DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Priority
As detailed on the Filing Receipt filed 9/6/2023, the instant application claims priority to as early as 1/29/2021. At this point in prosecution, all claims are accorded the earliest claimed priority date.
Information Disclosure Statement
The Information Disclosure Statement filed on 4/21/2023 is in compliance with the provisions of 37 CFR 1.97 and has been considered in full. A signed copy of the IDS is included with this Office Action.
Claim Status
Claims 1-15 are pending, and under examination.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 USC § 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
Claim 3 is rejected under 35 USC § 112(b) as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor, or a joint inventor, regards as the invention.
With respect to claim 3, there is uncertainty regarding the scope of the recited limitation of “further comprising generating by the computing device, applying” (lines 1-2). It is unclear what this limitation requires a user to “generat[e]”. In the interest of compact prosecution, the recited limitation is simply interpreted as “further comprising applying”.
For the above reasons, the claim is indefinite.
Claim Rejections - 35 USC § 101
35 USC § 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-15 are rejected under 35 USC § 101 because the claimed invention is directed to an abstract idea without significantly more (i.e., non-statutory subject matter).
"Claims directed to nothing more than abstract ideas, natural phenomena, and laws of nature are not eligible for patent protection" (MPEP 2106.04 § I). Abstract ideas include mathematical concepts (including formulas, equations and calculations), and procedures for evaluating, analyzing or organizing information, which are a type of mental process (MPEP 2106.04(a)(2)).
The claims as a whole, considering all claim elements both individually and in combination, do not amount to significantly more than an abstract idea.
Step 1: The Four Categories of Statutory Subject Matter (MPEP 2106.03)
The claims are directed to a method (claims 1-15), which falls under the ‘process’ category of statutory subject matter.
Step 2A, Prong One: Whether the Claims Set Forth or Describe a Judicial Exception (MPEP 2106.04 § II.A.1)
‘Mathematical concepts’ are relationships between variables and numbers, numerical formulas or equations, or acts of calculation, which need not be expressed in mathematical symbols (MPEP 2106.04(a)(2) § I). The claims recite elements which encompass mathematical concepts, at least under their broadest reasonable interpretation, including:
generating, by… applying the first dataset and the second dataset as inputs to a machine learning model, an output including a machine learning score (claim 1);
the machine learning model [is] trained to map the inputs to the output to minimize a cost function defined by the machine learning model and maximize the dissimilarity between the patient between the first point in time and the second point in time (claim 1);
generating, by… executing the machine learning model in view of the inputs, a first intermediate score associated with the first dataset and a second intermediate score associated with the second dataset (claim 2);
computing a difference between the first intermediate score and the second intermediate score to derive the machine learning score (claim 2);
applying at least a portion of the input data to the machine learning model to generate a distribution of intermediate scores (claim 3);
the machine learning model is trained to learn certain ones of the one or more outcome measures that represent a maximal dissimilarity of the patient from the first point in time to the second point in time (claim 4);
the machine learning model comprises a Siamese neural network that includes an input layer defining a node for each outcome measure of the one or more outcome measures, and an output layer that includes a node that provides intermediate scores including the first intermediate score and the second intermediate score (claim 5);
the cost function is defined as Jmin (s1,s2) = - mean(s2-s1) / std(s2-s1) wherein S2 corresponds to the second point in time and S1 correspond to the first point in time, and the cost function assists the machine learning model during training to maximize the difference between S2 and S1 (claim 6);
the function is a contrastive objective function (claim 7)
normalizing… the first dataset and the second dataset by rescaling each outcome measure from the first dataset and the second dataset to a range [0, 1] using the minimum and maximum values for each outcome measure (claim 8);
during machine learning… deriv[ing] an equation defining a plurality of computations performed by the machine learning model when executed, parameters of the equation being trained using the cost function to find the largest difference in patients between two points in time (claim 11);
the machine learning model during training modifies the parameters to minimize the cost function based on training data defining outcome measures fed to the machine learning model during training (claim 12); and
feeding incrementally the machine learning model with additional outcome measures training data and updating the parameters (claim 13).
The recited machine learning model, cost function, mathematical constraints, and acts of calculation constitute mathematical concepts.
‘Mental processes’ are processes that can be performed in the human mind at least with use of a physical aid, e.g., a slide rule or pen and paper (MPEP 2106.04(a)(2) § III). The claims recite elements that encompass processes that are practicably performable in the human mind, at least under their broadest reasonable interpretation, including:
determining… a suggested activity for the patient based on the outcome measure (claim 1), i.e., determining information based on an associated value;
for each feature in a feature matrix defined by the first dataset; and
appending an additional column to the first dataset to serve as a mask for identifying missing data, and assigning values of 1 and 0 (claims 14-15).
The recited steps of evaluating information, which are practicably performable in the human mind, constitute mental processes.
The following claim elements delimit embodiments of the identified mathematical concepts and mental processes:
the machine learning score infers improvement of the patient from the first point in time to the second point in time (claim 1);
the distribution of intermediate scores reflect[s] computed changes in each of the one or more outcome measures at respective points in time (claim 3);
greater scores of the distribution of intermediate scores reflect[] greater improvement of the patient made during the rehabilitation (claim 3);
the first dataset and the second dataset correspond to phases of rehabilitation of the patient (claim 7);
the objective function uses an assumption of patient improvement from the first point in time to the second point in time (claim 7);
the one or more outcome measures includes any metric configured as a numeric value informative as to a change in the patient (claim 10);
assigned values of 1 reflect population of a value for a given outcome measure (claim 14); and
assigned values of 0 reflect missing data (claim 15).
The above elements merely indicate that manipulated data (e.g., evaluated and calculated values) has certain representative significance, which does not alter the statutory characterization of the identified mathematical concepts and mental processes that operate upon said data.
Mathematical concepts and mental processes constitute enumerated groupings of abstract ideas (MPEP 2106.04(a)(2) §§ I and III). Hence, the claims recite elements that, individually and in combination, constitute an abstract idea. The claims must therefore be examined further to determine whether they integrate the abstract idea into a practical application (MPEP 2106.04(d)).
Step 2A, Prong Two: Whether the Claims Contain Additional Elements that Integrate the Judicial Exception(s) into a Practical Application (MPEP 2106.04 § II.A.2)
The claims recite additional elements that gather data necessary for performance of claimed method steps, including:
accessing… a first dataset of input data for one or more outcome measures derived from a patient at a first point in time of rehabilitation (claim 1); and
accessing… a second dataset of input data for one or more outcome measures derived from the patient at a second point in time of the rehabilitation (claim 1).
Necessary data gathering is considered to be insignificant pre-solution activity, and as such insufficient to integrate an abstract idea into a practical application (MPEP 2106.05(g)).
The claims further recite additional elements that require performance of claimed functions on a computer, including:
“by [a] computing device” (claims 1-3 and 8-9);
transmitting… [information] to an end user device (claim 9); and
“the computing device derives” (claim 11).
The claims do not describe any specific computational steps by which a computer performs or carries out functions drawn to the abstract idea, nor do they provide any details of how specific structures of a computer are used to implement these functions. The claims state nothing more than that a generic computer performs functions including those drawn to the abstract idea. Use of a computer in its ordinary capacity, or simply adding a general purpose computer to an abstract idea, does not integrate a judicial exception into a practical application. The above elements thus constitute mere instructions to apply the abstract idea using a computer, and do not integrate the abstract idea into a practical application. See MPEP 2106.04(d) § I and 2106.05(f)).
No further additional elements are recited.
When the claims are considered as a whole: they do not improve the functioning of a computer, other technology, or technical field (MPEP 2106.04(d)(1) and 2106.05(a)); they do not apply the abstract idea to effect a particular treatment or prophylaxis for a disease or medical condition (MPEP 2106.04(d)(2)); they do not implement the abstract idea with, or in conjunction with, a particular machine (MPEP 2106.05(b)); they do not effect a transformation or reduction of a particular article to a different state or thing (MPEP 2106.05(c)); and they do not apply or use the abstract idea in some other meaningful way beyond linking the use of the abstract idea to a particular technological environment and/or field of use (e.g., assessment of rehabilitative progress; MPEP 2106.05(e) and 2106.05(h)).
Hence, the recited abstract idea is not integrated into a practical application. See MPEP 2106.04(d) § I.
Because the claims recite an abstract idea, and do not integrate that abstract idea into a practical application, the claims are directed to the abstract idea. Claims that are directed to an abstract idea must be examined further to determine whether the additional elements besides the abstract idea render the claims significantly more than the abstract idea. Additional elements besides the abstract idea may constitute inventive concepts that are sufficient to render the claims significantly more (MPEP 2106.05).
Step 2B: Whether the Claims Contain Additional Elements that Amount to an Inventive Concept (MPEP 2106.05)
As noted above, several recited additional elements amount to insignificant extra-solution activity. Mere addition of insignificant extra-solution activity does not amount to an inventive concept that would render the claims significantly more than the recited abstract idea, particularly when the activities are well-understood or conventional (MPEP 2106.05(g)). The conventionality of recited additional elements that amount to insignificant extra-solution activity must be further considered.
Recited additional elements amounting to insignificant extra-solution activity encompass the following computer-implemented functions, which the courts have held as coextensive with a general-purpose computer and/or well-understood, routine and conventional:
Receiving, storing, and processing data, e.g., accessing first and second datasets of input data (In re Katz Interactive Call Processing Patent Litigation, 639 F.3d 1303, 1316 (Fed. Cir. 2011); EON Corp. IP Holdings LLC v. AT&T Mobility LLC, 785 F.3d 616, 622 (Fed. Cir. 2015)); and
Transmitting data over a network, e.g., to an end user device (buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355 (Fed. Cir. 2014); OIP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363 (Fed. Cir. 2015)).
Hence, the encompassed extra-solution activity is considered well-understood, routine and conventional. Well-understood, routine and conventional activity is insufficient to constitute an inventive concept that would render the claims significantly more than judicial exceptions (MPEP 2106.05(d)).
As noted above, several additional elements require performance of functions drawn to the abstract idea using a computer. Use of a computer in its ordinary capacity, or simply adding a general purpose computer to an abstract idea, does not provide significantly more (see, e.g., Alice Corp. v. CLS Bank, 573 U.S. 208, 223-24 (2014)). The direction of the recited computer-implemented functionality to the ordinary capacity of a computer, and the direction of the claimed computing device to a general purpose computer, must be further considered.
Recited computer-implemented functions encompass the following, which the courts have held as coextensive with a general-purpose computer and/or well-understood, routine and conventional:
Receiving, storing, and processing data (In re Katz Interactive Call Processing Patent Litigation, 639 F.3d 1303, 1316 (Fed. Cir. 2011); EON Corp. IP Holdings LLC v. AT&T Mobility LLC, 785 F.3d 616, 622 (Fed. Cir. 2015));
Updating a database (Ultramercial, Inc. v. Hulu, LLC, 772 F.3d 709, 715 (Fed. Cir. 2014));
Selecting information (Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354-55 (Fed. Cir. 2016)); and
Displaying the result of data analysis (TLI Communications LLC v. AV Auto. LLC, 823 F.3d 607, 612-13 (Fed. Cir. 2016)).
Additionally, the specification indicates that the claimed computing device may include personal computers, i.e., general purpose computers (para. 0068).
The requirement that functions of the method must be performed using a computer therefore does not amount to significantly more than mere instructions to implement the abstract idea using a computer, which are insufficient to constitute an inventive concept that would render the claims significantly more than the abstract idea (see MPEP 2106.05(f)).
When the claims are considered as a whole, they do not integrate the abstract idea into a practical application; they do not confine the use of the abstract idea to a particular technology; they do not solve a problem rooted in or arising from the use of a
particular technology; they do not improve a technology by allowing the technology to
perform a function that it previously was not capable of performing; and they do not
provide any limitations beyond generally linking the use of the abstract idea to a particular technological environment and/or field of use (e.g., assessment of rehabilitative progress; MPEP 2106.05(e) and 2106.05(h)).
Hence, the claims do not include additional elements that are sufficient to amount to significantly more than the abstract idea. See MPEP 2106.05.
Conclusion: Claims are Directed to Non-statutory Subject Matter
For these reasons, the claims, when the limitations are considered individually and as a whole, are directed to a judicial exception and lack an inventive concept. Hence, the claimed invention does not constitute significantly more than the judicial exception, so the claims are rejected under 35 USC § 101 as being directed to non-statutory subject matter.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 USC §§ 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 USC § 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 USC § 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 USC § 102(b)(2)(C) for any potential 35 USC § 102(a)(2) prior art against the later invention.
Claims 1, 4, 7 and 9-13 are rejected under 35 USC § 103 as being unpatentable over Vaccaro (US 2019/0019578; effectively filed 7/17/2017; on IDS filed 4/21/2023), in view of De Bruin (US 2011/0119212; effectively filed 2/20/2008; on IDS filed 4/21/2023).
Claim 1 recites a method of quantifying rehabilitative progress via artificial intelligence, comprising steps of: accessing, by a computing device, a first dataset of input data for one or more outcome measures derived from a patient at a first point in time of rehabilitation; accessing, by the computing device, a second dataset of the input data for the one or more outcome measures derived from the patient at a second point in time of the rehabilitation; and generating, by the computing device applying the first dataset and the second dataset as inputs to a machine learning model, an output including a machine learning score that infers improvement of the patient from the first point in time to the second point in time, the machine learning model trained to map the inputs to the output to minimize a cost function defined by the machine learning model and maximize the dissimilarity between the patient between the first point in time and the second point in time.
With respect to claim 1, Vaccaro discloses a system for tracking patient recovery following an orthopedic procedure, comprising: a physical sensor configured to collect pre-procedural and post-procedural walking parameters (para. 0018); a statistical computing engine configured to use said data to implement a predictive model of a patient’s post-procedural state (para. 0021), comprising a temporal trendline of predicted outcomes at multiple post-operative timepoints (para. 0036), wherein the model can be a classifier trained via machine learning approaches (para. 0034) and outcome measures include various symptom-specific outcome scores (para. 0067); a processor having a comparator configured to compare the patient’s actual post-procedural walking parameters to the modeled post-procedural trendline and outputting the results (paras. 0022), wherein the output infers improvement of the patient from the pre-procedural time point to the post-procedural time points (paras. 0035, 0058 and 0084; Figs. 4A-5B).
Vaccaro demonstrates quantification of outcome measures in terms of percent change over a post-procedural period, and further discloses utilizing the collected data to determine and learn which factors and treatments work best for which patients and pose the greatest benefits (para. 0102; Fig. 4A). In this way, Vaccaro discloses identifying variables that maximize dissimilarity of patients between pre- and post-treatment time points.
Vaccaro does not particularly disclose training a predictive model to minimize a cost function.
De Bruin discusses a medical digital expert system that predicts a patient’s response to treatments using pre-treatment information (Abstract), and teaches steps of: generating a training dataset comprising measured patient related clinical data feature information relating to treatment response (i.e., outcome measures), extracting features from the measured data, and processing the extracted feature dataset to derive a feature data scheme or model relating feature data and treatment response (paras. 0022-24).
De Bruin further teaches that the feature data scheme or model is preferably built by means of solving a numerical optimization problem based on optimality criteria, and teaches embodiments wherein the optimality criteria includes minimizing the probability of modeling cost, i.e., minimizing a cost function (para. 0033). De Bruin discusses model building criteria, and states that basis on cost minimization is probabilistically effective (para. 0137).
With respect to claim 4, Vaccaro demonstrates quantification of outcome measures in terms of percent change over a post-procedural period, and further discloses utilizing the collected data to determine and learn which factors and treatments work best for which patients and pose the greatest benefits (para. 0102; Fig. 4A). In this way, Vaccaro discloses identifying variables that maximize dissimilarity of patients between pre- and post-treatment time points.
With respect to claim 7, Vaccaro discloses collection of pre-procedural and post-procedural walking parameters from a patient (para. 0018). Vaccaro also discusses measurement of the patient over various periods of time, including pre-procedure, at medical appointments, between medical appointments, during procedures and post-operative(para. 0082). In other words, collecting input data corresponding to phases of rehabilitation of a patient.
Vaccaro further discloses generation of a predictive model of the patient’s post-procedural state (para. 0021) comprising a temporal trendline of predicted outcomes at multiple post-operative timepoints (para. 0036), wherein the model can be a classifier trained via machine learning approaches (para. 0034); and comparison of the patient’s actual post-procedural walking parameters to the modeled post-procedural trendline (para. 0022). Performance of the disclosed predictive temporal modeling via a machine learning classifier would necessarily implement an objective function that uses an assumption of patient improvement from the first point in time to the second point in time.
With respect to claim 9, Vaccaro discloses provision by the system of evidence-based medicine including provision of physical therapy instructions to a patient (para. 0097). discloses prompting and monitoring of the patient to improve patient compliance with physical activity recommendations (para. 0089).
With respect to claim 10, Vaccaro discusses collection of various numeric outcome measures, and configuration of walking parameters and pain input as a numerical value that represents the percentage of patients that typically report better or worse walking parameters and pain (e.g., paras. 0050-52 and 0069; Fig. 4A).
Vaccaro also exemplifies collection of movement data via a GPS sensor that tracks improvements in patient movement as a result of physical therapy, and states that collected data provides an understanding of patient outcomes, progress and condition (paras. 0084, 0086 and 0098).
In this way, Vaccaro is considered to disclose collection of outcome measures including metrics configured as numeric values informative as to changes in the patient.
With respect to claim 11, Vaccaro discloses generation of a predictive model of the patient’s post-procedural state (para. 0021), wherein the model can be a classifier trained via machine learning approaches (para. 0034). Vaccaro also demonstrates quantification of outcome measures in terms of percent change over a post-procedural period, and further discloses utilizing the collected data to determine and learn which factors and treatments work best for which patients and pose the greatest benefits (para. 0102; Fig. 4A).
In this way, Vaccaro discloses training parameters of a machine earning model to find the largest difference in patients between pre- and post-treatment time points. Vaccaro does not disclose training the model using a cost function.
De Bruin teaches building a feature model by means of solving a numerical optimization problem based on minimizing the probability of modeling cost, i.e., using a cost function (para. 0033). De Bruin also discusses model building criteria, and states that basis on cost minimization is probabilistically effective (para. 0137).
With respect to claim 12, Vaccaro discloses generation of a predictive model of the patient’s post-procedural state (para. 0021), wherein the model can be a classifier trained via machine learning approaches (para. 0034). As one of ordinary skill in the art would understand, training a machine learning classifier involves modifying model parameters to solve an objective function based on training data defining outcome measures. Vaccaro does not disclose training the model to minimize a cost function.
De Bruin teaches building a feature model by means of solving a numerical optimization problem based on minimizing the probability of modeling cost, i.e., training the model to minimize a cost function (para. 0033). De Bruin also discusses model building criteria, and states that basis on cost minimization is probabilistically effective (para. 0137).
With respect to claim 13, Vaccaro discloses that new patient data is added to the training and test sets as it is collected, and machine learning approaches are applied to continuously tune the models (para. 0024). In this way, Vaccaro discloses incrementally feeding a machine learning model with additional outcome measures training data and updating the parameters.
An invention would have been obvious to one of ordinary skill in the art if some teaching in the prior art would have led that person to combine prior art reference teachings to arrive at the claimed invention. Before the effective filing date of the claimed invention, said practitioner would have implemented minimization of a cost function as a training objective, as taught by De Bruin, for the predictive modeling techniques disclosed by Vaccaro, because De Bruin teaches cost minimization is the most probabilistically effective of discussed model building criteria (para. 0137). Said practitioner would have had a reasonable expectation of success because Vaccaro and De Bruin both concern predictive modeling of patient outcomes based on pre-treatment measurements.
In this way the disclosure of Vaccaro, in view of De Bruin, makes obvious the limitations of claims 1, 4, 7 and 9-13. Thus, the claimed invention is prima facie obvious.
Claims 2-3 are rejected under 35 USC § 103 as being unpatentable over Vaccaro, in view of De Bruin, as applied to claim 1 above, and further in view of Gossage (US 2012/0328606; effectively filed 5/18/2011; on IDS filed 4/21/2023).
With respect to claim 2, Vaccaro discloses outcome measures include various symptom-specific outcome scores (para. 0067). Vaccaro does not disclose generating a first intermediate score associated with the first dataset and a second intermediate score associated with the second dataset; and computing a difference between the first intermediate score and the second intermediate score to derive the machine learning score.
De Bruin discloses prediction of three scores and their summing to produce an overall score (para. 0220). De Bruin does not disclose generating a first intermediate score associated with the first dataset and a second intermediate score associated with the second dataset; and computing a difference between the first intermediate score and the second intermediate score to derive the machine learning score.
Gossage discusses methods of diagnosing and prognosing a pulmonary disease or disorder including predicting a prognosis relating to how the condition will respond to therapy, (Abstract and para. 0180). Gossage teaches embodiments wherein the diagnostic score is the variance between two calculated intermediate scores, e.g., a numeric value calculated by subtracting a first score from a second score (paras. 0026, 0096 and 0126), and the method is implemented via a neural network model (paras. 0177-78).
With respect to claim 3, Gossage teaches a method of monitoring the efficacy of a therapy, comprising: calculating a first score from measurements obtained before administration of a therapy; calculating a second score from measurements obtained after administration of a therapy; and comparing the two scores, wherein a first score greater than the second indicates efficacy (para. 0016). In other words, a greater difference between the scores (i.e., distribution) reflects greater improvement of the patient made during the treatment.
An invention would have been obvious to one of ordinary skill in the art if some teaching in the prior art would have led that person to combine prior art reference teachings to arrive at the claimed invention. Before the effective filing date of the claimed invention, said practitioner would have implemented generating and computing a difference between first and second intermediate scores, as taught by Gossage, in combination with the patient recovery analysis techniques taught by Vaccaro, in view of De Bruin, because Gossage indicates suitability of this statistical technique for predictive modeling of patient response via machine learning approaches. Said practitioner would have had a reasonable expectation of success because Vaccaro and Gossage both concern predictive modeling of patient response via machine learning approaches.
In this way the disclosure of Vaccaro, in view of De Bruin and Gossage, makes obvious the limitations of claims 2-3. Thus, the claimed invention is prima facie obvious.
Claim 5 is rejected under 35 USC § 103 as being unpatentable over Vaccaro, in view of De Bruin and Gossage, as applied to claims 1-2 above, and further in view of Bhagwat (PLoS Computational Biology 14(9): e1006376, 25 pages; published 9/14/2018).
With respect to claim 5, Vaccaro discloses generation of a predictive model of a patient’s post-procedural state (para. 0021), wherein the model can be a classifier trained via machine learning approaches (para. 0034). Vaccaro does not particularly disclose generation of a Siamese neural network model.
De Bruin teaches that the feature data scheme or model may be determined using an neural network (para. 0038), but does not particularly teach employment of a Siamese neural network model.
Bhagwat discusses a computational framework for modeling and predicting symptom trajectories of patients with Alzheimer’s disease, based on measured data, using a longitudinal Siamese neural network (pg. 1, Abstract). Bhagwat teaches that their Siamese neural network architecture processes input measurements from the same subject at two time points and produces a ‘difference embedding’, representing change in the measurements between the time points, which is then concatenated with clinical scores to predict a trajectory (pg. 7, Fig. 3 and para. 3 – pg. 8, para. 1).
Bhagwat presents findings that their longitudinal Siamese neural network architecture (LSN) exhibits highly accurate performance, and outperforms all four compared reference models in accurately predicting patient trajectory classes based on features from two timepoints (pg. 1, Abstract; pg. 10, para. 4).
An invention would have been obvious to one of ordinary skill in the art if some teaching in the prior art would have led that person to combine prior art reference teachings to arrive at the claimed invention. Before the effective filing date of the claimed invention, said practitioner would have implemented a Siamese neural network, as taught by Bhagwat, to model patient recovery as disclosed by Vaccaro, in view of De Bruin and Gossage, because Bhagwat teaches that their Siamese neural network architecture exhibits highly accurate performance, and outperforms all four compared reference models in accurately predicting patient trajectory classes based on features from two timepoints (pg. 1, Abstract; pg. 10, para. 4). Said practitioner would have had a reasonable expectation of success because Vaccaro and Bhagwat both concern predictive modeling of patient outcome trajectories, based on pre-treatment measurements, via machine learning approaches.
In this way the disclosure of Vaccaro, in view of De Bruin, Gossage and Bhagwat, makes obvious the limitations of claim 5. Thus, the claimed invention is prima facie obvious.
Claim 6 is rejected under 35 USC § 103 as being unpatentable over Vaccaro, in view of De Bruin, as applied to claim 1 above, and further in view of Zhang (Statistics in Biopharmaceutical Research 2(2): 292-299; published 2010).
With respect to claim 6, Vaccaro discloses training a predictive model via machine learning approaches (para. 0034). Training a model via machine learning approaches necessarily involves training the model to solve an objective function.
Vaccaro also demonstrates quantification of outcome measures in terms of percent change over a post-procedural period, and discloses utilizing the collected data to determine and learn which factors and treatments work best for which patients and pose the greatest benefits (para. 0102; Fig. 4A). In this way, Vaccaro discloses identifying variables that maximize dissimilarity of patients between pre- and post-treatment time points. One of ordinary skill in the art would find it obvious to formalize this contemplated goal as a model training objective. However, Vaccaro does not disclose implementing the recited cost function for this purpose.
De Bruin teaches building a feature model by means of solving a numerical optimization problem based on minimizing the probability of modeling cost (para. 0033). De Bruin does not teach the recited cost function.
Zhang discusses strictly standardized mean difference (SSMD), a statistical metric that quantifies differences (effect sizes) between two groups, defined as the ratio of mean to standard deviation of the difference between the two groups (pg. 292, Abstract; pg. 294, l. column). Zhang discusses numerous statistical advantages that SSMD has over traditional comparison metrics, particularly noting that SSMD is robust to arbitrarily increasing sample size due to incorporation of variability information (unlike t-statistics and p-values), is applicable to measure the magnitude of difference between non-independent groups (unlike Cohen’s d), and has clear and meaningful probability interpretations (pg. 294, r. column – pg. 297, l. column).
SSMD as described can be represented as mean (s2-s1) / std (s2-s1).
As one of ordinary skill in the art would understand, maximization of a given parameter is conventionally implemented as a machine learning objective by training to the model to minimize the inverse (-) of that parameter. Thus, Zhang provides an advantageous dissimilarity metric that one of ordinary skill in the art, wanting to implement dissimilarity maximization between paired time series data an objective function, would formalize as Jmin(s1,s2) = - mean (s2-s1) / std (s2-s1).
An invention would have been obvious to one of ordinary skill in the art if some teaching in the prior art would have led that person to combine prior art reference teachings to arrive at the claimed invention. Before the effective filing date of the claimed invention, said practitioner would have implemented SSMD, as taught by Zhang, with the patient recovery analysis techniques taught by Vaccaro, in view of De Bruin, because Zhang teaches that SSMD has numerous statistical advantages over traditional comparison metrics including suitability for paired data (pg. 294, r. column – pg. 297, l. column). Said practitioner would have had a reasonable expectation of success because Vaccaro and Zhang both discuss statistical comparison of paired data.
In this way the disclosure of Vaccaro, in view of De Bruin and Zhang, makes obvious the limitations of claim 6. Thus, the claimed invention is prima facie obvious.
Claim 8 is rejected under 35 USC § 103 as being unpatentable over Vaccaro, in view of De Bruin, as applied to claim 1 above, and further in view of Torres (US 2017/0344706; effectively filed 11/11/2011; on IDS filed 4/21/2023).
With respect to claim 8, Vaccaro states that normalized universal standards for subjective patient data are hard to establish (para. 0076). Vaccaro does not disclose normalization of input data wherein each outcome measure is rescaled to a range [0,1] using the minimum and maximum values for each outcome measure.
De Bruin discusses unity normalization of weight vectors as a constituent of partial least squares regression (para. 0170). De Bruin does not teach normalization of input data wherein each outcome measure is rescaled to a range [0,1] using the minimum and maximum values for each outcome measure.
Torres discusses systems and methods for analyzing a neurological disorder in a subject and/or data compression, which can be used in a medical context, e.g., to facilitate diagnosis and treatment (Abstract; para. 0005). Torres further discusses application to assessing patient response to treatment, e.g., improvement in the condition of the patient, and predicting the probable course and outcome of a disease/disorder or likelihood of recovery (paras. 0043-45).
Torres particularly teaches techniques for normalizing, by a computing device, of raw sensor data specifying a bodily rhythm created by a human subject to define a waveform capturing rates of changes in sensor data (para. 0006). Torres exemplifies a unity-based normalization technique wherein each data point is rescaled to a range [0,1] using the minimum and maximum values for each measured parameter (para. 0071).
Torres states that data normalization is very important when dealing with parameters of different units and scales, and is performed to standardize the different resolution, scales or units of time series waveforms defined by different types of data (para. 0069-70).
An invention would have been obvious to one of ordinary skill in the art if some teaching in the prior art would have led that person to combine prior art reference teachings to arrive at the claimed invention. Before the effective filing date of the claimed invention, said practitioner would have implemented unity normalization, as taught by Torres, with the patient recovery analysis techniques taught by Vaccaro, in view of De Bruin, because Torres indicates that such normalization allows for collective analysis of measured data having different resolutions, scales and units (para. 0069-70). Thus, Torres indicates that normalizing the input data of Vaccaro in this way would allow for application of the disclosed analytical methodology to a plurality of different measured parameters. Said practitioner would have had a reasonable expectation of success because Vaccaro and Torres both concern computerized analysis of changes in bodily movement parameters collected from human subjects via sensors.
In this way the disclosure of Vaccaro, in view of De Bruin and Torres, makes obvious the limitations of claim 8. Thus, the claimed invention is prima facie obvious.
Claims 14-15 are rejected under 35 USC § 103 as being unpatentable over Vaccaro, in view of De Bruin, as applied to claim 1 above, and further in view of Jarrett (IEEE Journal of Biomedical and Health Informatics 24(2): 424-436; published February 2020).
With respect to claims 14-15, Vaccaro discloses training a predictive model via machine learning approaches (para. 0034). Training a model via machine learning approaches necessarily involves rendering the input data as a feature matrix. Vaccaro also discloses imputation of missing predictor data (para. 0034).
However, Vaccaro does not disclose appending an additional column to the feature matrix to serve as a mask for identifying missing data; assigning a value of 1 to reflect population of a value for a given outcome measure; and assigning a value of 0 to reflect missing data.
De Bruin teaches transforming input data into a discriminative feature vector (paras. 045-47), i.e., a feature matrix. De Bruin also discusses identifying missing data (para. 0082), but does not teach appending an additional column to the feature matrix to serve as a mask for identifying missing data; assigning a value of 1 to reflect population of a value for a given outcome measure; and assigning a value of 0 to reflect missing data.
Jarrett discusses prediction of disease trajectories via MissingNet, a missingness-aware neural network architecture (pg. 424, Abstract). Jarrett teaches implementation of a binary mask of missing-value indicators, which take a value of 1 if and only if corresponding data is missing from the feature matrix (pg. 427, l. column). Where data is populated in the feature matrix, the described binary mask would take a value of 0. Assignment of ‘1’ to reflect missing data and ‘0’ to reflect population of data (as disclosed) is not considered patentably distinct from assignment of ‘1’ to reflect population of data and ‘0’ to reflect missing data (as claimed).
Jarrett teaches that prior methods rely on the assumption that the timing and frequency of covariate measurements is uninformative, while MissingNet learns correlations between patterns of data missingness and patient progression which may be additionally informative (pg. 427, l. column).
An invention would have been obvious to one of ordinary skill in the art if some teaching in the prior art would have led that person to combine prior art reference teachings to arrive at the claimed invention. Before the effective filing date of the claimed invention, said practitioner would have implemented analysis of a binary mask of missingness indicators alongside the feature matrix, as taught by Jarrett, using the patient recovery analysis techniques taught by Vaccaro, in view of De Bruin, because Jarrett teaches that patterns of missingness in clinical time-series data may be informative features and can be considered by the model in the form of an additional binary missingness mask. Said practitioner would have had a reasonable expectation of success because Vaccaro and Jarrett both concern predictive modeling of patient trajectories, via machine learning approaches, based on clinical time-series data.
In this way the disclosure of Vaccaro, in view of De Bruin and Jarrett, makes obvious the limitations of claims 14-15. Thus, the claimed invention is prima facie obvious.
Conclusion
At this point in prosecution, no claim is allowed.
The following prior art, made of record and not relied upon, is considered pertinent to applicant's disclosure:
Zhang et al (Proceedings of the 2016 IEEE International Conference on Acoustics, Speech and Signal Processing 2832-2836; published 5/19/2016) discloses a Siamese neural network-based gait recognition framework that robustly extracts model features despite large intra-class variations in walking parameters from the same person (pg. 2832, Abstract – r. column).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Theodore C. Striegel whose telephone number is (571)272-1860. The examiner can normally be reached Mon-Fri 12pm-8pm ET.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Olivia M. Wise can be reached at (571)272-2249. 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.
/T.C.S./Examiner, Art Unit 1685
/JESSE P FRUMKIN/Primary Examiner, Art Unit 1685 July 11, 2026
Prosecution Insights