DETAILED ACTION
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 .
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.
Notice of Amendment
In response to the amendment(s) filed on 9/17/26, amended claim(s) 1, 8-10, and 17-20, canceled claim(s) 4-5 and 13-14, and new claim(s) 21 is/are acknowledged. The following new and/or reiterated ground(s) of rejection is/are set forth:
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claim(s) 1-3, 6-12, and 15-21 is/are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
For claim 1, the claim language “computing a forward pass through the machine learning model for the individual windows, wherein the machine learning model is trained to classify, based on the acceleration information included in the individual windows, the individual windows as goal-directed movement (GDM) windows reflecting GDMs performed by the user or as non-GDM windows, the non-GDM windows including windows reflecting passive actions of the limbs and windows corresponding to task-free periods during which the user does not perform a GDM,” and “obtaining, via the machine learning model, labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows” does not appear to be described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for pre-AIA the inventor(s), at the time the application was filed, had possession of the claimed invention. A claim may lack written description when the specification does not disclose the computer and the algorithm (i.e., the necessary steps and/or flowcharts) that perform the claimed function in sufficient detail such that one of ordinary skill in the art can reasonably conclude that the inventor invented the claimed subject matter. See MPEP 2161.01(I). Here, the claim recites the function of computing a forward pass through the machine learning model for the individual windows, wherein the machine learning model is trained to classify, based on the acceleration information included in the individual windows, the individual windows as goal-directed movement (GDM) windows reflecting GDMs performed by the user or as non-GDM windows, the non-GDM windows including windows reflecting passive actions of the limbs and windows corresponding to task-free periods during which the user does not perform a GDM; and obtaining, via the machine learning model, labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows, but the specification never discloses the necessary steps and/or flowcharts of how this occurs. The term “machine learning model” is treated as a black box and the specification does not describe the specifics of how to achieve the above-recited function(s) with this algorithm. For example, how many and what types of layers are there? How is the data propagated? What logics are programmed to help the machine learning algorithm make a decision? Is the training supervised or unsupervised? What are the weightings? Are other training concepts used such as regression? What strategies are employed to minimize the loss function? How is dimensionality reduced? How is the clustering problem solved? How are relationships between datasets discovered and/or established? It is not enough that a skilled artisan could devise a way to accomplish the function because this is not relevant to the issue of whether the inventor has shown possession of the claimed invention. See MPEP 2161.01(I). Therefore, adequate disclosure is needed.
For claim 10, the claim language “computing a forward pass through the machine learning model for the individual windows, wherein the machine learning model is trained to classify, based on the acceleration information included in the individual windows, the individual windows as goal-directed movement (GDM) windows reflecting GDMs performed by the user or as non-GDM windows, the non-GDM windows including windows reflecting passive actions of the limbs and windows corresponding to task-free periods during which the user does not perform a GDM,” and “obtaining, via the machine learning model, labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows” does not appear to be described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for pre-AIA the inventor(s), at the time the application was filed, had possession of the claimed invention. A claim may lack written description when the specification does not disclose the computer and the algorithm (i.e., the necessary steps and/or flowcharts) that perform the claimed function in sufficient detail such that one of ordinary skill in the art can reasonably conclude that the inventor invented the claimed subject matter. See MPEP 2161.01(I). Here, the claim recites the function of computing a forward pass through the machine learning model for the individual windows, wherein the machine learning model is trained to classify, based on the acceleration information included in the individual windows, the individual windows as goal-directed movement (GDM) windows reflecting GDMs performed by the user or as non-GDM windows, the non-GDM windows including windows reflecting passive actions of the limbs and windows corresponding to task-free periods during which the user does not perform a GDM; and obtaining, via the machine learning model, labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows, but the specification never discloses the necessary steps and/or flowcharts of how this occurs. The term “machine learning model” is treated as a black box and the specification does not describe the specifics of how to achieve the above-recited function(s) with this algorithm. For example, how many and what types of layers are there? How is the data propagated? What logics are programmed to help the machine learning algorithm make a decision? Is the training supervised or unsupervised? What are the weightings? Are other training concepts used such as regression? What strategies are employed to minimize the loss function? How is dimensionality reduced? How is the clustering problem solved? How are relationships between datasets discovered and/or established? It is not enough that a skilled artisan could devise a way to accomplish the function because this is not relevant to the issue of whether the inventor has shown possession of the claimed invention. See MPEP 2161.01(I). Therefore, adequate disclosure is needed.
For claim 18, the claim language “computing a forward pass through the machine learning model for the individual windows, wherein the machine learning model is trained to classify, based on the acceleration information included in the individual windows, the individual windows as goal-directed movement (GDM) windows reflecting GDMs performed by the user or as non-GDM windows, the non-GDM windows including windows reflecting passive actions of the limbs and windows corresponding to task-free periods during which the user does not perform a GDM,” and “obtaining, via the machine learning model, labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows” does not appear to be described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for pre-AIA the inventor(s), at the time the application was filed, had possession of the claimed invention. A claim may lack written description when the specification does not disclose the computer and the algorithm (i.e., the necessary steps and/or flowcharts) that perform the claimed function in sufficient detail such that one of ordinary skill in the art can reasonably conclude that the inventor invented the claimed subject matter. See MPEP 2161.01(I). Here, the claim recites the function of computing a forward pass through the machine learning model for the individual windows, wherein the machine learning model is trained to classify, based on the acceleration information included in the individual windows, the individual windows as goal-directed movement (GDM) windows reflecting GDMs performed by the user or as non-GDM windows, the non-GDM windows including windows reflecting passive actions of the limbs and windows corresponding to task-free periods during which the user does not perform a GDM; and obtaining, via the machine learning model, labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows, but the specification never discloses the necessary steps and/or flowcharts of how this occurs. The term “machine learning model” is treated as a black box and the specification does not describe the specifics of how to achieve the above-recited function(s) with this algorithm. For example, how many and what types of layers are there? How is the data propagated? What logics are programmed to help the machine learning algorithm make a decision? Is the training supervised or unsupervised? What are the weightings? Are other training concepts used such as regression? What strategies are employed to minimize the loss function? How is dimensionality reduced? How is the clustering problem solved? How are relationships between datasets discovered and/or established? It is not enough that a skilled artisan could devise a way to accomplish the function because this is not relevant to the issue of whether the inventor has shown possession of the claimed invention. See MPEP 2161.01(I). Therefore, adequate disclosure is needed.
Dependent claim(s) 2-3, 6-9, 11-12, 15-17, and 19-21 the deficiencies of independent claim(s) 1, 10, and 18. Thus, claim(s) 1-3, 6-12, and 15-21 is/are rejected under 35 U.S.C. 112(a).
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.
Claim(s) 1-3, 6-12, and 15-21 is/are rejected under 35 U.S.C. 101 because the claimed invention, considering all claim elements both individually and in combination as a whole, do not amount to significantly more than a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea).
Claim 1 is a claim to a process, machine, manufacture, or composition of matter and therefore meets one of the categorical limitations of 35 U.S.C. 101. However, claim 1 meets the first prong of the step 2A analysis because it is directed to a/an abstract idea, as evidenced by the claim language of “obtaining sensor data from individual devices worn on individual limbs of a user, the devices generating sensor data indicative of, at least, acceleration information associated with the limbs,” “adjusting the obtained sensor data for input into a machine learning model, wherein the machine learning model is a deep learning model, wherein adjusting the obtained sensor data comprises separating the obtained sensor data into individual windows of sensor data, and wherein the individual windows include acceleration information associated with a threshold amount of time,” “computing a forward pass through the machine learning model for the individual windows, wherein the machine learning model is trained to classify, based on the acceleration information included in the individual windows, the individual windows as goal-directed movement (GDM) windows reflecting GDMs performed by the user or as non-GDM windows, the non-GDM windows including windows reflecting passive actions of the limbs and windows corresponding to task-free periods during which the user does not perform a GDM,” and “obtaining, via the machine learning model, labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows.” This claim language, under the broadest, reasonable interpretation, encompasses subject matter that may be performed by a human using mental steps or with pen and paper that can involve basic critical thinking, which are types of activities that have been found by the courts to represents abstract ideas (i.e., the mental comparison in Ambry Genetics, or the diagnosing an abnormal condition by performing clinical tests and thinking about the results in Grams). The claim language also meets prong 2 of the step 2A analysis because the above-recited claim language does not integrate the abstract idea into a practical application. That is, there appears to be no tangible improvement in a technology, effect of a particular treatment or prophylaxis, a particular machine or manufacture that is integrated, or transformation/reduction of a particular article to a different state or thing as a result of this claimed subject matter. As a result, step 2A is satisfied and the second step, step 2B, must be considered.
With regard to the second step, the claim does not appear to recite additional elements that amount to significantly more. The additional elements are “one or more processors.” However, these elements are not “significantly more” because they are well-known, routine, and/or conventional as evidenced by Bilski and Alice, which have held that generic computer structures do not otherwise transform a patent-ineligible claim into a patent-eligible one. Therefore, these elements do not add significantly more and thus the claim as a whole does not amount to significantly more than a judicial exception.
Additionally, the ordered combination of elements do not add anything significantly more to the claimed subject matter. Specifically, the ordered combination of elements do not have any function that is not already supplied by each element individually. That is, the whole is not greater than the sum of its parts.
In view of the above, independent claim 1 fails to recite patent-eligible subject matter under 35 U.S.C. 101. Independent claim(s) 10 and 18 fail to recite patent-eligible subject matter for similar, if not the exact same, reasoning as that of independent claim 1. Independent claims 10 and 19 differ in that they recite “non-transitory computer storage media.” However, this claimed subject matter is also not “significantly more” because it is well-known, routine, and/or conventional as evidenced by Bilski and Alice, which have held that generic computer structures do not otherwise transform a patent-ineligible claim into a patent-eligible one. Dependent claim(s) 2-3, 6-9, 11-12, 15-17, and 19-21 the deficiencies of independent claim(s) 1, 10, and 18 by merely reciting additional abstract ideas and/or further limitations on abstract ideas already recited. Thus, claim(s) 1-3, 6-12, and 15-21 is/are rejected under 35 U.S.C. 101.
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.
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 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claim(s) 1-3, 8-12, 17-18, and 20-21 is/are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Patent Application Publication No. 2021/0275109 to Sheth et al. (hereinafter “Sheth”) in view of U.S. Patent Application Publication No. 2020/0375501 to Motamedi et al. (hereinafter “Motamedi”) and U.S. Patent Application Publication No. 2007/0185697 to Tan et al. (hereinafter “Tan”).
For claim 1, Sheth discloses a method (Abstract) implemented by a system of one or more processors (14) (Figs. 1-2 and 6), and the method comprising:
obtaining sensor data from individual devices (12) (Fig. 1) (para [0043]) worn on individual limbs of a user (as can be seen in Fig. 1), the devices generating sensor data indicative of, at least, acceleration information associated with the limbs (para [0059]);
adjusting the obtained sensor data (para [0061]) for input into a model (para [0100]), wherein the adjusting the obtained sensor data comprises separating the obtained sensor data into individual windows of sensor data (para [0122]), and wherein the individual windows include acceleration information associated with a threshold amount of time (para [0122]);
computing information through the model for the individual windows (para [0122]) (also see para [0101]-[0105]), wherein the model is configured to classify (Fig. 10) (para [0108]), based on the acceleration information included in the individual windows (para [0122]), between goal-directed movement and non-goal directed movements (Fig. 10) (para [0108]) (also see Fig. 11, which shows the “diffusion maps” (see para [0108]) and how para [0122] discusses “determining a specific limb activity signature for each individual based on diffusion map analysis,” and then later discusses deviations from these maps) (also see more generally para [0102]-[0122]); and
obtaining, via the model, an output for the individual windows (para [0102]-[0103] and [0105], but see more generally [0101]-[0108] for context) (alternatively, para [0123], but see more generally [0109]-[0123] for context).
Sheth does not expressly disclose that the model is a machine learning model, wherein the machine learning model is a deep learning model; and computing a forward pass through the machine learning model.
However, Motamedi teaches that the model is a machine learning model (para [0056]), wherein the machine learning model is a deep learning model (“convolutional neural network,” para [0043]); computing a forward pass through the machine learning model (para [0056]); wherein the machine learning model is trained to output information indicative of goal-directed movements performed by the user (para [0061]).
It would have been obvious to a skilled artisan to modify Sheth such that the model is a machine learning model, wherein the machine learning model is a deep learning model; and computing a forward pass through the machine learning model, in view of the teachings of Motamedi, for the obvious advantage of providing a more advanced algorithm that can handle complex processing with big data (such as data that involves full body/limb movement over time) to get more accurate classification of abnormalities, and Sheth wants to accurately classify abnormalities of human motion.
Sheth ad Motamedi do not expressly disclose that the machine learning model is trained to classify the individual windows, that the non-GDM windows include windows reflecting passive actions of the limbs and windows corresponding to task-free periods during the user does not perform a GDM; and that the output for the individual windows is labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows.
However, Tan teaches machine learning model is trained to classify individual windows (para [0052]), that the non-GDM windows include windows reflecting passive actions of the limbs and windows corresponding to task-free periods during the user does not perform a GDM (para [0052]) (also see para [0037] and [0049]); and that the output for the individual windows is labels for the individual windows (para [0052]), the labels distinguishing the GDM windows from the non-GDM windows (para [0052]) (also see para [0027]-[0028] for a further description of different tasks).
It would have been obvious to a skilled artisan to modify Sheth such that that the machine learning model is trained to classify the individual windows, that the non-GDM windows include windows reflecting passive actions of the limbs and windows corresponding to task-free periods during the user does not perform a GDM; and that the output for the individual windows is labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows, in view of the teachings of Tan, for the obvious advantage of identifying particular time periods as active time periods or resting time periods, which helps a health-care provider distinguish between intended movement and non-intended or lack of movement, which improves signal-to-noise for the data set.
For claim 2, Sheth further discloses wherein the sensor data is generated via individual inertial measurement units included in individual devices (para [0059]), and wherein the acceleration information reflects tri-axis acceleration data (para [0059]).
For claim 3, Sheth further discloses wherein the individual devices include two devices worn on respective wrists of the user (para [0059]).
For claim 8, Sheth, as modified, further discloses wherein the labels further identify whether the individuals windows reflect unimanual actions, bimanual actions, passive actions, or are task-free (see para [0052] of Tan).
For claim 9, Sheth, as modified, further discloses causing presentation of an interactive user interface comprising summary information associated with the information indicative of labels (see para [0028] of Tan).
For claim 10, Sheth discloses a system Abstract) comprising one or more processors (14) (Figs. 1-2 and 6) and non-transitory computer storage media storing instructions that when executed by the one or more processors, cause the one or more processors to perform operations (para [0050]) comprising:
obtaining sensor data from individual devices (12) (Fig. 1) (para [0043]) worn on individual limbs of a user (as can be seen in Fig. 1), the devices generating sensor data indicative of, at least, acceleration information associated with the limbs (para [0059]);
adjusting the obtained sensor data (para [0061]) for input into a model (para [0100]), wherein the adjusting the obtained sensor data comprises separating the obtained sensor data into individual windows of sensor data (para [0122]), and wherein the individual windows include acceleration information associated with a threshold amount of time (para [0122]);
computing information through the model for the individual windows (para [0122]) (also see para [0101]-[0105]), wherein the model is configured to classify (Fig. 10) (para [0108]), based on the acceleration information included in the individual windows (para [0122]), between goal-directed movement and non-goal directed movements (Fig. 10) (para [0108]) (also see Fig. 11, which shows the “diffusion maps” (see para [0108]) and how para [0122] discusses “determining a specific limb activity signature for each individual based on diffusion map analysis,” and then later discusses deviations from these maps) (also see more generally para [0102]-[0122]); and
obtaining, via the model, an output for the individual windows (para [0102]-[0103] and [0105], but see more generally [0101]-[0108] for context) (alternatively, para [0123], but see more generally [0109]-[0123] for context).
Sheth does not expressly disclose that the model is a machine learning model, wherein the machine learning model is a deep learning model; and computing a forward pass through the machine learning model.
However, Motamedi teaches that the model is a machine learning model (para [0056]), wherein the machine learning model is a deep learning model (“convolutional neural network,” para [0043]); computing a forward pass through the machine learning model (para [0056]); wherein the machine learning model is trained to output information indicative of goal-directed movements performed by the user (para [0061]).
It would have been obvious to a skilled artisan to modify Sheth such that the model is a machine learning model, wherein the machine learning model is a deep learning model; and computing a forward pass through the machine learning model, in view of the teachings of Motamedi, for the obvious advantage of providing a more advanced algorithm that can handle complex processing with big data (such as data that involves full body/limb movement over time) to get more accurate classification of abnormalities, and Sheth wants to accurately classify abnormalities of human motion.
Sheth ad Motamedi do not expressly disclose that the machine learning model is trained to classify the individual windows, that the non-GDM windows include windows reflecting passive actions of the limbs and windows corresponding to task-free periods during the user does not perform a GDM; and that the output for the individual windows is labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows.
However, Tan teaches machine learning model is trained to classify individual windows (para [0052]), that the non-GDM windows include windows reflecting passive actions of the limbs and windows corresponding to task-free periods during the user does not perform a GDM (para [0052]) (also see para [0037] and [0049]); and that the output for the individual windows is labels for the individual windows (para [0052]), the labels distinguishing the GDM windows from the non-GDM windows (para [0052]) (also see para [0027]-[0028] for a further description of different tasks).
It would have been obvious to a skilled artisan to modify Sheth such that that the machine learning model is trained to classify the individual windows, that the non-GDM windows include windows reflecting passive actions of the limbs and windows corresponding to task-free periods during the user does not perform a GDM; and that the output for the individual windows is labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows, in view of the teachings of Tan, for the obvious advantage of identifying particular time periods as active time periods or resting time periods, which helps a health-care provider distinguish between intended movement and non-intended or lack of movement, which improves signal-to-noise for the data set.
For claim 11, Sheth further discloses wherein the sensor data is generated via individual inertial measurement units included in individual devices (para [0059]), and wherein the acceleration information reflects tri-axis acceleration data (para [0059]).
For claim 12, Sheth further discloses wherein the individual devices include two devices worn on respective wrists of the user (para [0059]).
For claim 17, Sheth, as modified, further discloses wherein the labels further identify whether the individuals windows reflect unimanual actions, bimanual actions, passive actions, or are task-free (see para [0052] of Tan).
For claim 18, Sheth discloses a non-transitory computer storage medium storing instructions that when executed by a system of one or more processors (14) (Figs. 1-2 and 6), cause the processors to perform operations (para [0050]) comprising:
obtaining sensor data from individual devices (12) (Fig. 1) (para [0043]) worn on individual limbs of a user (as can be seen in Fig. 1), the devices generating sensor data indicative of, at least, acceleration information associated with the limbs (para [0059]);
adjusting the obtained sensor data (para [0061]) for input into a model (para [0100]), wherein the adjusting the obtained sensor data comprises separating the obtained sensor data into individual windows of sensor data (para [0122]), and wherein the individual windows include acceleration information associated with a threshold amount of time (para [0122]);
computing information through the model for the individual windows (para [0122]) (also see para [0101]-[0105]), wherein the model is configured to classify (Fig. 10) (para [0108]), based on the acceleration information included in the individual windows (para [0122]), between goal-directed movement and non-goal directed movements (Fig. 10) (para [0108]) (also see Fig. 11, which shows the “diffusion maps” (see para [0108]) and how para [0122] discusses “determining a specific limb activity signature for each individual based on diffusion map analysis,” and then later discusses deviations from these maps) (also see more generally para [0102]-[0122]); and
obtaining, via the model, an output for the individual windows (para [0102]-[0103] and [0105], but see more generally [0101]-[0108] for context) (alternatively, para [0123], but see more generally [0109]-[0123] for context).
Sheth does not expressly disclose that the model is a machine learning model, wherein the machine learning model is a deep learning model; and computing a forward pass through the machine learning model.
However, Motamedi teaches that the model is a machine learning model (para [0056]), wherein the machine learning model is a deep learning model (“convolutional neural network,” para [0043]); computing a forward pass through the machine learning model (para [0056]); wherein the machine learning model is trained to output information indicative of goal-directed movements performed by the user (para [0061]).
It would have been obvious to a skilled artisan to modify Sheth such that the model is a machine learning model, wherein the machine learning model is a deep learning model; and computing a forward pass through the machine learning model, in view of the teachings of Motamedi, for the obvious advantage of providing a more advanced algorithm that can handle complex processing with big data (such as data that involves full body/limb movement over time) to get more accurate classification of abnormalities, and Sheth wants to accurately classify abnormalities of human motion.
Sheth ad Motamedi do not expressly disclose that the machine learning model is trained to classify the individual windows, that the non-GDM windows include windows reflecting passive actions of the limbs and windows corresponding to task-free periods during the user does not perform a GDM; and that the output for the individual windows is labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows.
However, Tan teaches machine learning model is trained to classify individual windows (para [0052]), that the non-GDM windows include windows reflecting passive actions of the limbs and windows corresponding to task-free periods during the user does not perform a GDM (para [0052]) (also see para [0037] and [0049]); and that the output for the individual windows is labels for the individual windows (para [0052]), the labels distinguishing the GDM windows from the non-GDM windows (para [0052]) (also see para [0027]-[0028] for a further description of different tasks).
It would have been obvious to a skilled artisan to modify Sheth such that that the machine learning model is trained to classify the individual windows, that the non-GDM windows include windows reflecting passive actions of the limbs and windows corresponding to task-free periods during the user does not perform a GDM; and that the output for the individual windows is labels for the individual windows, the labels distinguishing the GDM windows from the non-GDM windows, in view of the teachings of Tan, for the obvious advantage of identifying particular time periods as active time periods or resting time periods, which helps a health-care provider distinguish between intended movement and non-intended or lack of movement, which improves signal-to-noise for the data set.
For claim 20, Sheth, as modified, further discloses wherein the labels further identify whether the individuals windows reflect unimanual actions, bimanual actions, passive actions, or are task-free (see para [0052] of Tan).
For claim 21, Sheth, as modified, further discloses wherein the individual devices include a first device worn on a first limb of the user and a second device worn on a second limb of the user (para [0055] and [0059] of Sheth), and wherein the windows of the sensor data obtained from the first device are classified separately from the windows of sensor data obtained from the second device (para [0107]-[0108] of Sheth).
Claim(s) 6, 15, and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sheth in view of Motamedi, and further in view of U.S. Patent Application Publication No. 2021/0298648 to Hefner et al. (hereinafter “Hefner”).
For claim 6, Sheth and Motamedi do not expressly disclose wherein the machine learning model is a transformer-based deep learning model.
However, Hefner teaches wherein the machine learning model is a transformer-based deep learning model (para [0060]).
It would have been obvious to a skilled artisan to modify Sheth wherein the machine learning model is a transformer-based deep learning model, in view of the teachings of Hefner, because such a modification is a simple substitution of the type of model that would lead to the predictable result of analyzing the data to arrive at an output for diagnosis.
For claim 15, Sheth and Motamedi do not expressly disclose wherein the machine learning model is a transformer-based deep learning model.
However, Hefner teaches wherein the machine learning model is a transformer-based deep learning model (para [0060]).
It would have been obvious to a skilled artisan to modify Sheth wherein the machine learning model is a transformer-based deep learning model, in view of the teachings of Hefner, because such a modification is a simple substitution of the type of model that would lead to the predictable result of analyzing the data to arrive at an output for diagnosis.
For claim 19, Sheth and Motamedi do not expressly disclose wherein the machine learning model is a transformer-based deep learning model or an explainable convolutional neural network.
However, Hefner teaches wherein the machine learning model is a transformer-based deep learning model (para [0060]).
It would have been obvious to a skilled artisan to modify Sheth wherein the machine learning model is a transformer-based deep learning model or an explainable convolutional neural network, in view of the teachings of Hefner, because such a modification is a simple substitution of the type of model that would lead to the predictable result of analyzing the data to arrive at an output for diagnosis.
Claim(s) 7 and 16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sheth in view of Motamedi, and further in view of U.S. Patent Application Publication No. 2024/0290494 to Kwon et al. (hereinafter “Kwon”).
For claim 7, Sheth, as modified, further discloses wherein the machine learning model is an convolutional neural network (see para [0043] of Motamedi).
Sheth and Motamedi do not expressly disclose wherein the convolutional neural network is an explainable convolutional neural network.
However, Kwon teaches an explainable convolutional neural network (claims 1 and 14).
It would have been obvious to a skilled artisan to modify Sheth wherein the convolutional neural network is an explainable convolutional neural network, in view of the teachings of Kwon, because such a modification is the simple substitution of one model for another model that would lead to the predictable result of analyzing the data to arrive at an output for diagnosis.
For claim 16, Sheth, as modified, further discloses wherein the machine learning model is an convolutional neural network (see para [0043] of Motamedi).
Sheth and Motamedi do not expressly disclose wherein the convolutional neural network is an explainable convolutional neural network.
However, Kwon teaches an explainable convolutional neural network (claims 1 and 14).
It would have been obvious to a skilled artisan to modify Sheth wherein the convolutional neural network is an explainable convolutional neural network, in view of the teachings of Kwon, because such a modification is the simple substitution of one model for another model that would lead to the predictable result of analyzing the data to arrive at an output for diagnosis.
Response to Arguments
Applicant’s arguments filed 9/17/26 have been fully considered.
With respect to the claim objection, Applicant’s amendments and arguments are persuasive and thus the objection is withdrawn.
With respect to the 112 rejections, the written description requirement used in the above rejection is pulled directly from MPEP 2161.01(I) and is therefore not “heightened” (i.e., the examiner is not making up or embellishing the requirement). Of course, every application is different and every technology area is different. So the requirement needs to applied judiciously to the art that is being examined. Here, the technology area includes machine learning and artificial intelligence and so the written description requirement needs to be applied to this technology. The disclosure of the “weightings” is just an example of one question that could be answered that would be favorable towards the written description being met in this technology area. However, such a question (and all of the questions listed in the rejection) are just concrete examples meant to illustrate and guide Applicant to come up with some answers that could potentially help overcome the rejection. This is why the rejection uses the language “For example” before listing the questions. The examiner could not tell if Applicant is arguing that the claimed “model” is a conventional model known in the art (and the examiner does not want to assume that Applicant is making such an admission on the record), but if Applicant is, then the examiner requests that Applicant state clearly on the record that the claimed model is a well-known and/or conventional model. Evidence of conventionality does support an argument traversing the written description requirement because well-known structures do not need to be described, but the examiner also cautions Applicant because the same evidence could then be used in a 102/103 rejection. At this point in prosecution, the examiner does not find that Applicant is making such an admission.
With respect to the 101 rejections, the Ex parte Desjardins decision found the claims were found patentable because the machine learning model improved computer technology. Specifically, in that case, the Court found evidence that the machine learning model reduced storage requirements while preserving performance. The examiner did not find any evidence in the instant application that the machine learning models are reducing storage requirements while preserving performance. If Applicant has evidence of such an improvement, the examiner invites Applicant to submit such evidence so that the Examiner may review it when considering whether the withdraw the 101 rejection or not. With respect to whether the claim limitations can be practiced in the human mind or not, the claim language is still broad enough to encompass subject matter that can be practiced in the human mind. For example, “separating the obtained sensor data into individual windows of sensor data” is broad enough to encompass separating two pieces of data into two windows. That is, “sensor data” is broad enough to just be two values and if those two values are separated, i.e., one value in one window and one value in another window, then that scenario is covered by the scope of the claim language. Separating two pieces of data is not too complex for the human mind. Similarly, “computing a forward pass …” is also broad enough to encompass a scenario where it is not too complex for the human mind based on the same reasoning as above.
With respect to the 103 rejections, Applicant’s arguments have been considered but are moot because the arguments do not address the new grounds of rejection necessitated by Applicant’s amendments presented in the response filed 9/17/26.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DANIEL LEE CERIONI whose telephone number is (313) 446-4818. The examiner can normally be reached M - F 8:00 AM - 5:00 PM PT.
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, Jennifer Robertson can be reached at (571) 272-5001. 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.
/DANIEL L CERIONI/Primary Examiner, Art Unit 3791