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 .
This final office action is in response to the response filed 1 September 2026.
Claims 1-20 are pending. Claims 1 and 12 are independent claims.
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.
Claims 1-20 remain rejected under 35 U.S.C. 101 because the claimed invention is directed to abstract idea without significantly more.
Step 1:
According to Step 1 of the two Step analysis, claims 1-11 are directed toward a method (process). Claims 12-20 are directed toward a system (machine). Therefore, each of these claims falls within one of the four statutory categories.
Claim 1:
Step 2A, Prong 1:
The claim recites in part:
generating a base frequency domain representation of the time series data (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation of time series data and generating a base frequency domain representation)
adding a first noise to the base frequency domain representation to obtain a first noisy frequency domain representation (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to add noise to the frequency domain representation to create a first noisy frequency domain representation)
obtaining a first noisy time domain representation of the time series data by applying an inverse discrete Fourier transform to the first noisy frequency domain representation, the first noisy time domain representation including a set of values for a set of time points, wherein each value in the set of values corresponds to a time point in the set of time points (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation by applying a discrete Fourier transform to the first noisy frequency representation to obtain a first noisy time domain representation)
adding a second noise to one or more values in the set of values of the first noisy time domain representation of the time series data to generate a second noisy time domain representation (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to add noise to the values in the first noisy time domain representation to generate a second noisy time domain representation)
for each of one or more time points in the second noisy time domain representation, replacing the corresponding value with an anomalous value to generate an anomalous training data set including one or more anomalous values, the anomalous training data set having a corresponding anomalous label for a time period including the one or more anomalous values (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses a judgement to replace the values of the second noisy time domain representation with anomalous values to generate an anomalous training data set)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional element:
receiving time series data for training the machine learning model
This amounts to extra-solution activity of gathering data for use in the claimed process. As described in MPEP 2106.05(g), limitations that amount to merely adding insignificant extra-solution activity to a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
Further, the courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
The claim recites the additional element:
training the machine learning model using the anomalous training data set and the corresponding anomalous label
This element is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f))
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do no integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional element:
receiving time series data for training the machine learning model
This amounts to extra-solution activity of gathering data for use in the claimed process. As described in MPEP 2106.05(g), limitations that amount to merely adding insignificant extra-solution activity to a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
Further, the courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
The claim recites the additional element:
training the machine learning model using the anomalous training data set and the corresponding anomalous label
This element is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f))
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 2:
With respect to claim 2, the claim depends upon claim 1. The analysis of claim 1 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
wherein adding the first noise to the base frequency domain representation to obtain the first noisy frequency domain representation further comprises: for each of one or more frequencies in the base frequency domain representation, increasing or decreasing at least one of: a corresponding phase value or a corresponding amplitude value by a random amount (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to add noise to the frequency domain representation to create a first noisy frequency domain representation by increasing/decreasing a phase value or corresponding amplitude value by a random amount)
Step 2A, Prong 2:
There are no additional elements considered under Step 2A, Prong 2.
Step 2B:
There are no additional elements considered under Step 2B.
Claim 3:
With respect to claim 3, the claim depends upon claim 1. The analysis of claim 1 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
wherein the corresponding anomalous label indicate at least one anomaly type of: a global point anomaly, a contextual point anomaly, a shapelet anomaly, a seasonal anomaly, or a trend anomaly (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses a judgement to replace the values of the second noisy time domain representation with anomalous values to generate an anomalous label, wherein the label corresponds to an anomaly type)
Step 2A, Prong 2:
There are no additional elements considered under Step 2A, Prong 2.
Step 2B:
There are no additional elements considered under Step 2B.
Claim 4:
With respect to claim 4, the claim depends upon claim 3. The analysis of claim 3 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
wherein the anomaly type is determined based on at least one of: (i) a predetermined anomaly type selected randomly from a distribution or (ii) a predefined anomaly insertion type (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses a judgement to determine the anomaly type based upon a random selection from a distribution or a predetermined anomaly insertion type)
Step 2A, Prong 2:
There are no additional elements considered under Step 2A, Prong 2.
Step 2B:
There are no additional elements considered under Step 2B.
Claim 5:
With respect to claim 5, the claim depends upon claim 1. The analysis of claim 1 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
obtaining a second noisy frequency domain representation of the time series data by applying a discrete Fourier transform to the second noisy time domain representation, the second noisy frequency domain representation including a corresponding phase and corresponding amplitude for each frequency (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation by applying a discrete Fourier transform to the second noisy time domain representation to obtain a second noisy time domain representation)
for each of one or more frequencies in the second noisy frequency domain representation, replacing at least one of: the corresponding phase of the corresponding amplitude with a second anomalous value (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation to add noise to the frequency domain representation to create a first noisy frequency domain representation by increasing/decreasing a phase value or corresponding amplitude value by a random amount)
Step 2A, Prong 2:
There are no additional elements considered under Step 2A, Prong 2.
Step 2B:
There are no additional elements considered under Step 2B.
Claim 6:
With respect to claim 6, the claim depends upon claim 1. The analysis of claim 1 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
wherein the one or more values in the set of values of the first noisy time domain representation are determined based on a position of the one or more corresponding time points, and wherein the position is selected randomly from a distribution (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation of a position randomly selected from a distribution to determine the values in the set of values of the first noisy time domain representation)
Step 2A, Prong 2:
There are no additional elements considered under Step 2A, Prong 2.
Step 2B:
There are no additional elements considered under Step 2B.
Claim 7:
With respect to claim 7, the claim depends upon claim 1. The analysis of claim 1 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
wherein a number of anomalous values in the anomalous training data set: (i) is within a specified range or (ii) results from a probabilistic determination as to whether each of a second set of time points in the second noisy time domain representation are to be anomalous (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation of a range or results of a probabilistic determination)
Step 2A, Prong 2:
There are no additional elements considered under Step 2A, Prong 2.
Step 2B:
There are no additional elements considered under Step 2B.
Claim 8:
With respect to claim 8, the claim depends upon claim 1. The analysis of claim 1 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
replacing one or more adjoining values of one or more adjoining time points with one or more second anomalous values, wherein a number of the one or more adjoining time points is predetermined or selected randomly from a distribution (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an evaluation of adjoining time points via predetermination or random selection for replacing adjoining values in adjoining time points)
Step 2A, Prong 2:
There are no additional elements considered under Step 2A, Prong 2.
Step 2B:
There are no additional elements considered under Step 2B.
Claim 9:
With respect to claim 9, the claim depends upon claim 1. The analysis of claim 1 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
wherein the anomalous value is chosen from a range of possible values, each respective possible value being selected randomly from a distribution (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an opinion to randomly select a value from a distribution)
Step 2A, Prong 2:
There are no additional elements considered under Step 2A, Prong 2.
Step 2B:
There are no additional elements considered under Step 2B.
Claim 10:
With respect to claim 10, the claim depends upon claim 1. The analysis of claim 1 is incorporated herein by reference.
Step 2A, Prong 1:
The claims are directed toward the abstract idea identified with respect to claim 1.
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional element:
wherein the anomalous training data set represents access request to a resource, and wherein the machine learning model is trained to identify anomalous access requests to access the resource
This element is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f))
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do no integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional element:
wherein the anomalous training data set represents access request to a resource, and wherein the machine learning model is trained to identify anomalous access requests to access the resource
This element is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f))
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 11:
With respect to claim 11, the claim depends upon claim 1. The analysis of claim 1 is incorporated herein by reference.
Step 2A, Prong 1:
The claim recites:
generating additional anomalous training data sets, each having one or more anomalous labels for corresponding time periods (mental process; As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation of data and generating anomalous training data sets based upon labels)
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional element:
wherein the additional anomalous training data sets are used to train the training machine learning model
This element is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f))
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do no integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional element:
wherein the additional anomalous training data sets are used to train the training machine learning model
This element is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f))
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 12:
With respect to claim 12, the claim recites the limitations substantially similar to those in claim 1. The analysis of claim 1 is incorporated herein by reference.
Step 2A, Prong 1:
The claim is directed toward the abstract idea identified with respect to claim 1.
Step 2A, Prong 2:
The judicial exception is not integrated into a practical application.
The claim recites the additional elements:
one or more storage media configured to store computer-executed instructions
one or more processors configured to access the one or more storage media and execute the computer-executable instructions
These elements are recited at a high-level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do no integrate the judicial exception into a practical application.
Step 2B:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception.
The claim recites the additional elements:
one or more storage media configured to store computer-executed instructions
one or more processors configured to access the one or more storage media and execute the computer-executable instructions
These elements are recited at a high-level of generality such that it amounts to no more than mere instructions to apply the exception using a generic computer component (See MPEP 2106.05(f)).
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claims 13-20:
With respect to claims 13-20, the claims recite the limitations substantially similar to those in claims 2-9, respectively. Claims 13-20 are rejected under similar rationale.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
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.
Claims 1, 3-12, and 14-20 remain rejected under 35 U.S.C. 103 as being unpatentable over Ma et al. (US 2016/0019271, published 21 January 2016) and further in view of Dutta et al. (US 2023/0186075, filed 28 March 2022, hereafter Dutta) and further in view of Cheon et al. (US 2024/0394509, filed 25 May 2023, hereafter Cheon).
As per independent claim 1, Ma discloses a method comprising:
receiving time series data (Figure 8; paragraphs 0011 and 0104: Here, a base data set is received. The base dataset includes real data and time dependent characteristics (paragraph 0010))
generating a base frequency domain representation of the time series data (paragraph 0012: Here, the base dataset is decomposed into dynamic components and static components. Additionally, the base dynamic component is analyzed to determine the mean and variance of the component (paragraph 0078))
adding a first noise to the base frequency domain representation to obtain a first noisy frequency domain representation (paragraph 0012-0013 and 0105: Here, the decomposed dynamic components are replaced with random components and merged with the static components of the base dataset. The result is a noisy synthetic dataset. Additionally, this noisy synthetic dataset may be generated in accordance with a normal distribution (paragraphs 0077-0079))
obtaining a first noisy time domain representation of the time series data, the first noisy time domain representation including a set of values for a set of time points, wherein each value in the set of values corresponds to a time point in the set of time points (paragraph 0013: Here, the base dataset is decomposed and dynamic and static components are identified. These static components maintain time dependent information from the real base dataset to the synthetic noisy dataset. The dynamic components are replaced with noisy synthetic data and merged with the static components of the base dataset)
adding a second noise to one or more values in the set of values for the first noisy time domain representation of the time series data to generate a second noisy time domain representation (paragraph 0106: Here, based upon receiving user feedback via a graphical user interface, a second set of noise may be applied to the synthetic dataset to introduce anomalies into the dataset)
for each of one or more time points in the second noisy time domain representation, replacing the corresponding value with an anomalous value to generate an anomalous training data set including one or more anomalous values (paragraph 0106: Here, based upon receiving user feedback via a graphical user interface, a second set of noise may be applied to the synthetic dataset to introduce anomalies into the dataset)
However, Ma fails to specifically disclose:
applying an inverse discrete Fourier transform to the first noisy frequency domain representation
the anomalous training data set having a corresponding anomalous label for a time period including one or more anomalous values
training the machine learning model using the anomalous training data set and the corresponding anomalous label
However, Dutta, which is analogous to the claimed invention because it is directed toward training a machine learning model to identify anomalies, discloses:
the anomalous training data set having a corresponding anomalous label for a time period including one or more anomalous values (paragraph 0005: Here, a machine learning model is trained using a training data set, such as time series data, in which some subset of data samples are identified as anomalies. This includes the training data set including a ground-truth anomaly label)
training the machine learning model using the anomalous training data set and the corresponding anomalous label (paragraph 0005: Here, a machine learning model is trained using a training data set, such as time series data, in which some subset of data samples are identified as anomalies. This includes the training data set including a ground-truth anomaly label)
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Dutta with Ma, with a reasonable expectation of success, as it would have allowed for generating synthetic datasets (Ma: Figure 8) for use in training a machine learning model to detect anomalies (Dutta: Abstract).
Further, Cheon, which is analogous to the claimed invention because it is directed toward training machine learning models, discloses applying an inverse discrete Fourier transform to the first noisy frequency domain representation (paragraph 0178: Here, an inverse discrete Fourier transform is applied to decompose a sequence of time series values into components of different frequencies in the frequency domain. This data may be used to train a machine learning model (paragraph 0179)). It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Cheon with Ma-Dutta, with a reasonable expectation of success, as it would have allowed for decomposing the time series data into individual components corresponding to frequency domains (Cheon: paragraph 0178).
As per dependent claim 3, Ma, Dutta, and Cheon disclose the limitations similar to those in claim 1, and the same rejection is incorporated herein. Ma discloses wherein the anomaly type is one of: a global point anomaly, a contextual point anomaly, a shapelet anomaly, a seasonal anomaly (paragraphs 0011 and 0104), or a trend anomaly (paragraphs 0011 and 0104).
Ma fails to specifically disclose an anomalous label. However, Dutta, which is analogous to the claimed invention because it is directed toward training a machine learning model to identify anomalies, discloses an anomalous label (paragraph 0005: Here, a machine learning model is trained using a training data set, such as time series data, in which some subset of data samples are identified as anomalies. This includes the training data set including a ground-truth anomaly label).
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Dutta with Ma, with a reasonable expectation of success, as it would have allowed for generating synthetic datasets (Ma: Figure 8) for use in training a machine learning model to detect anomalies (Dutta: Abstract).
As per dependent claim 4, Ma, Dutta, and Cheon, disclose the limitations similar to those in claim 3, and the same rejection is incorporated herein. Ma disclose wherein the anomaly type is determined based on at least one of: (i) a predetermined anomaly type selected randomly from a distribution or (ii) a predefined anomaly insertion type (Figure 3, item 304; paragraph 0069: Here, the synthetic information incorporates predefined anomaly insertion types, seasonal anomaly and trend anomaly).
As per dependent claim 5, Ma, Dutta, and Cheon disclose the limitations similar to those in claim 1, and the same rejection is incorporated herein. Ma discloses for each of one or more frequencies in the second noise frequency domain representation, replacing at least one of: the corresponding phrase or the corresponding amplitude with a second anomalous value (paragraph 0106: Here, based upon receiving user feedback via a graphical user interface, a second set of noise may be applied to the synthetic dataset to introduce anomalies into the dataset).
Ma fails to specifically disclose obtaining a second noisy frequency domain representation of the time series data by applying a discrete Fourier transform to the second noisy time domain representation, the second noisy frequency domain representation including a corresponding phase and corresponding amplitude for each frequency.
Further, Cheon, which is analogous to the claimed invention because it is directed toward training machine learning models, discloses obtaining a second noisy frequency domain representation of the time series data by applying a discrete Fourier transform to the second noisy time domain representation, the second noisy frequency domain representation including a corresponding phase and corresponding amplitude for each frequency (paragraph 0178: Here, a discrete Fourier transform is applied to decompose a sequence of time series values into components of different frequencies in the frequency domain. This data may be used to train a machine learning model (paragraph 0179)). It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Cheon with Ma-Dutta-Cheon, with a reasonable expectation of success, as it would have allowed for decomposing the time series data into individual components corresponding to frequency domains (Cheon: paragraph 0178).
As per dependent claim 6, Ma, Dutta, and Cheon disclose the limitations similar to those in claim 1, and the same rejection is incorporated herein. Ma discloses wherein the one or more values in the set of values of the first noisy time domain representation are determined based on a position of the one or more corresponding time points ((paragraph 0013: Here, the base dataset is decomposed and dynamic and static components are identified. These static components maintain time dependent information from the real base dataset to the synthetic noisy dataset. The dynamic components are replaced with noisy synthetic data and merged with the static components of the base dataset. The synthetic noisy values correspond to the time series data points).
Ma fails to specifically disclose a position randomly selected from the distribution. However, Cheon, discloses a position randomly selected from the distribution (paragraph 0253: Here, a generator may provide noise, random input, using pseudo-random values based upon a seed to generate synthetic data). It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Cheon with Ma-Dutta-Cheon, with a reasonable expectation of success, as it would have allowed for generating synthetic noise data using random inputs as a seed (Cheon: paragraph 00253).
As per dependent claim 7, Ma, Dutta, and Cheon disclose the limitations similar to those in claim 1, and the same rejection is incorporated herein. Dutta discloses wherein a number of anomalous values in the anomalous training data set: (i) is within a specified range or (ii) results from a probabilistic determination as to whether each of a second set of time points in the second noisy time domain representation are to be anomalous (paragraph 0005: Here, a machine learning model is trained using a training data set, such as time series data, in which some subset of data samples are identified as anomalies. This includes the training data set including a ground-truth anomaly label. This trained model may then be used to identify anomalies based on a probability (accuracy threshold))
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Dutta with Ma, with a reasonable expectation of success, as it would have allowed for generating synthetic datasets (Ma: Figure 8) for use in training a machine learning model to detect anomalies (Dutta: Abstract).
As per dependent claim 8, Ma, Dutta, and Cheon disclose the limitations similar to those in claim 1, and the same rejection is incorporated herein. Ma discloses replacing one or more adjoining values of the one or more adjoining time points with one or more second anomalous values, wherein a number of the one or more adjoining time points is predetermined of selected randomly from a distribution (paragraph 0012-0013 and 0105: Here, the decomposed dynamic components are replaced with random components and merged with the static components of the base dataset. The result is a noisy synthetic dataset. Additionally, this noisy synthetic dataset may be generated in accordance with a normal distribution (paragraphs 0077-0079). In this instance, it is predetermined that the synthetic values will replace all values in the time series).
As per dependent claim 9, Ma, Dutta, and Cheon disclose the limitations similar to those in claim 1, and the same rejection is incorporated herein. Dutta discloses wherein the anomalous value is chosen from a range of possible values, each respective possible value being selected randomly from a distribution (paragraph 0079: Here, the data points may be randomly sampled from historical data representing arbitrarily selected time periods).
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Dutta with Ma, with a reasonable expectation of success, as it would have allowed for generating synthetic datasets (Ma: Figure 8) for use in training a machine learning model to detect anomalies (Dutta: Abstract).
As per dependent claim 10, Ma, Dutta, and Cheon disclose the limitations similar to those in claim 1, and the same rejection is incorporated herein. Dutta, discloses wherein the anomalous training data set represents access requests to a resource and wherein the machine learning model is trained to identify anomalous access requests to access the resource (paragraph 0067: Here, the anomalous data pattern may include unusual data access patterns/requests).
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Dutta with Ma, with a reasonable expectation of success, as it would have allowed for generating synthetic datasets (Ma: Figure 8) for use in training a machine learning model to detect anomalies (Dutta: Abstract).
As per dependent claim 11, Ma, Dutta, and Cheon disclose the limitations similar to those in claim 1, and the same rejection is incorporated herein. Dutta, discloses generating additional anomalous training data sets, each having one or more anomalous labels for corresponding time periods, wherein the additional anomalous training data sets are used to train the machine learning model (paragraph 0005: Here, a machine learning model is trained using a training data set, such as time series data, in which some subset of data samples are identified as anomalies. This includes the training data set including a ground-truth anomaly label)
It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Dutta with Ma, with a reasonable expectation of success, as it would have allowed for generating synthetic datasets (Ma: Figure 8) for use in training a machine learning model to detect anomalies (Dutta: Abstract).
With respect to claim 12, the claim recites the limitations substantially similar to those in claim 1, and the same rejection is incorporated herein by reference.
Ma further discloses one or more storage media configured to store computer-executable instructions and one or more processors configured to access the one or more storage media and execute the computer-executable instructions (Figure 10, items 1002 and 1004; paragraphs 0122-0124).
With respect to claims 14-20, the claims recite the limitations substantially similar to those in claims 3-9, respectively. Claims 14-20 are rejected under similar rationale.
Claims 2 and 13 remain rejected under 35 U.S.C. 103 as being unpatentable over Ma, Dutta, and Cheon and further in view of Moffatt (US 7813433, patented 12 October 2010).
As per dependent claim 2, Ma, Dutta, and Cheon disclose the limitations similar to those in claim 1, and the same rejection is incorporated herein. Ma fails to specifically disclose for each of one or more frequencies in the base frequency domain representation, increasing or decreasing at least one of: a corresponding phase value or a corresponding amplitude value by a random amount.
However, Moffatt, which is analogous to the claimed invention because it is directed toward modifying frequencies, discloses for each of one or more frequencies in the base frequency domain representation, increasing or decreasing at least one of: a corresponding phase value or a corresponding amplitude value by a random amount (claim 1: Here, amplitude and/or phase values are randomly increased to reduce interference). It would have been obvious to one of ordinary skill in the art at the time of the applicant’s effective filing date to have combined Moffat with Ma-Dutta-Cheon, with a reasonable expectation of success, as it would have allowed for reducing interference (Moffatt: claim 1).
With respect to claim 13, the claims recite the limitations substantially similar to those in claim 2. Claim 13 is rejected under similar rationale.
Response to Arguments
Applicant's arguments with respect to the rejection of claims under 35 USC 101 have been fully considered but they are not persuasive.
With respect to Step 2A, Prong One, the applicant argues that the claims do not recite a mental process because “a Fourier transform… cannot be done mentally, or even generating a base a frequency domain representation (page 9).” The applicant alleges that the “meaning of what can be done mentally…appears to encompass all software and is therefore not consistent with the USPTO examples or guidance (page 9-10).” The examiner respectfully disagrees.
Instead, the examiner interprets limitations under the broadest reasonable interpretation and determines whether the limitation covers performance of a mental process including an observation, evaluation, judgment, opinion but for the recitation of generic computer components. Specifically, the mental process may be performed with the aid of pencil and paper.
In this instance, generating a base frequency domain representation of the time series data is a mental process. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation of time series data and generating a base frequency domain representation.
As defined by the specification, time series data “may refer to any collection of data values that correspond to a period of time (paragraph 0029; emphasis added).” Additionally, the specification does not appear to define the specific scope of the term “base frequency domain representation;” instead, the specification provides examples of base frequency domain representations (paragraphs 0082-0085). However, the specification states the “time series may be converted into the respective frequency domain and either used as a base frequency domain representation or one or more base parameters… of the frequency domain representation may be modified to thereby create a base frequency domain representation (paragraph 0085).”
Therefore, the broadest reasonable interpretation of the term “base frequency domain representation” includes conversion of the time domain series or modification of one or more base parameters (paragraph 0085). Because the specification does not define the scope of the term ‘base frequency domain representation,” the broadest reasonable interpretation of observing the time domain series or base parameters and performing a conversion/modification of these parameters may be done mentally or with the aid of a pencil and paper.
Further, applying a Fourier transform is a mental process. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses an observation of time series data and generating a base frequency domain representation.
In this instance, under its broadest reasonable interpretation applying a Fourier transform is an evaluation, performed with the aid of a pencil and paper. Although the applicant argues that a Fourier transform cannot be performed mentally or with the aid of pencil and paper, the examiner respectfully disagrees. The Fourier transform and inverse Fourier transform were originally defined and used/evaluated by the French mathematician Jean Baptiste Joseph Fourier in the early 1800s (see: Highlights in the History of the Fourier Transform, 2016, page 6). For at least this reason, this argument is not persuasive.
With respect to Step 2A, Prong 2, the applicant argues that the claims recite an improvement to a technology or technical field (pages 10-11). Specifically, “solutions described herein can generate a complex and/or realistic time series where the position and anomaly type of each anomaly is known while also retaining the base signal pattern. Such complex time series data that includes anomalies may be useful for training models to detect anomalies in future unlabeled time series. Anomalies may be anomalous access request to a resource. For example, anomalies may relate to cybersecurity for a network computer, where such time series data relates to data read operations, data write operations, communications from/to a device, communications between devices, etc. Thus, precise anomaly detection may be useful in many scenarios (page 10).”
As an initial matter, much of the “improvement” is based upon limitations that are unclaimed. Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993).
Additionally, at least a portion of the argued improvements are directed to the mental process considered under Step 2A, Prong 1. Therefore, the argument conflates a mental process performed using a computer with a technological solution. In this instance, the claim recites the additional element considered under Step 2A, Prong 2:
receiving time series data for training the machine learning model
This amounts to extra-solution activity of gathering data for use in the claimed process. As described in MPEP 2106.05(g), limitations that amount to merely adding insignificant extra-solution activity to a judicial exception do not amount to significantly more than the exception itself, and cannot integrate a judicial exception into a practical application.
Further, the courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Further, the claim recites the additional element considered under Step 2A, Prong 2:
training the machine learning model using the anomalous training data set and the corresponding anomalous label
This element is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f))
For this reason, this argument is not persuasive.
With respect to Step 2B, the applicant argues that the claims recite significantly more than the judicial exception (page 11). To support this rationale, the applicant argues that the claims recite the following elements that cannot be practically performed mentally, “generating a base frequency domain representation of the time series data… obtaining a first noisy time domain representation of the time series data by applying an inverse discrete Fourier transform to the first noisy frequency domain representation… and training the machine learning model using the anomalous training data set and the corresponding anomalous label (page 11).” This argument is not persuasive.
First, the examiner has addressed the applicant’s arguments regarding “generating a base frequency domain representation of the time series data” and the use of an “inverse discrete Fourier transform” above with respect to Step 2A, Prong 1. As noted above, the argument that these limitations are not mental processes is not persuasive.
Additionally, it is noted that the examiner does not allege that “training the machine learning model using the anomalous training data set and the corresponding anomalous label” is a mental process. Instead, as noted above with respect to Step 2A, Prong 2, and Step 2B, “training the machine learning model using the anomalous training data set and the corresponding anomalous label” is an additional element.
Specifically, this element is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)). For this reason, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception, and this argument is not persuasive.
Applicant's arguments with respect to the rejection of claims under 35 USC 103 have been fully considered but they are not persuasive.
The applicant’s initial argument with respect to the rejection of claims under 35 USC 103 is that no reference teaches or suggests generating a base frequency domain representation of the time series data (pages 13-14). The examiner respectfully disagrees.
Ma discloses receiving time series data (Figure 8; paragraphs 0011 and 0104) and generating a base frequency domain representation of the time series data (paragraph 0012). Specifically, a base data including real data and time dependent characteristics (paragraph 0010) is received. The base dataset is decomposed into dynamic components and static components. Additionally, the base dynamic component is analyzed to determine the mean and variance of the component (paragraph 0078).
As defined by the specification, time series data “may refer to any collection of data values that correspond to a period of time (paragraph 0029; emphasis added).” Additionally, the specification does not appear to define the specific scope of the term “base frequency domain representation;” instead, the specification provides examples of base frequency domain representations (paragraphs 0082-0085). However, the specification states the “time series may be converted into the respective frequency domain and either used as a base frequency domain representation or one or more base parameters… of the frequency domain representation may be modified to thereby create a base frequency domain representation (paragraph 0085).”
Therefore, modifying the base data by decomposing into dynamic and static components is consistent with the applicant’s definition of “generating a base frequency domain representation of the time series data” as defined by the specification. For this reason, this argument is not persuasive.
The applicant further argues that the prior art fails to disclose adding a first noise to the base frequency domain representation to obtain a first noisy frequency domain representation (pages 14-15). The examiner respectfully disagrees.
Ma discloses adding a first noise to the base frequency domain representation to obtain a first noisy frequency domain representation (paragraph 0012-0013 and 0105). Specifically, the decomposed dynamic components are replaced with random components and merged with the static components of the base dataset. The result is a noisy synthetic dataset. Additionally, this noisy synthetic dataset may be generated in accordance with a normal distribution (paragraphs 0077-0079)).
The applicant’s specification defines “noise” as “a random variation or fluctuation that interferes with a base signal. Noise can be added to a base signal by adding or subtracting random values at one or more data points (paragraph 0011; emphasis added).” As noted by the examiner, Ma discloses “randomly resampling data points from the base dataset (paragraph 0105).”
Therefore, replacing decomposed dynamic components with random components and merging with the static components of the base dataset is consistent with “adding a first noise to the base frequency domain representation to obtain a first noisy frequency domain representation” as defined by the specification. For this reason, this argument is not persuasive.
The applicant further argues that no reference teaches or suggests “obtaining a first noisy time domain representation of the time series dataset (page 15).” The examiner respectfully disagrees.
Ma discloses obtaining a first noisy time domain representation of the time series data, the first noisy time domain representation including a set of values for a set of time points, wherein each value in the set of values corresponds to a time point in the set of time points (paragraph 0013). In this instance, the base dataset is decomposed and dynamic and static components are identified. These static components maintain time dependent information from the real base dataset to the synthetic noisy dataset. The dynamic components are replaced with noisy synthetic data and merged with the static components of the base dataset.
The applicant alleges that the examiner “double-maps paragraph [0013] of Ma to meet two separate claim elements: the recited “first noisy frequency domain representation” and the “first noisy time domain representation.”
While it is true that both limitations, “adding a first noise…” and “obtaining a first noisy…” include citations to paragraph 0013, it is possible for a single paragraph to teach multiple limitations. In this instance, as noted by the examiner, a first noisy frequency domain representation is obtained by adding a first noise to the base frequency domain representation to obtain a first noisy frequency domain representation (paragraph 0012-0013 and 0105). Specifically, the decomposed dynamic components are replaced with random components of the base dataset. The result is a noisy synthetic dataset. Additionally, this noisy synthetic dataset may be generated in accordance with a normal distribution (paragraphs 0077-0079)).
This noisy synthetic dataset (first noisy frequency domain representation) may be merged with the static data components, the static data components maintaining the time dependent information, to obtain a first noisy time domain representation (paragraph 0013). For this reason, this argument is not persuasive.
The applicant further argues that the prior art fails to teach or suggest “adding a second noise to one or more values in the set of values of the first noisy time domain representation of the time series data to generate a second noisy time domain representation (pages 15-16). The examiner respectfully disagrees.
Ma discloses adding a second noise to one or more values in the set of values for the first noisy time domain representation of the time series data to generate a second noisy time domain representation (paragraph 0106). In this instance, based upon receiving user feedback via a graphical user interface, a second set of noise may be applied to the synthetic dataset to introduce anomalies into the dataset.
Specifically, moving a particular data point from a first location in the graph to a second location in the graph introduces an anomaly (noise) into the synthetic dynamic component. As noted previously by the examiner, the applicant’s specification defines “noise” as “a random variation or fluctuation that interferes with a base signal. Noise can be added to a base signal by adding or subtracting random values at one or more data points (paragraph 0011; emphasis added).” As noted by the examiner, Ma discloses “randomly resampling data points from the base dataset (paragraph 0105).”
Therefore, moving a particular data point from a first location in the graph to a second location in the graph constitutes subtracting a value at one position and adding a value at another position. This is consistent with “adding a second noise to one or more values in the set of values for the first noisy time domain representation of the time series data to generate a second noisy time domain representation” as defined by the specification. For this reason, this argument is not persuasive.
Finally, the applicant argues that the combination of references is improper because there is no motivation to combine Ma with any reference that would alter frequency parameters (pages 16-17). To support this position, the applicant argues that “adding noise in the frequency domain to any of Ma’s components would render Ma unsatisfactory for its intended purpose by fracturing the time dependent characteristics… that Ma wishes to retain (page 17).”
However, it is noted that Ma discloses the addition of noise. First, Ma discloses adding a first noise to the base frequency domain representation to obtain a first noisy frequency domain representation (paragraph 0012-0013 and 0105). Specifically, the decomposed dynamic components are replaced with random components and merged with the static components of the base dataset. The result is a noisy synthetic dataset. Additionally, this noisy synthetic dataset may be generated in accordance with a normal distribution (paragraphs 0077-0079)). Additionally, Ma discloses adding a second noise to one or more values in the set of values for the first noisy time domain representation of the time series data to generate a second noisy time domain representation (paragraph 0106). In this instance, based upon receiving user feedback via a graphical user interface, a second set of noise may be applied to the synthetic dataset to introduce anomalies into the dataset. For these reasons, this argument is not persuasive.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Highlights in the History of the Fourier Transform (2016): Discloses the history of the Fourier transform and equations for performing a Fourier transfer and inverse Fourier transform (page 6)
Niu et al. (US 12511576): Discloses generating a synthetic time series for training a machine learning model, the machine learning model to provide inferences to detect anomalies in time series data (column 3, lines 31-43)
Li (US 2024/0135683): Discloses a generative model to generate synthetic data (paragraph 0003) and using a trained model to perform classification, including anomaly detection (paragraph 0050)
Nguyen et al. (US 2023/0316105): Discloses generating synthetic data based on input data and the output of the knowledge model as training data for an identifying anomalies in time series data (paragraphs 0006-0007)
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 KYLE R STORK whose telephone number is (571)272-4130. The examiner can normally be reached 8am - 2pm; 4pm - 6pm.
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, Omar Fernandez Rivas can be reached at 571/272-2589. 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.
/KYLE R STORK/Primary Examiner, Art Unit 2128
Prosecution Insights