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 .
Claims 1-18 are presented for examination.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on March 9, 2023 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Objections
Claims 6, 12 and 18 are objected to because of the following informalities:
Claims 6, 12, and 18: It appears in light of the specification that “a peripheral region of data of interest in the intermediate output and the low-dimensional feature value” should read “a peripheral region of data of interest in the intermediate output of the low-dimensional feature value.”
Appropriate correction is required.
Specification
The disclosure is objected to because it contains an embedded hyperlink and/or other form of browser-executable code ([0004]). Applicant is required to delete the embedded hyperlink and/or other form of browser-executable code; references to websites should be limited to the top-level domain name without any prefix such as http:// or other browser-executable code. See MPEP § 608.01.
The title of the invention is not descriptive. A new title is required that is clearly indicative of the invention to which the claims are directed.
The disclosure is objected to because of the following informalities:
[0005]: "are employed" should read "is employed"
[0011]: "in which in which" should read "in which"
[0017]: "multi-dimensionality input data" should read "multi-dimensional input data"
[0029]: "adjusted such the" should read "adjusted such that the"
[0096]: "The accordingly" should read "This accordingly"
Appropriate correction is required.
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 1-18 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The analysis of the claims will follow the 2019 Revised Patent Subject Matter Eligibility Guidance (“2019 PEG”).
Claim 1
Step 1: The claim recites a non-transitory recording medium, and is therefore directed to the statutory category of articles of manufacture.
Step 2A Prong 1: The claim recites:
estimating, as a probability distribution, a low-dimensional feature value obtained by encrypting input data, the low-dimensional feature value having a lower dimensionality than the input data; This limitation could encompass mentally estimating a probability distribution of a low-dimensional feature value obtained by encrypting input data.
generating output data by decrypting a feature value resulting from adding noise to the low-dimensional feature value; This limitation could encompass mentally decrypting a feature value using a decryption algorithm to generate output data.
adjusting respective parameters of the encrypting, the estimating, and the decrypting, based on a cost including an error between the input data and the output data and including an entropy of the probability distribution; This limitation could encompass mentally adjusting parameters of encrypting, estimating, and decrypting based on a cost.
wherein, in a determination as to whether or not target input data is normal, a determination standard for the determination is controlled based on information obtained from another probability distribution estimated by encrypting the target input data with the parameters after the adjusting; This limitation could encompass mentally controlling a determination standard for a determination based on information obtained from another probability distribution.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. The claim further recites “A non-transitory recording medium storing a determination control program that causes a computer to execute a process comprising… [the method].” However, this limitation amounts to mere instructions to apply the judicial exception using a generic computer (MPEP § 2106.05(f)).
Step 2B: The claim does not contain significantly more than the judicial exception. The non-transitory recording medium limitation amounts to mere instructions to apply the judicial exception using a generic computer (MPEP § 2106.05(f)) as stated above. As an ordered whole, the claim is directed to the abstract idea of determining whether or not data is normal based on an estimated probability distribution. Nothing in the claim provides significantly more than this. As such, the claim is not patent eligible.
Claim 2
Step 1: An article of manufacture, as above.
Step 2A Prong 1: The claim recites:
a probability distribution resulting from mixing a plurality of distributions is estimated as the probability distribution; This limitation could encompass mentally estimating a probability distribution resulting from mixing a plurality of distributions.
from among a plurality of clusters equivalent to the plurality of distributions, which cluster the low-dimensional feature value belongs to is identified based on the information obtained from the other probability distribution; This limitation could encompass mentally identifying which cluster the low-dimensional feature value belongs to based on the information obtained from the other probability distribution.
a determination standard corresponding to the cluster that is identified is set from among a determination standard for each cluster; This limitation could encompass mentally setting a determination standard corresponding to the cluster that is identified.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See analysis of claim 1.
Step 2B: This claim does not contain significantly more than the judicial exception. See analysis of claim 1.
Claim 3
Step 1: An article of manufacture, as above.
Step 2A Prong 1: The claim recites:
wherein the cost is a weighted sum of the error and the entropy, and the parameters are adjusted so as to minimize the cost; This limitation could encompass mentally adjusting the parameters to minimize a cost that is a weighted sum of the error and the entropy.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See analysis of claim 1.
Step 2B: This claim does not contain significantly more than the judicial exception. See analysis of claim 1.
Claim 4
Step 1: An article of manufacture, as above.
Step 2A Prong 1: The claim recites:
wherein the noise is a random number based on a distribution having no inter-dimensional correlation, and having a mean of 0; This limitation merely further limits the noise added to the feature value that is decrypted to generate output data, and the decryption step is still mentally performable when random noise based on a distribution having no inter-dimensional correlation and a mean of 0 is added to the feature value.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See analysis of claim 1.
Step 2B: This claim does not contain significantly more than the judicial exception. See analysis of claim 1.
Claim 5
Step 1: An article of manufacture, as above.
Step 2A Prong 1: The claim recites:
wherein the determination is performed by comparing an entropy of the other probability distribution for the target input data against the determination standard; This limitation encompasses mentally comparing an entropy of the other probability distribution for the target input data against the determination standard.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See analysis of claim 1.
Step 2B: This claim does not contain significantly more than the judicial exception. See analysis of claim 1.
Claim 6
Step 1: An article of manufacture, as above.
Step 2A Prong 1: The claim recites:
wherein the determination is performed by comparing the determination standard against a difference between an expected value of entropy and an entropy of a conditional probability for an intermediate output of the low-dimensional feature value under data of a peripheral region of data of interest in the intermediate output and the low-dimensional feature value; This limitation encompasses mentally comparing the determination standard against a difference between the two entropy values.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. See analysis of claim 1.
Step 2B: This claim does not contain significantly more than the judicial exception. See analysis of claim 1.
Claims 7-12
Step 1: The claims recite a determination control device, and are therefore directed to the statutory category of machines.
Step 2A Prong One: The claims recite the same judicial exception as claims 1-6, respectively.
Step 2A Prong Two: The judicial exception is not integrated into a practical application. The analysis at this step corresponds to that of claims 1-6, respectively, except insofar as claims 7-12 recite a determination control device rather than a non-transitory recording medium. The claims further recite “A determination control device comprising: a memory; and a processor coupled to the memory, the processor being configured to execute processing including… [the method]. However, this limitation amounts to mere instructions to apply the judicial exception using a generic computer (MPEP § 2106.05(f)).
Step 2B: The claims do not contain significantly more than the judicial exception. The analysis at this step mirrors that of claims 1-6, respectively, except insofar as claims 7-12 recite a determination control device rather than a non-transitory recording medium. The determination control device limitation amounts to mere instructions to apply the judicial exception using a generic computer (MPEP § 2106.05(f)) as stated above.
Claims 13-18
Step 1: The claims recite a determination control method, and are therefore directed to the statutory category of processes.
Step 2A Prong One: The claims recite the same judicial exception as claims 1-6, respectively.
Step 2A Prong Two: The judicial exception is not integrated into a practical application. The analysis at this step corresponds to that of claims 1-6, respectively, except insofar as claims 13-18 recite a determination control method rather than a non-transitory recording medium, and therefore recite no additional elements beyond the judicial exception.
Step 2B: The claims do not contain significantly more than the judicial exception. The analysis at this step mirrors that of claims 1-6, respectively, except insofar as claims 13-18 recite a determination control method rather than a non-transitory recording medium, and therefore recite no additional elements beyond the judicial exception.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1-4, 7-10, and 13-16 are rejected under 35 U.S.C. 103 as being unpatentable over Zong et al. (NPL: Deep Autoencoding Gaussian Mixture Model for Unsupervised Anomaly Detection) (“Zong”) in view of Schroers et al. (US20200366914) (“Schroers”), and further in view of Segev (US10692004).
Regarding claim 1, Zong discloses “A non-transitory recording medium storing a determination control program that causes a computer to execute a process (Zong, 4.3 DAGMM Configuration: “All the DAGMM instances are implemented by tensorflow”) comprising:
PNG
media_image1.png
130
814
media_image1.png
Greyscale
estimating, as a probability distribution, a low-dimensional feature value obtained by encrypting input data, the low-dimensional feature value having a lower dimensionality than the input data (Zong, 3.2 Compression Network: “The low-dimensional representations provided by the compression network contains two sources of features: (1) the reduced low-dimensional representations learned by a deep autoencoder; and (2) the features derived from reconstruction error. Given a sample x, the compression network computes its low-dimensional representation z as follows
where zc is the reduced low-dimensional representation learned by the deep autoencoder, zr includes the features derived from the reconstruction error, θe and θd are the parameters of the deep autoencoder, x' is the reconstructed counterpart of x, h(·) denotes the encoding function, g(·) denotes the decoding function, and f(·) denotes the function of calculating reconstruction error features”; the examiner notes that z corresponds to “a low dimensional feature value obtained by encrypting input data” because it is obtained by encoding input data x using encoding function h(·), and z has lower dimensionality than the input data because the autoencoder reduces the dimensionality of the input data to obtain feature value z; Zong, 3.3 Estimation Network: “Given the low-dimensional representations z and an integer K as the number of mixture components, the estimation network makes membership prediction as follows [see eq (4)]… Given a batch of N samples and their membership prediction,
∀
1 ≤ k ≤ K, we can further estimate the parameters in GMM as follows [see eq (5)]…With the estimated parameters, sample energy can be further inferred by
PNG
media_image2.png
82
794
media_image2.png
Greyscale
where
∙
denotes the determinant of a matrix”; the examiner notes that the equation inside -log() corresponds to an estimated probability distribution of low-dimensional feature value z);
generating output data by decrypting a feature value… (Zong, 3.2 Compression Network, eq (1); the examiner notes that x' corresponds to output data generated by decoding a feature value zc using decoding function g(·)); and
PNG
media_image3.png
73
808
media_image3.png
Greyscale
adjusting respective parameters of the encrypting, the estimating, and the decrypting, based on a cost including an error between the input data and the output data and including an entropy of the probability distribution (Zong, 3.4 Objective Function: “Given a dataset of N samples, the objective function that guides DAGMM training is constructed as follows.
; the examiner notes that θe and θd are parameters of the encrypting and decrypting respectively (see Zong, 3.2, eq(1)) and θm is a parameter of the estimating (see Zong, 3.3, eq(4)),
L
(
x
i
,
x
i
'
)
corresponds to an error between the input data and the output data (see Zong, 3.4:
L
(
x
i
,
x
i
'
)
is the loss function that characterizes the reconstruction error caused by the deep autoencoder in the compression network), and E(zi ) corresponds to an entropy of the probability distribution (see Zong, 3.3, eq(6)));
… another probability distribution estimated by encrypting the target input data with the parameters after the adjusting (Zong, 3.3 Estimation Network: “In addition, during the testing phase with the learned GMM parameters, it is straightforward to estimate sample energy, and predict samples of high energy as anomalies by a pre-chosen threshold”; the examiner notes that during the testing phase, the adjusted parameters after training are used to encrypt target input data to estimate another probability distribution (repeating the steps from 3.2 and 3.3)).
Zong does not appear to explicitly disclose “a feature value resulting from adding noise to the low-dimensional feature value” or “wherein, in a determination as to whether or not target input data is normal, a determination standard for the determination is controlled based on information obtained from another probability distribution…”
However, Schroers discloses “generating output data by decrypting a feature value resulting from adding noise to… [a] low-dimensional feature value” (Schroers, [0030]: “This procedure requires a differentiable approximation of the quantization operation performed in the bottleneck and, in one implementation, additive uniform noise is used for this purpose. Adopting the notation 𝒰 for an independent uniform noise of width 1, the density function pŷ of the random variable ỹ = y + 𝒰 (-½, ½) becomes a continuous differentiable relaxation of the probability mass function pŷ” and Fig. 3; the examiner notes that x̃ in Fig. 3, line 4 is output data resulting from decoding ỹ, and y is a low-dimensional feature value because it results from encoding input image x, see Fig.3, line 10).
Schroers and the instant application both relate to neural networks and are analogous. It would have been obvious to one of ordinary skill in the art, prior to the effective filing date of the claimed invention, to have modified the step of generating output data by decrypting a feature value disclosed by Zong, to have the feature value that is decrypted result from adding noise to the low-dimensional feature value as disclosed by Schroers, and one would have been motivated to do so for the purpose of improving the latent representations by using noise to approximate the quantization operation (see Schroers, [0028-0030]).
Zong as modified by Schroers does not appear to explicitly disclose “wherein, in a determination as to whether or not target input data is normal, a determination standard for the determination is controlled based on information obtained from another probability distribution…”
However, Segev discloses “wherein, in a determination as to whether or not target input data is normal, a determination standard for the determination is controlled based on information obtained from… [a] probability distribution…” (Segev, Fig. 4B: “Apply Gaussian Mixture with EM Algorithm to the distribution of MDDPs -> Set the threshold that classifies each MDDP or NAMDDP as either normal or abnormal; and (128): “Another method for identifying abnormal data MDDPs in the embedding matrix Ψ (steps 330′ in FIGS. 3A and 3B) can be based on Gaussian mixture fit and threshold estimation as shown schematically herein with respect to steps 506B-510B in FIG. 4B. The average distance Dnn between each MDDP in the embedded space and its nearest neighbors may be determined (step 506B, FIG. 4B), followed by performing a Gaussian mixture approximation to the distribution of Dnn (step 508B, FIG. 4B) using, for example, the expectation-maximization function (see DLR). Namely the Gaussian mixture can be represented by P(Dnn) = [equation] where wi are the Gaussian weights and G(x|μi ,σi) are the Gaussians densities, with μi is the Gaussian mean of Gaussian i and σi as the standard deviation of Gaussian i; and (133) The threshold Dnnt (also: “T”) can be defined, for example, to be the distance for which P(i|x, μi, σi, maxi(μi+σi)) ≈ 0.9… A new MDPP is declared as an anomalous MDDP if its Dnn is larger than Dnnt; the examiner notes that threshold Dnnt corresponds to a determination standard for the determination, which is set based on Gaussian i, mean μi, and standard deviation σi obtained from probability distribution P(Dnn)).
Segev and the instant application both relate to anomaly detection and neural networks and are analogous. It would have been obvious to one of ordinary skill in the art, prior to the effective filing date of the claimed invention, to have modified the combination of Zong and Schroers to include controlling a determination standard based on information obtained from a probability distribution as disclosed by Segev, using the “another” probability distribution estimated by encrypting the target input data with the parameters after the adjusting disclosed by Zong, and one would have been motivated to do so for the purpose of determining whether or not data is normal by how closely it fits the identified Gaussian (see Segev, (130-131)).
Regarding claim 2, the rejection of claim 1 is incorporated. Zong as modified by Schroers further discloses “wherein: a probability distribution resulting from mixing a plurality of distributions is estimated as the probability distribution” (Zong, 3.3 Estimation Network: “Given the low-dimensional representations for input samples, the estimation network performs density estimation under the framework of GMM [Gaussian Mixture Model]”); and
from among a plurality of clusters equivalent to the plurality of distributions, which cluster the low-dimensional feature value belongs to is identified based on the information obtained from the other probability distribution (Zong, 3.3 Estimation Network: “Given the low-dimensional representations z and an integer K as the number of mixture components, the
PNG
media_image4.png
28
534
media_image4.png
Greyscale
estimation network makes membership prediction as follows.
where γ^ is a K-dimensional vector for the soft mixture-component membership prediction, and p is the output of a multi-layer network parameterized by θm. Given a batch of N samples and their membership prediction,
∀
1 ≤ k ≤ K, we can further estimate the parameters in GMM as follows…[see eq (5)] where
γ
k
^
is the membership prediction for the low-dimensional representation zi , and
ϕ
k
^
,
μ
k
^
,
Σ
k
^
are mixture probability, mean, covariance for component k in GMM, respectively…In addition, during the testing phase with the learned GMM parameters, it is straightforward to estimate sample energy, and predict samples of high energy as anomalies by a pre-chosen threshold.”; the examiner notes that component k in γ^ from the learned GMM parameters corresponds to the cluster the low-dimensional feature value belongs to, and the learned parameters are used to estimate the other probability distribution).
Zong as modified by Schroers does not appear to explicitly disclose “a determination standard corresponding to the cluster that is identified is set from among a determination standard for each cluster.”
However, Segev discloses “a determination standard corresponding to… [a] cluster that is identified is set from among a determination standard for each cluster” (Segev, Fig. 4B: “Apply Gaussian Mixture with EM Algorithm to the distribution of MDDPs -> Set the threshold that classifies each MDDP or NAMDDP as either normal or abnormal; and (128): “Namely the Gaussian mixture can be represented by… [equation] where wi are the Gaussian weights and G(x|μi ,σi) are the Gaussians densities, with μi is the Gaussian mean of Gaussian i and σi as the standard deviation of Gaussian i. The EM algorithm is an iterative algorithm that intends to maximize the likelihood of the Gaussian mixture. At each iteration, the parameters μi and σi of the Gaussian mixture may be determined until convergence is reached; and (133) The threshold Dnnt (also: “T”) can be defined, for example, to be the distance for which P(i|x, μi, σi, maxi(μi+σi)) ≈ 0.9… A new MDPP is declared as an anomalous MDDP if its Dnn is larger than Dnnt; the examiner notes that the formula for threshold Dnnt corresponds to “a determination standard for each cluster” and the threshold Dnnt calculated for the identified Gaussian i corresponds to “a determination standard corresponding to a cluster that is identified”).
Segev and the instant application both relate to anomaly detection and are analogous. It would have been obvious to one of ordinary skill in the art, prior to the effective filing date of the claimed invention, to have modified the combination of Zong and Schroers to include setting a determination standard corresponding to the cluster that is identified from among a determination standard for each cluster as disclosed by Segev, and one would have been motivated to do so for the purpose of determining whether or not data is normal by how closely it fits the identified Gaussian (see Segev, (130-131)).
PNG
media_image3.png
73
808
media_image3.png
Greyscale
Regarding claim 3, the rejection of claim 1 is incorporated. Zong as modified by Schroers and Segev further discloses “wherein the cost is a weighted sum of the error and the entropy, and the parameters are adjusted so as to minimize the cost” (Zong, 3.4 Objective Function: “Given a dataset of N samples, the objective function that guides DAGMM training is constructed as follows.
; the examiner notes that L() corresponds to the error, E(zi) corresponds to the entropy, and the objective function adjusts the parameters to minimize the cost, see Zong, 3.4: “…Therefore, a compression network of lower reconstruction error is always desired…By minimizing the sample energy, we look for the best combination of compression and estimation networks that maximize the likelihood to observe input samples”).
Regarding claim 4, the rejection of claim 1 is incorporated. Zong as modifed by Schroers and Segev further discloses “wherein the noise is a random number based on a distribution having no inter-dimensional correlation, and having a mean of 0” (Schroers, [0030]: “This procedure requires a differentiable approximation of the quantization operation performed in the bottleneck and, in one implementation, additive uniform noise is used for this purpose. Adopting the notation 𝒰 for an independent uniform noise of width 1, the density function pŷ of the random variable ỹ = y + 𝒰 (-½, ½) becomes a continuous differentiable relaxation of the probability mass function pŷ”; the examiner notes that the noise is based on a uniform distribution that is independent (no inter-dimensional correlation), and has a mean of 0 (given the mean formula for uniform distributions:
1
/
2
+
-
1
/
2
2
=
0
)).
Regarding claim 7, Zong discloses “A determination control device comprising: a memory; and a processor coupled to the memory, the processor being configured to execute processing including… [the method]” (Zong, 4.3 DAGMM Configuration: “All the DAGMM instances are implemented by tensorflow”). The remainder of the limitations correspond to those of claim 1, and the claim is rejected for the same reasons given for claim 1 above.
Regarding claim 8, the rejection of claim 7 is incorporated. Claim 8 is a device claim corresponding to non-transitory recording medium claim 2 and is rejected for the same reasons as given for claim 2 above.
Regarding claim 9, the rejection of claim 7 is incorporated. Claim 9 is a device claim corresponding to non-transitory recording medium claim 3 and is rejected for the same reasons given for claim 3 above.
Regarding claim 10, the rejection of claim 7 is incorporated. Claim 10 is a device claim corresponding to non-transitory recording medium claim 4 and is rejected for the same reasons as given for claim 4 above.
Regarding claim 13, claim 13 is a method claim corresponding to claim 1 and is rejected for the same reasons given for claim 1 above.
Regarding claim 14, the rejection of claim 13 is incorporated. Claim 14 is a method claim corresponding to non-transitory recording medium claim 2 and is rejected for the same reasons given for claim 2 above.
Regarding claim 15, the rejection of claim 13 is incorporated. Claim 15 is a method claim corresponding to non-transitory recording medium claim 3 and is rejected for the same reasons as given for claim 3 above.
Regarding claim 16, the rejection of claim 13 is incorporated. Claim 16 is a method claim corresponding to non-transitory recording medium claim 4 and is rejected for the same reasons as given for claim 4 above.
Claims 5, 11, and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Zong in view of Schroers and Segev, and further in view of Liao et al. (NPL: A Unified Unsupervised Gaussian Mixture Variational Autoencoder for High Dimensional Outlier Detection).
Regarding claim 5, the rejection of claim 1 is incorporated. Zong as modifed by Schroers and Segev further discloses “an entropy of the other probability distribution for the target input data” (Zong, 3.3 Estimation Network: “In addition, during the testing phase with the learned GMM parameters, it is straightforward to estimate sample energy, and predict samples of high energy as anomalies by a pre-chosen threshold”; the examiner notes that the sample energy estimated during the testing phase corresponds to “an entropy of the other probability distribution for the target input data”), but does not appear to disclose the further limitations of the claim.
However, Liao discloses “wherein… [a] determination is performed by comparing an entropy of… [a] probability distribution for…target input data against… [a] determination standard (Liao, IV. A.: “In the testing phase, the sample density estimation works as the outlier detector which predicts outliers when the sample density is beyond a learned threshold, which can be obtained during the training phase”; the examiner notes that the sample density estimation corresponds to an entropy of a probability distribution, see eq (3), and the learned threshold corresponds to a determination standard).
Liao and the instant application both relate to autoencoders for outlier detection and are analogous. It would have been obvious to one of ordinary skill in the art, prior to the effective filing date of the claimed invention, to have modified the combination of Zong, Schroers, and Segev, to include performing the determination by comparing an entropy of the other probability distribution for the target input data against the determination standard, as disclosed by Liao, and one would have been motivated to do so for the purpose of improving the accuracy of anomaly detection (see Liao, Abstract).
Regarding claim 11, the rejection of claim 7 is incorporated. Claim 11 is a device claim corresponding to non-transitory recording medium claim 5 and is rejected for the same reasons given for claim 5 above.
Regarding claim 17, the rejection of claim 13 is incorporated. Claim 17 is a method claim corresponding to non-transitory recording medium claim 5 and is rejected for the same reasons given for claim 5 above.
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP § 2146 et seq. for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The filing of a terminal disclaimer by itself is not a complete reply to a nonstatutory double patenting (NSDP) rejection. A complete reply requires that the terminal disclaimer be accompanied by a reply requesting reconsideration of the prior Office action. Even where the NSDP rejection is provisional the reply must be complete. See MPEP § 804, subsection I.B.1. For a reply to a non-final Office action, see 37 CFR 1.111(a). For a reply to final Office action, see 37 CFR 1.113(c). A request for reconsideration while not provided for in 37 CFR 1.113(c) may be filed after final for consideration. See MPEP §§ 706.07(e) and 714.13.
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The actual filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/apply/applying-online/eterminal-disclaimer.
Claims 1, 3-4, 7, 9-10, 13, and 15-16 are provisionally rejected on the ground of nonstatutory obviousness-type double patenting as being unpatentable over claims 1, 3-4, 6, 8-9, 11, and 13-14 respectively of copending Application No. 18/180,401 (reference application) in view of Segev and further in view of Zong. Although the claims at issue are not identical, they are not patentably distinct from each other. This is a provisional nonstatutory double patenting rejection because the patentably indistinct claims have not in fact been patented. The claims of the instant application and the claims of the reference patent are compared in the table below (only the non-transitory recording medium claims are shown for conciseness).
Instant Application
Reference Application (18/180,401)
1. A non-transitory recording medium storing a determination control program that causes a computer to execute a process comprising:
estimating, as a probability distribution, a low-dimensional feature value obtained by encrypting input data, the low-dimensional feature value having a lower dimensionality than the input data;
generating output data by decrypting a feature value resulting from adding noise to the low-dimensional feature value; and
adjusting respective parameters of the encrypting, the estimating, and the decrypting, based on a cost including an error between the input data and the output data and including an entropy of the probability distribution
wherein, in a determination as to whether or not target input data is normal, a determination standard for the determination is controlled based on information obtained from another probability distribution estimated by encrypting the target input data with the parameters after the adjusting.
1. A non-transitory computer-readable recording medium storing an abnormality determination program for causing a computer to execute processing comprising:
estimating a low-dimensional feature quantity with a lower dimensionality than input data obtained by encoding the input data as a conditional probability distribution using a condition based on data in a peripheral area of data of interest in the input data; and
adjusting parameters of each of the encoding and the estimating and decoding of a feature quantity obtained by adding a noise to the low-dimensional feature quantity, based on a cost that includes output data obtained by the decoding, an error between the output data and the input data, and entropy of the conditional probability distribution,
wherein, in determining whether input data to be determined is normal using the adjusted parameters, the determination is performed based on the conditional probability distribution based on data of a peripheral area of the input data to be determined.
3. The non-transitory recording medium of claim 1, wherein the cost is a weighted sum of the error and the entropy, and the parameters are adjusted so as to minimize the cost.
3. The non-transitory computer-readable recording medium according to claim 1, wherein the cost is a weighted sum of the error and the entropy, and the parameters are adjusted so as to minimize the cost.
4. The non-transitory recording medium of claim 1, wherein the noise is a random number based on a distribution having no inter-dimensional correlation, and having a mean of 0.
4. The non-transitory computer-readable recording medium according to claim 1, wherein the noise is a random number based on a distribution in which respective dimensions are uncorrelated with each other and a mean is 0.
Regarding instant claims 1, 7, and 13, reference claims 1, 6, and 11 recite all of the limitations of instant claims 1, 7, and 13 respectively, except for “a determination standard for the determination is controlled based on information obtained from another probability distribution estimated by encrypting the target input data with the parameters after the adjusting.” However, Segev discloses “a determination standard for…[a] determination is controlled based on information obtained from… [a] probability distribution…” (Segev, Fig. 4B: “Apply Gaussian Mixture with EM Algorithm to the distribution of MDDPs -> Set the threshold that classifies each MDDP or NAMDDP as either normal or abnormal; and (128): “Another method for identifying abnormal data MDDPs in the embedding matrix Ψ (steps 330′ in FIGS. 3A and 3B) can be based on Gaussian mixture fit and threshold estimation as shown schematically herein with respect to steps 506B-510B in FIG. 4B. The average distance Dnn between each MDDP in the embedded space and its nearest neighbors may be determined (step 506B, FIG. 4B), followed by performing a Gaussian mixture approximation to the distribution of Dnn (step 508B, FIG. 4B) using, for example, the expectation-maximization function (see DLR). Namely the Gaussian mixture can be represented by P(Dnn) = [equation] where wi are the Gaussian weights and G(x|μi ,σi) are the Gaussians densities, with μi is the Gaussian mean of Gaussian i and σi as the standard deviation of Gaussian i; and (133) The threshold Dnnt (also: “T”) can be defined, for example, to be the distance for which P(i|x, μi, σi, maxi(μi+σi)) ≈ 0.9… A new MDPP is declared as an anomalous MDDP if its Dnn is larger than Dnnt; the examiner notes that threshold Dnnt corresponds to a determination standard for the determination, which is set based on Gaussian i, mean μi, and standard deviation σi obtained from probability distribution P(Dnn)).
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the method of claims 1, 6, 11 of the reference patent to include a determination standard for the determination being controlled based on information obtained from another probability distribution, as disclosed by Segev, and one would have been motivated to do so for the purpose of determining whether or not data is normal by how closely it fits the identified Gaussian (see Segev, (130-131)).
Segev does not appear to explicitly disclose that the probability distribution is “estimated by encrypting the target input data with the parameters after the adjusting.”
However, Zong discloses “[a] probability distribution estimated by encrypting… target input data with… parameters after… adjusting” (Zong, 3.3 Estimation Network: “In addition, during the testing phase with the learned GMM parameters, it is straightforward to estimate sample energy, and predict samples of high energy as anomalies by a pre-chosen threshold”; the examiner notes that during the testing phase, the adjusted parameters from training (see 3.4 eq 7) are used to encrypt target input data (see 3.2 eq 1) to estimate another probability distribution and its corresponding sample energy (see 3.3 eq 6)).
It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to have modified the combination of the reference patent and Segev to include the probability distribution being estimated by encrypting the target input data with the parameters after the adjusting, as disclosed by Zong, and one would have been motivated to do so for the purpose of improving the accuracy of unsupervised anomaly detection (see Zong, Abstract).
Regarding instant claims 3, 9, and 15, claims 3, 8, and 13 of the reference patent recite the same further limitations as claims 3, 9, and 15 respectively of the instant application.
Regarding instant claims 4, 10, and 16, claims 4, 9, and 14 of the reference patent recite the same further limitations as claims 4, 10, and 16 respectively of the instant application.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure: Ballé et al. “Variational Image Compression with a Scale Hyperprior”; Zhou et al. “Multi-scale and Context-adaptive Entropy Model for Image Compression.”
Any inquiry concerning this communication or earlier communications from the examiner should be directed to GWYNEVERE A DETERDING whose telephone number is (571)272-7657. The examiner can normally be reached Mon-Fri. 9am-5pm.
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, Kamran Afshar can be reached at (571) 272-7796. 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.
/G.A.D./Examiner, Art Unit 2125
/KAMRAN AFSHAR/Supervisory Patent Examiner, Art Unit 2125