linking the use of a judicial exception to a particular technological environment or field of use. This limitation does not integrate the abstract idea into a practical application. See MPEP 2106.05(h)).
Claim 10 recites the following additional elements:
wherein the learning model is a machine learning, ML, model (This limitation amount to generally linking the use of a judicial exception to a particular technological environment or field of use. This does not amount to significantly more than judicial exception. See MPEP 2106.05(h)).
15. Independent claim 11 is directed to a device, and falls into one of the four statutory categories.
With regards to claim 11, it is substantially similar to claim 1, and is rejected in the same manner and reasoning applying.
16. Dependent claim 12 is directed to a device, and falls into one of the four statutory categories.
With regards to claim 12, it is substantially similar to claim 2, and is rejected in the same manner and reasoning applying.
17. Dependent claim 13 is directed to a device, and falls into one of the four statutory categories.
With regards to claim 13, it is substantially similar to claim 3, and is rejected in the same manner and reasoning applying.
18. Dependent claim 14 is directed to a device, and falls into one of the four statutory categories.
With regards to claim 14, it is substantially similar to claim 4, and is rejected in the same manner and reasoning applying.
19. Dependent claim 15 is directed to a device, and falls into one of the four statutory categories.
With regards to claim 15, it is substantially similar to claim 5, and is rejected in the same manner and reasoning applying.
20. Dependent claim 16 is directed to a device, and falls into one of the four statutory categories.
With regards to claim 16, it is substantially similar to claim 6, and is rejected in the same manner and reasoning applying.
21. Dependent claim 17 is directed to a device, and falls into one of the four statutory categories.
With regards to claim 17, it is substantially similar to claim 7, and is rejected in the same manner and reasoning applying.
22. Dependent claim 18 is directed to a device, and falls into one of the four statutory categories.
With regards to claim 18, it is substantially similar to claim 8, and is rejected in the same manner and reasoning applying.
23. Dependent claim 19 is directed to a device, and falls into one of the four statutory categories.
With regards to claim 19, it is substantially similar to claim 9, and is rejected in the same manner and reasoning applying.
24. Independent claim 20 is directed to a machine, and falls into one of the four statutory categories.
With regards to claim 20, it is substantially similar to claim 1, and is rejected in the same manner and reasoning applying.
Claim 20 further recites “A non-transitory computer-readable medium storing instructions” this limitation is directed to generic computer software. This does not integrate the abstract idea into a practical application nor does it amount to significantly more than judicial exception. See MPEP 2106.05(h).
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
25. Claims 1, 9, 10, 11, 19 and 20 are rejected under 35 U.S.C 102(a)(1) as being anticipated by Sudharsan (US20150006456)
Regarding claim 1, Sudharsan teaches a method for dynamically selecting a learning model (Systems and methods are provided for selecting one or more models for predicting medical conditions, abstract; Further details regarding assessing performance of machine learning algorithms and selecting models are provided accordingly to exemplary embodiments [0045]; The method further includes applying the selected one or models and generating a notification when the application of the selected one or more model indicates an intervention is necessary [0011]) for a sensor device (For example, blood glucose levels may be monitored from continuous glucose monitoring (CGM) devices for hypoglycemic patients [0066]) comprising at least one sensor (Additionally, information from one or more sensors may be received such as blood glucose levels from a continuous glucose monitoring (CGM) device [0084], Fig. 4; execute software allowing for monitoring of patient related data including ... utilizing associated sensors to monitor patient conditions [0026]),
the learning model (Models may be stored in the logical library (LL) independently and/or may be categorized per one or more criteria [0077]; LL 410 may be queried and models applicable for both diseases and models that take ... symptoms, or any other suitable factors, into account, may be retrieved and provided as output 412 [0088], Fig. 4) being configured for determining output data (The output 412 may be utilized to generate notifications when a medical intervention may be necessary [0089], Fig. 4) based on sensor data from the at least one sensor (A data storage/stream service (SS) 402 may store and/or receive data related to the patient ... Additional received data by SS 402 may include self-reported data such as blood glucose levels ... Additionally, information from one or more sensors may be received such as blood glucose levels from a continuous glucose monitoring (CGM) device [0084], Fig. 4),
the method being performed in a model determiner comprising a processor and memory (For instance, at least one processor device and a memory may be used to implement the above-described embodiments [0096]),
the method comprising the steps of: detecting a need for a new learning model for the sensor device based on performance (This may include two aspects including estimating the performance of different models in order to choose the best one and estimating a model's prediction error (generalization error) on new data [0059]. The Examiner notes model's prediction error detects a need, and a new learning model is the best one chosen) of a currently used learning model (Again, test error here may be ErrT=E[L(G, Ĝ(X))|T], the population mis-classification error of the classifier trained on T, and Err may be the expected misclassification error of not predicting hypoglycemia [0056]. The Examiner notes the classifier is the currently used learning model) in the sensor device (For example, blood glucose levels may be monitored from continuous glucose monitoring (CGM) devices for hypoglycemic patients [0066]);
determining at least one feature candidate based on sensor data from the at least one sensor (For example, for diabetes patients, the only relevant aspects of data for providing to machine learning algorithms may be the blood glucose levels and the associated time stamps [0036]. The Examiner notes blood glucose levels is a feature candidate based on sensor data, and the instant specification discloses: “Relevant sensor: sensor 3 a-b in the sensor device 2 that provides relevant information. This corresponds to a feature candidate for the new model” (US20230153628 [0081])),
wherein each one of the at least one feature candidate is associated with a different source of sensor data (Additional received data by SS 402 may include self-reported data such as blood glucose levels and associated time stamps self-reported by the patient (SMBG) [0084]);
selecting a new learning model, from a set of candidate learning models (Additionally, new possible models may be automatically evaluated and considered to be added to the pool of models for selection [0090]), based on the at least one feature candidate (selecting a first type of model when the extracted metadata indicates that the patient's blood glucose level is continuous glucose monitor data and selecting a second type of model when the extracted metadata indicates that the patient's blood glucose level is self-monitored blood glucose data [0012]) and input features (A first aspect of the process is determining a quantitative response. For example, there may be a target variable Y, a vector of inputs X, and a prediction model f̂(X) that has been estimated from a training set T [0046].The Examiner notes a vector of inputs X are input features) of each one of the candidate learning models (In such a scenario, rules may be stored in databases 108 that allow for one or more models to be selected from the library of models based on the received data [0082]; Additionally, new possible models may be automatically evaluated and considered to be added to the pool of models for selection [0090]. The Examiner notes new possible models are candidate learning models); and
using the selected new learning model for operations (For example, the selected models may be applied to received data or the extracted metadata to predict medical occurrences [0079]) on the sensor device (For example, blood glucose levels may be monitored from continuous glucose monitoring (CGM) devices for hypoglycemic patients [0066]).
Regarding claim 9, Sudharsan teaches the method according to claim 1, Sudharsan teaches wherein the step of detecting a need for a new learning model for the sensor device is also based on user input (One or more of user or client devices 102 may be further configured to execute software allowing for monitoring of patient related data including the ability to receive user input or utilizing associated sensors to monitor patient conditions. For example, user or client devices 102 may contain an application which allows it to receive data from a paired and/or integrated blood sugar level monitor and then transmit the data to other entities within environment 100 [0026]).
Regarding claim 10, Sudharsan teaches the method according to claim 1, Sudharsan teaches wherein the learning model is a machine learning, ML, model (FIG. 1 is a schematic diagram of an exemplary network environment in which various systems may select one or more models predicting medical conditions [0024]; In embodiments, machine learning algorithms learn all possible patterns within the patient data points ... The prediction variable is the variable that is to be predicted utilizing the machine learning algorithms. Every algorithm has parametric values that define a configuration of the algorithm that best predicts the prediction variable. Such configurations are called models [0038]).
Regarding claim 11, claim 11 is similar to claim 1. It is rejected in the same manner and reasoning applying.
Regarding claim 19, claim 19 is similar to claim 9. It is rejected in the same manner and reasoning applying.
Regarding claim 20, claim 20 is similar to claim 1. It is rejected in the same manner and reasoning applying.
Further Sudharsan teaches a non-transitory computer-readable medium storing instructions for dynamically selecting a learning model (According to some embodiments, a non-transitory computer readable medium is disclosed as storing instructions that, when executed by a computer, cause the computer to perform a method, ... selecting the one or more models from a library of models based on the extracted metadata. The method further includes applying the selected one or models and generating a notification when the application of the selected one or more model indicates an intervention is necessary [0011])
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
26. Claims 2, 3, 5, 12, 13 and 15 are rejected under 35 U.S.C 103 as being unpatentable over Sudharsan (US20150006456) in view of Batalov (US11106994 filed 09/30/2016)
Regarding claim 2, Sudharsan teaches the method according to claim 1, Sudharsan does not explicitly teach the limitations of claim 2.
Batalov teaches further comprising the step of: determining a number of correct classifications and number of misclassifications, during a time period (In one aspect, graph 200 depicts identification of a number of predictions that fall into the two types of correct class predictions and the two types of incorrect class predictions (col. 7, lines 19-21); The Y-axis depicts the number of observations at each output value, col. 6, lines 47-48. The Examiner notes the observations occur over a time period), of the currently used learning model (FIG. 2 illustrates a graph 200 or histogram depicting idealized machine learning performance using a classification model, col. 6, lines 12-14); and
wherein, in the step of detecting a need for a new learning model (For example, two computing applications might have very different requirements for their ML model. One computing application might need to be extremely sure about the positive predictions actually being positive (high precision), and be able to afford to misclassify some positive examples as negative (moderate recall), col. 8, lines 37-43), the performance is based on the number of correct classifications and the number of misclassifications (In one aspect, a ML model that has a greater predictive accuracy, as compared to other ML models, may predict a more accurate number of true and false outcomes, col. 7, lines 4-7).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Sudharsan to incorporate the teachings of Batalov for the benefit reducing the use of essential computing resources, reducing a substantial amount of manual processes used by a classification ML (Machine Learning) model which is unable to auto tune a classification ML model, and reducing unnecessary costs required by a computer environment (Batalov, col. 5, lines 16-20)
Regarding claim 3, Sudharsan and Batalov teaches the method according to claim 2, Batalov teaches wherein the step of determining further comprises determining a number of no classifications, during the time period (unable to distinguish between the positive classes and negative classes, col. 7, lines 16-17. The Y-axis depicts the number of observations at each output value, col. 6, lines 47-48. The Examiner notes the observations occur over a time period and being unable to distinguish between the positive classes and negative classes indicates no classification were made), of the currently used learning model (FIG. 2 illustrates a graph 200 or histogram depicting idealized machine learning performance using a classification model, col. 6, lines 12-14); and
wherein, in the step of detecting a need for a new learning model (For example, two computing applications might have very different requirements for their ML model. One computing application might need to be extremely sure about the positive predictions actually being positive (high precision), and be able to afford to misclassify some positive examples as negative (moderate recall), col. 8, lines 37-43), the performance is based on the number of no classifications (In one aspect, an extremely poor performing ML model, as compared to the perfect ML model, may be unable to distinguish between the positive classes and negative classes, col. 7, lines 14-17).
The same motivation to combine dependent claim 2 applies here.
Regarding claim 5, Sudharsan teaches the method according to claim 1, Sudharsan does not explicitly teach the limitations of claim 5.
Batalov teaches wherein the step of selecting a new learning model comprises selecting the candidate learning model (In one aspect, upon discovery, calculation, identification, and/or selection of a machine learning model, the weight adjusted classification threshold may take into account the performance baseline, col. 4, lines 40-43), from the set of candidate learning models (Accordingly, the total value of a model may be compared to the total value of other models to find ... increased total value and/or improved performance, col. 5, lines 59-62), that has the greatest number of overlapping features with the at least one feature candidate (However, ML models are subject to errors, thus graph 200 depicts that the two histograms overlap at certain predictive outcome scores. In one aspect, an extremely poor performing ML model, as compared to the perfect ML model, may be unable to distinguish between the positive classes and negative classes, and both classes may have mostly overlapping histograms, col. 7, lines 12-18).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Sudharsan to incorporate the teachings of Batalov for the benefit reducing the use of essential computing resources, reducing a substantial amount of manual processes used by a classification ML (Machine Learning) model which is unable to auto tune a classification ML model, and reducing unnecessary costs required by a computer environment (Batalov, col. 5, lines 16-20)
Regarding claim 12, claim 12 is similar to claim 2. It is rejected in the same manner and reasoning applying.
Regarding claim 13, claim 13 is similar to claim 3. It is rejected in the same manner and reasoning applying.
Regarding claim 15, claim 15 is similar to claim 5. It is rejected in the same manner and reasoning applying.
27. Claims 4 and 14 are rejected under 35 U.S.C 103 as being unpatentable over Sudharsan (US20150006456) in view of Al-Ali et al. (US20170055882 filed 08/31/2016)
Regarding claim 4, Sudharsan teaches the method according to claim 1, Sudharsan does not explicitly teach the limitations of claim 4.
Al-Ali teaches in the determining at least one feature candidate step (Once the wireless sensor is affixed to the patient's body, ..., sensor data corresponding to the patient's motion (e.g., acceleration and angular velocity) are obtained [0014]), the sensor data is normalised over time for each sensor prior to determining the at least one feature candidate (The result is a normalized set of values that can be further processed by the method 1200 [0175]; At block 1206, the normalized set of values is processed to determine features that are useful in determining whether a patient is falling [0176]).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Sudharsan to incorporate the teachings of Al-Ali for the benefit of a relatively low cost extender/repeater which can be used to receive signals transmitted from one or more wireless sensors over the wireless communications link(s) using a shorter-range, lower-power-consuming transmission mode (Al-Ali [0089])
Regarding claim 14, claim 14 is similar to claim 4. It is rejected in the same manner and reasoning applying.
28. Claims 6 and 16 are rejected under 35 U.S.C 103 as being unpatentable over Sudharsan (US20150006456) in view of Batalov (US11106994 filed 09/30/2016) and further in view of Bhowan et al. (US20170286866)
Regarding claim 6, Sudharsan teaches the method according to claim 1, Sudharsan does not explicitly teach the limitations of claim 6.
Batalov teaches wherein the step of selecting a new learning model (Also, one or more alternative functions may also be optimized ... based on a defined or selected machine learning model to increase the total value of the machine learning model and decrease performance outcome errors, col. 4, 36-40) comprises the steps of: obtaining a candidate feature weight for each feature candidate (features of the one or more data points may be classified according to the weight adjusted classification threshold in the classification model, as in block 530, col. 12, lines 47-50);
finding, for each candidate learning model (Accordingly, the total value of a model may be compared to the total value of other models to find ... increased total value and/or improved performance, col. 5, lines 59-62),
a set of overlapping features consisting of any feature candidates overlapping the input features (thus graph 200 depicts that the two histograms overlap at certain predictive outcome scores, col. 7, lines 13-15);
and selecting the new learning model to be the candidate learning model having the greatest weighted candidate feature score (For example, a perfect model may have the two histograms at two different ends of the x-axis with no overlap showing that actual positive predictive outcomes all received high scores (e.g., classified as 1) and actual negative predictive outcomes all received low scores (classified as “0”), col. 7, lines 7-12).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Sudharsan to incorporate the teachings of Batalov for the benefit reducing the use of essential computing resources, reducing a substantial amount of manual processes used by a classification ML (Machine Learning) model which is unable to auto tune a classification ML model, and reducing unnecessary costs required by a computer environment (Batalov, col. 5, lines 16-20)
Sudharsan and Batalov does not explicitly teach adding up, for each candidate learning model, the candidate feature weights of the candidate features forming part of the set of overlapping features, yielding a weighted candidate feature score;
Bhowan teaches adding up (For example, the merged edge weight of an edge may be combined by simple addition of the edge weights of the edges from each of the graphs [0011]), for each candidate learning model (a machine learning model may be dependent upon the presence of one or more other features [0009]), the candidate feature weights of the candidate features forming part of the set of overlapping features (the method combines the edge weights of overlapping edges from different graphs (i.e., edges that are shared by two or more of the graphs) to form a merged edge weight for each edge in the merged feature graph [0011]), yielding a weighted candidate feature score (The features of the merged feature graph may be sorted, e.g., into a ranked list, according to a relevance score value. In various embodiments, the relevance score may be based on the edge weights of the edges [0076]);
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Sudharsan and Batalov to incorporate the teachings of Bhowan for the benefit of improving the performance in automated and semi-automated tasks such as problem solving, decision making and prediction (Bhowan [0001-0002])
Regarding claim 16, claim 16 is similar to claim 6. It is rejected in the same manner and reasoning applying.
29. Claims 7 and 17 are rejected under 35 U.S.C 103 as being unpatentable over Sudharsan (US20150006456) in view of Batalov (US11106994 filed 09/30/2016) and further in view of Kelly et al. ("Neural NILM: Deep Neural Networks Applied to Energy Disaggregation" Proceedings of the 2nd ACM international conference on embedded systems for energy-efficient built environments. 2015).
Regarding claim 7, Sudharsan teaches the method according to claim 1, Sudharsan teaches wherein the step of selecting a new learning model comprises the steps of: obtaining, for each candidate learning model (Further details regarding assessing performance of machine learning algorithms and selecting models are provided accordingly to exemplary embodiments [0045]),
Sudharsan does not explicitly teach an input feature weight for each input feature finding, for each candidate learning model, a set of overlapping features being input features overlapping the at least one feature candidate; adding up, for each candidate learning model, the input feature weights for the input features of the overlapping features, yielding a weighted input feature score; and selecting the new learning model to be the candidate learning model having the greatest weighted input feature score.
Batalov teaches and selecting the new learning model to be the candidate learning model having the greatest weighted input feature score (For example, a perfect model may have the two histograms at two different ends of the x-axis with no overlap showing that actual positive predictive outcomes all received high scores (e.g., classified as 1) and actual negative predictive outcomes all received low scores (classified as “0”), col. 7, lines 7-12).
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Sudharsan to incorporate the teachings of Batalov for the benefit reducing the use of essential computing resources, reducing a substantial amount of manual processes used by a classification ML (Machine Learning) model which is unable to auto tune a classification ML model, and reducing unnecessary costs required by a computer environment (Batalov, col. 5, lines 16-20)
Sudharsan and Batalov does not explicitly teach an input feature weight for each input feature; finding, for each candidate learning model, a set of overlapping features being input features overlapping the at least one feature candidate; adding up, for each candidate learning model, the input feature weights for the input features of the overlapping features, yielding a weighted input feature score;
Kelly teaches an input feature weight for each input feature (The weight on the connection from input i to neuron h is denoted by wih (so w is the ‘weights matrix’), pg. 56, right col., second to the last para.);
finding, for each candidate learning model, a set of overlapping features being input features overlapping the at least one feature candidate (We measure the overlap and normalize the overlap to [0,1], pg. 61, left col., second para.);
adding up, for each candidate learning model, the input feature weights for the input features of the overlapping features (The weighted sum (also called the ‘network input’) of the inputs into neuron ... (pg. 56, right col., second to the last para.); (overlapping sub-regions of the input image, pg. 57, left col., section 2.3), yielding a weighted input feature score (Each artificial neuron calculates a weighted sum of its inputs pg. 56, right col., second to the last para.);
It would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention to have modified the method of Sudharsan and Batalov to incorporate the teachings of Kelly in order to enable neural network algorithms to generalize well to an unseen data (Kelly, Abstract).
Regarding claim 17, claim 17 is similar to claim 7. It is rejected in the same manner and reasoning applying.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MORIAM MOSUNMOLA GODO whose telephone number is (571)272-8670. The examiner can normally be reached Monday-Friday 8:00am-5:00pm 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, Michelle T. Bechtold can be reached on (571) 431-0762. 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.
/M.G./Examiner, Art Unit 2148
Prosecution Insights