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 .
Claim Objections
Claims 1 and 3 is objected to because of the following informalities:
Claim 1 recites “at least one processor that is configured to execute the instructions to” and the claim should add a “:” and recite “at least one processor that is configured to execute the instructions to:”.
Claim 3 recites “at least one processor is configured to execute the instructions to calculate, as the classification index,” and the claim should recited “at least one processor is configured to execute the instructions to calculate as the classification index”.
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 rejected under 35 U.S.C. 101 because the claimed invention is directed to abstract idea without significantly more. The claim(s) recite(s) significantly more. The subject matter eligibility test for products and process is describe below for claim 1 in view of dependent claims.
Regarding claim 1:
Step 1: Is the claim to a process machine manufacture or composition of matter?
Yes – Claim 1 recites an apparatus, which is a system that falls under the statutory categories.
Step 2A Prong 1: Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes – The claim recites the following:
“calculate a classification index indicating a likelihood of a class to which the series data belong, on the basis of at least two elements of the plurality of elements;” - The limitations recites a mathematical process for calculating a classification index (see MPEP 2106.04(a)(2)I).
“select output, from N classes that are classification candidates of the series data (where N is a natural number), K classes to which the series data are likely to belong (where K is a natural number that is less than or equal to N and that is greater than or equal to 1), on the basis of the classification index.”- The limitations recites a mental process of selecting N classes that are candidates and selecting K classes and to output them based on the classification index (see MPEP 2106.04(a)(2)III). The limitations also recite a mathematical concept of N and K being natural numbers where K is less than or equal to N and that is greater than or equal to 1 (see MPEP 2106.04(a)(2)I).
Step 2 Prong 2: Does the claim recite additional elements that integrate the judicial exception into a particular application? No –
The claim includes the additional element(s):
“An information processing apparatus comprising: at least one memory that is configured to store instructions; and at least one processor that is configured to execute the instructions to obtain a plurality of elements included in series data;”
The additional elements fall under “apply it” as using a generic computer to execute instructions. See Mere Instructions to Apply an Exemption (see MPEP 2106.05(f)).
The additional elements fall under Insignificant Extra-Solution Activity as mere data gathering. See MPEP 2106.5(g). The judicial exemptions do not integrate into a practical application nor provide an improvement. The process does not provide an inventive concept nor provides a practical application.
Step 2B: Does the claim recite additional elements that amount to significantly more than the judicial exception?
No - The claim does not include additional elements that are sufficient to amount to a significantly more than the judicial exemption. As an order whole, the claim is directed processing data and calculating a classification index. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements of processing data. The method does not improve on the function of a computer, transforms an article into another article, nor is it applied by a particular machine, making the claim not patent eligible.
Regarding claim 2:
Step 2A Prong 1: Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes – The claim recites the following:
““calculate a classification index indicating a likelihood of a class to which the series data belong, on the basis of at least two elements of the plurality of elements;” - The limitations recites a mathematical process for calculating a classification index of each of the N classes (see MPEP 2106.04(a)(2)I).
Step 2A Prong 2, Step 2B: The additional element(s):
No additional elements. The judicial exemptions do not integrate into a practical application nor provide an improvement. The process does not provide an inventive concept nor provides a practical application.
Regarding claim 3:
Step 2A Prong 1: Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes – The claim recites the following:
“The information processing apparatus according to claim 2, wherein at least one processor is configured to execute the instructions to calculate, as the classification index, a likelihood ratio that is a ratio between a first likelihood indicating a likelihood that the series data belong to a class and a second likelihood indicating a likelihood that the series data do not belong to a class.” - The limitations recites a mathematical process for calculating a ratio between a first and second likelihood that will be the classification index (see MPEP 2106.04(a)(2)I).
Step 2A Prong 2, Step 2B: The additional element(s):
No additional elements. The judicial exemptions do not integrate into a practical application nor provide an improvement. The process does not provide an inventive concept nor provides a practical application.
Regarding claim 4:
Step 2A Prong 2, Step 2B: The additional element(s):
“The information processing apparatus according to claim 2,wherein the at least one processor is configured to execute the instructions to select the K classes in order from an earliest arrival time when the classification index reaches a predetermined threshold, on the basis of the classification index calculated for each of the N classes.”
The additional elements fall under “apply it” as using a generic computer to execute instructions to select the earliest K classes with the earliest arrival time. See Mere Instructions to Apply an Exemption (see MPEP 2106.05(f)).
Regarding claim 5:
Step 2A Prong 2, Step 2B: The additional element(s):
“The information processing apparatus according to claim 2, wherein at least one processor is configured to execute the instructions to select the K classes in descending order of the classification index in a predetermined time range, on the basis of the classification index calculated for each of the N classes.”
The additional elements fall under “apply it” as using a generic computer to execute instructions to select the K classes in descending order of the classification index. See Mere Instructions to Apply an Exemption (see MPEP 2106.05(f)).
Regarding claim 6:
Step 2A Prong 1: Does the claim recite an abstract idea, law of nature, or natural phenomenon?
at least one processor is configured to execute the instructions to calculate the classification index by considering continuity of the at least two elements.”
The limitations recites a mental process of calculating the classification index by considering continuity (see MPEP 2106.04(a)(2)III).
Step 2A Prong 2, Step 2B: The additional element(s):
“The information processing apparatus according to claim1, wherein the acquisition unit at least one processor is configured to execute the instructions to sequentially obtain the plurality of elements, and ”
The additional elements fall under “apply it” as using a generic computer to execute instructions to sequentially obtain the plurality of elements. See Mere Instructions to Apply an Exemption (see MPEP 2106.05(f)).
Regarding claim 7:
Step 2A Prong 2, Step 2B: The additional element(s):
“The information processing apparatus according to claims1 wherein the at least one processor is configured to execute the instructions to rank and outputs output the selected K classes in descending order of a possibility that the serial data belong to the class.”
The additional elements fall under “apply it” as using a generic computer to execute instructions to rank the output the k classes in descending order based on the possibility that it belongs to a class. See Mere Instructions to Apply an Exemption (see MPEP 2106.05(f)).
Regarding claim 8:
Step 2A Prong 1: Does the claim recite an abstract idea, law of nature, or natural phenomenon?
“The information processing apparatus according to claim1wherein the at least one processor is configured to execute the instructions to perform weighting about ease of selection, on each of the N classes, and then selects the K classes.”
The limitations recites a mathematical process of weighting the ease of selection of K classes (see MPEP 2106.04(a)(2)III).
Step 2A Prong 2, Step 2B: The additional element(s):
No additional elements. The judicial exemptions do not integrate into a practical application nor provide an improvement. The process does not provide an inventive concept nor provides a practical application.
Claims 9 recite a method and is analogous to the system of claim 1. Therefore, the rejections of claim 1 above applies to claims 9.
Claims 10 recite a CRM claim and is analogous to the system of claim 1. Therefore, the rejections of claim 1 above applies to claims 10.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1, 2, 6, 9, and 10 are rejected under 35 U.S.C. 103 as being unpatentable over Ariyoshi et al. (US20150112900A1) (“Ariyoshi”) in view of Takimoto (US20180246846A1) (“Takimoto”).
Regarding claim 1 and analogous claims 9 and 10, Ariyoshi teaches An information processing apparatus comprising: at least one memory that is configured to store instructions;
and at least one processor that is configured to execute the instructions to obtain a plurality of elements included in series data (Ariyoshi para 0045, FIG. 1 is a diagram showing a time-series data prediction algorithm performed by a time-series data prediction device according to an embodiment of the present invention. In the present embodiment, a case of time-series data of the observation value of the energy demand in a house will be described as an example [An information processing apparatus comprising].
Para 0061, The acquisition unit 32 reads time-series data from the time-series data storage device 2. According to the setting file, the acquisition unit 32 divides the read time-series data into prediction data used for the prediction of energy demand and learning data used for the generation of a prediction model, and further divides the learning data into training data used for the model learning of the latest prediction model (no-cluster prediction model and cluster-specific prediction model) and test data used for the evaluation of a prediction model [obtain a plurality of elements included in series data].
Para 0116, In addition, the process of the operation of the timeseries data prediction device 3 is stored in a computer-readable recording medium in the form of a program, and the processing described above is performed by reading and executing the program using a computer system. The 'computer system' referred to herein includes a CPU, various memories or an OS, and hardware, such as peripheral devices [and at least one processor that is configured to execute the instructions].
Para 0168 line 1-4, Examples of the "computer-readable recording medium" include portable media, such as a flexible disk, a magneto-optical disc, a ROM, and a CD-ROM, and a storage device, such as a hard disk built into a computer system [at least one memory that is configured to store instructions;]);
calculate a classification index indicating a likelihood of a class to which the series data belong, on the basis of at least two elements of the plurality of elements (Ariyoshi Para 0146, As shown in FIG. 11, when generating a prediction result based on the cluster-specific prediction model, the evaluation unit 35 generates a classification model for calculating the cluster proximity index that is a probability that the test data for evaluation belongs to each class (hereinafter, referred to as a "class belonging probability") [calculate a classification index indicating a likelihood of a class to which the series data belong]. A class is a number of the cluster. Assuming that the class belonging probability of a class c (c=1, 2, . . . , C) is P.sub.c and the cluster of the class c is S.sub.c, the class belonging probability P.sub.c of the class c indicates a probability that test data (test data for evaluation) belongs to the cluster S.sub.c. Sparse Logistic Regression (SLR) using the probability derivation expression of logistic regression analysis is used for the generation of the classification model.
Para 149, After reading the training data for prediction from the storage unit 31 (step S510), the evaluation unit 35 generates a classification model using the feature amount for clustering extracted from the read training data for prediction (step S520).
Para 0153, The evaluation unit 35 generates a derivation expression of the class belonging probability shown in the following Expression (21), that is, a classification model, using θ determined so as to maximize Expression (20). In addition, t indicates transposition, and Xn is input data of the classification target.
Para 00152, From the above, the relationship between the input x and the output y when there are N pieces of training data for prediction with a known determination result is expressed by the following Expression (20).
Para 0154, Using Expression (21) that is the classification model generated in step S520, the evaluation unit 35 calculates class belonging probabilities P n (I) to P n (c) of the test data for evaluation with the feature amount for clustering acquired from the test data for evaluation as Xn (step S530) [on the basis of at least two elements of the plurality of elements; and]);
Ariyoshi does not explicitly teach select and output, from N classes that are classification candidates of the series data (where N is a natural number), K classes to which the series data are likely to belong (where K is a natural number that is less than or equal to N and that is greater than or equal to 1), on the basis of the classification index.
However Takimoto teaches select and output, from N classes that are classification candidates of the series data (where N is a natural number), K classes to which the series data are likely to belong (where K is a natural number that is less than or equal to N and that is greater than or equal to 1), on the basis of the classification index (Takimoto FIG. 7A
PNG
media_image1.png
224
458
media_image1.png
Greyscale
para 0034, As display which allows the classifier and the user to interactively update a classification standard, a distribution chart can be produced. The distribution chart shows a data distribution of the original feature space or indicates a projection result obtained by the projection in a feature space of equal to or lower than three-dimension through processing such as principal component analysis (PCA) to visualize the data distribution of the original feature space. FIG. 5 is a diagram illustrating an example of a distribution chart 500. In the distribution chart 500 in FIG. 5, data is described in a high-dimensional feature quantity. In the distribution chart 500, a distribution of data classified into three classes, i.e., class 1 to class 3, is visualized and expressed in the three-dimensional feature space through supervised dimensionality reduction.
Para 038 line 10-18, In step S602, the data evaluation unit 202 specifies a class to which the data belongs and reliability thereof with respect to all of the input data. Herein, the class corresponds to a type of abnormality. Further, the reliability is a value indicating likelihood that the data belongs to the class, and the reliability can be expressed by probability of the data belonging to the class. The processing in step S602 is an example of class determination processing or reliability determination processing [on the basis of the classification index]
Para 0045, FIG. 7A is a diagram illustrating an example of a distribution chart 700 displayed in step S604. The distribution chart 700 in FIG. 7A is a two-dimensional graph in which a horizontal axis indicates a class type and a vertical axis indicates reliability with respect to the class. In addition, the reliability is normalized to 0 to 1. In the distribution chart 700 in FIG. 7A, data is classified into five classes, so that values corresponding to the five classes are assigned [select and outputs, output, from N classes that are classification candidates of the series data]. Further, dot images corresponding to data are plotted in the distribution chart 700. Black dot images correspond to label-instructed data, whereas white dot images correspond to label-uninstructed data. The label-instructed data accord with determination of a class corresponding to the user operation, and is data to which a class label has been instructed (given). The label-uninstructed data is data to which the class label has not been instructed. [K classes to which the series data are likely to belong (where K is a natural number that is less than or equal to N and that is greater than or equal to 1),]
Para 0046, As described above, because the information processing apparatus 100 displays the label-instructed data and the label-uninstructed data in different colors such as black and white, the user can easily distinguish between the data in the distribution chart. In addition, because the user can distinguish between both data if the information processing apparatus 100 displays the label-instructed data and the label-uninstructed data in different display forms, a specific display form is not limited to the display form described in the present exemplary embodiment. (Examiner Notes: The system displays all the selected classes and what class the data belongs in)).
Ariyoshi and Takimoto are considered to be analogous to the claim invention because they are in the same field of machine learning. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filling date of the claimed invention to have modified Ariyoshi to incorporate the teachings of Takimoto and select the class that the data series likely belongs by determining its reliability. Doing so to determine the reliability of the likelihood that the input data belong to a class and displaying the output to a user to support the user operations (Takimoto para 0023, An information processing apparatus according to a first exemplary embodiment uses a plurality of data pieces expressed by a plurality of feature quantities as learning data and generates a classifier that identifies a class to which the learning data belong. Further, when the classifier is generated, the information processing apparatus of the present exemplary embodiment visualizes data appropriate for supporting the user operation of assigning labels for assorting data into classes.
para 0038, In step S602, the data evaluation unit 202 specifies a class to which the data belongs and reliability thereof with respect to all of the input data. Herein, the class corresponds to a type of abnormality. Further, the reliability is a value indicating likelihood that the data belongs to the class, and the reliability can be expressed by probability of the data belonging to the class. The processing in step S602 is an example of class determination processing or reliability determination processing.).
Regarding claim 2, Ariyoshi in view of Takimoto teach the information processing apparatus according to claim 1.
Ariyoshi further teaches wherein the calculation unit calculates at least one processor is configured to execute the instructions to calculate the classification index for each of the N classes (Ariyoshi FIG. 11,
PNG
media_image2.png
411
640
media_image2.png
Greyscale
Para 0146, As shown in FIG. 11, when generating a prediction result based on the cluster-specific prediction model, the evaluation unit 35 generates a classification model for calculating the cluster proximity index that is a probability that the test data for evaluation belongs to each class (hereinafter, referred to as a "class belonging probability") [calculate the classification index for each of the N classes]. A class is a number of the cluster. Assuming that the class belonging probability of a class c ( c= 1, 2, ... , C) is Pc and the cluster of the class c is Sc, the class belonging probability Pc of the class c indicates a probability that test data (test data for evaluation) belongs to the cluster Sc. Sparse Logistic Regression (SLR) using the probability derivation expression of logistic regression analysis is used for the generation of the classification model.
Para 0116, Para 0116, In addition, the process of the operation of the timeseries data prediction device 3 is stored in a computer-readable recording medium in the form of a program, and the processing described above is performed by reading and executing the program using a computer system. The 'computer system' referred to herein includes a CPU, various memories or an OS, and hardware, such as peripheral devices [wherein the calculation unit calculates at least one processor is configured to execute the instructions])).
Regarding claim 6, Ariyoshi in view of Takimoto teach the information processing apparatus according to claim 1.
Ariyoshi teaches wherein the at least one processor is configured to execute the instructions to sequentially obtain the plurality of elements, and the calculation unit calculates at least one processor is configured to execute the instructions to calculate the classification index by considering continuity of the at least two elements (Ariyoshi Para 0135, FIG. 10 is a flowchart showing the operation in the prediction model selection process of the evaluation unit 35, and shows the detailed process of step S150 in FIG. 4.
0136 First, the evaluation unit 35 extracts a feature amount for clustering and a feature amount, which is used as an input parameter of the prediction model, from test data for evaluation (step S410). The evaluation unit 35 calculates a prediction result of the time-series data of the prediction target period of2 days based on the cluster-specific prediction model using the extracted feature amount of the test data for evaluation as an input parameter in the no-cluster prediction model generated in step S140 (step S420). That is, the evaluation unit 35 calculates a predicted value by using the feature amount extracted from the test data for evaluation (timeseries data Xn_7 , Xn_ 6, and Xn_ 5) as an input parameter in the approximation model of each element of the prediction target period of 2 days that configures a no-cluster prediction model, and obtains prediction results Xn_4' and Xn_3' of the timeseries data of the prediction target period of 2 days that is the calculated predicted value of each element.
Para 0149, After reading the training data for prediction from the storage unit 31 (step S510), the evaluation unit 35 generates a classification model using the feature amount for clustering extracted from the read training data for prediction (step S520).
Para 0154 line 1-6 Using Expression (21) that is the classification model generated in step S520, the evaluation unit 35 calculates class belonging probabilities
P
n
(
1
)
to
P
n
(
c
)
of the test data for evaluation with the feature amount for clustering acquired from the test data for evaluation as Xn (step S530) [at least one processor is configured to execute the instructions to calculate the classification index by considering continuity of the at least two elements]).
Claim 3 is rejected under 35 U.S.C. 103 as being unpatentable over Ariyoshi in view of Takimoto and further in view of B. Tang, S. Kay and H. He, "Toward Optimal Feature Selection in Naive Bayes for Text Categorization," in IEEE Transactions on Knowledge and Data Engineering, vol. 28, no. 9, pp. 2508-2521, 1 Sept. 2016, doi: 10.1109/TKDE.2016.2563436 (“Tang”).
Regarding claim 3, Ariyoshi in view of Takimoto teach the information processing apparatus according to claim 1.
Ariyoshi and Takimoto are combined with the same rationale used in claim 1 and analogous claims 9 and 10.
Ariyoshi does not explicitly teach wherein
Tang teaches wherein between a first likelihood indicating a likelihood that the series data belong to a class and a second likelihood indicating a likelihood that the series data do not belong to a class ((Tang page 2511 3.1 Divergence Measures for Binary Hypothesis Testing para 1, Considering a two-class classification problem first, each class is represented by a particular distribution, saying
P
1
=
p
(
x
|
c
1
;
θ
1
)
for class
c
1
and
P
2
=
p
(
x
|
c
2
;
θ
2
)
for class
c
2
. A test procedure for classification can be considered as a binary hypothesis testing such that if a sample is drawn from P1 we accept the hypothesis
H
1
(reject the hypothesis
H
2
), and if a sample is drawn from P2 we accept H2 (reject H1). In other words, we have
p
x
c
1
=
p
x
H
1
and
p
x
c
2
=
p
x
H
2
, and we also denote
p
x
H
i
as the class conditional probability distribution in the rest of paper.
Para 3,
PNG
media_image3.png
192
781
media_image3.png
Greyscale
[a likelihood ratio that is a ratio between a first likelihood indicating a likelihood that the series data belong to a class and a second likelihood indicating a likelihood that the series data do not belong to a class]).
Ariyoshi and Tang are considered to be analogous to the claim invention because they are in the same field of machine learning. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filling date of the claimed invention to have modified Ariyoshi to incorporate the teachings of Tang and calculate a ration between a likelihood of data being in a class and not belonging to a class. Doing so to determine a measure discrimination in favor for one class (Tang page 2521 3.1 Divergence Measures for Binary Hypothesis Testing para 3,
PNG
media_image3.png
192
781
media_image3.png
Greyscale
)
Claim(s) 4 is rejected under 35 U.S.C. 103 as being unpatentable over Ariyoshi in view of Takimoto and further in view of Mori, U., Mendiburu, A., Keogh, E. et al. Reliable early classification of time series based on discriminating the classes over time. Data Min Knowl Disc 31, 233–263 (2017), (“Mori”).
Regarding claim 4, Ariyoshi in view of Takimoto teach the information processing apparatus according to claim 2.
Ariyoshi and Takimoto are combined with the same rationale used in claim 1 and analogous claims 9 and 10.
Ariyoshi does not explicitly teach wherein the at least one processor is configured to execute the instructions to select the K classes in order from an earliest arrival time when the classification index reaches a predetermined threshold, on the basis of the classification index calculated for each of the N classes.
Mori teaches wherein the at least one processor is configured to execute the instructions to select the K classes in order from an earliest arrival time when the classification index reaches a predetermined threshold, on the basis of the classification index calculated for each of the N classes (Mori Page 238, Figure 2 shows an example timeline with four relevant timestamps: t1, t2, t3 and t4. These instants are ordered in time and, at each of them, a class or set of classes becomes safe. For example, at instant t3 we can begin to discriminate class C5 from the rest of the classes. If we consider the classes that have appeared earlier in the timeline, we can conclude that at timestamp t3, classes C2, C4, C1 and C5 can be predicted accurately. As will be shown in the next steps, predictions will only be provided at the timestamps that appear in the timeline or later and, by means of this mechanism, a large number of calculations can be avoided when making predictions for new time series.
Page 238, Fig. 2
PNG
media_image4.png
101
668
media_image4.png
Greyscale
[ to select the K classes in order from an earliest arrival time when the classification index reaches a predetermined threshold]
Page 244 4.2 Prediction phase para 2, Moreover, the series that are assigned to a safe class must pass the test of the prediction reliability. The differences between the predicted class probabilities obtained from the classification will be compared to the thresholds extracted in Step 2. Only if they are larger, will a final class be given. If not, the series will not be classified and will continue in the process and wait until enough new data points are available.
Page 238,
PNG
media_image5.png
348
656
media_image5.png
Greyscale
[on the basis of the classification index calculated for each of the N classes]).
Ariyoshi and Mori are considered to be analogous to the claim invention because they are in the same field of machine learning. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filling date of the claimed invention to have modified Ariyoshi to incorporate the teachings of Mori and select the classes in order of the earliest one to predicted accurately. Doing so to determine what classes can be predicted accurately (Mori Page 238, Figure 2 shows an example timeline with four relevant timestamps: t1, t2, t3 and t4. These instants are ordered in time and, at each of them, a class or set of classes becomes safe. For example, at instant t3 we can begin to discriminate class C5 from the rest of the classes. If we consider the classes that have appeared earlier in the timeline, we can conclude that at timestamp t3, classes C2, C4, C1 and C5 can be predicted accurately.
As will be shown in the next steps, predictions will only be provided at the timestamps
that appear in the timeline or later and, by means of this mechanism, a large number of calculations can be avoided when making predictions for new time series.)
Claim(s) 5 is rejected under 35 U.S.C. 103 as being unpatentable over Ariyoshi in view of Takimoto and further in view of Mori and further in view of Goa et al. (US20020152069A1) (“Goa”).
Regarding claim 5, Ariyoshi in view of Takimoto teach the information processing apparatus according to claim 2.
Ariyoshi and Takimoto are combined with the same rationale used in claim 1 and analogous claims 9 and 10.
Ariyoshi and Mori are combined with the same rationale used in claim 5.
Ariyoshi does not explicitly teach wherein the at least one processor is configured to execute the instructions to select the K classes in descending order of the classification index in a predetermined time range, on the basis of the classification index calculated for each of the N classes.
However Mori teaches wherein the at least one processor is configured to execute the instructions to select the K classes [in descending order] of the classification index in a predetermined time range, on the basis of the classification index calculated for each of the N classes (Mori Page 238, Fig. 2
PNG
media_image4.png
101
668
media_image4.png
Greyscale
244 4.2 Prediction phase para 2, Moreover, the series that are assigned to a safe class must pass the test of the prediction reliability. The differences between the predicted class probabilities obtained from the classification will be compared to the thresholds extracted in Step 2. Only if they are larger, will a final class be given. If not, the series will not be classified and will continue in the process and wait until enough new data points are available.
Page 238,
PNG
media_image5.png
348
656
media_image5.png
Greyscale
[index in a predetermined time range]).
Goa teaches [wherein the at least one processor is configured to execute the instructions to select the K classes] in descending order [of the classification index in a predetermined time range, on the basis of the classification index calculated for each of the N classes] (Goa Para 0046, In some embodiments of the invention, the observation vectors can be frames of speech and the feature vectors can be acoustic feature vectors. The pattern can be a time waveform cor/responding to speech, such that x.sub.1, x.sub.2, . . . x.sub.N can be represented as {right arrow over (x)}(t), where t is time.
Para 0054, Reference should now be had to FIG. 5, which depicts a partial flow chart showing optional sub-steps which can be performed during the combination of the feature vectors in block 110 of FIG. 1. Specifically, all classes can be ranked for each of the feature vectors, per block 136, and then a merged rank list can be generated by picking that class from among each of the feature vectors which yields the highest rank. This permits discriminating among the correct and incorrect ones of the classes. As shown in block 136, we can rank the likelihoods obtained from each model in a descending order such that all the classes are ranked for each feature vector. Per block 138, one can select the state with the highest rank from the rank ordered list R.sub.1, R.sub.2, . . . of each model corresponding to each feature vector so as to obtain the merged rank list.) [in descending order].
Ariyoshi and Goa are considered to be analogous to the claim invention because they are in the same field of machine learning. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filling date of the claimed invention to have modified Ariyoshi to incorporate the teachings of Goa and rank the output in deciding order. Doing so to that one can select the state with the highest rank from the rank order list (Goa Para 0054, Reference should now be had to FIG. 5, which depicts a partial flow chart showing optional sub-steps which can be performed during the combination of the feature vectors in block 110 of FIG. 1. Specifically, all classes can be ranked for each of the feature vectors, per block 136, and then a merged rank list can be generated by picking that class from among each of the feature vectors which yields the highest rank. This permits discriminating among the correct and incorrect ones of the classes. As shown in block 136, we can rank the likelihoods obtained from each model in a descending order such that all the classes are ranked for each feature vector. Per block 138, one can select the state with the highest rank from the rank ordered list R.sub.1, R.sub.2, . . . of each model corresponding to each feature vector so as to obtain the merged rank list.).
Claim(s) 7 is rejected under 35 U.S.C. 103 as being unpatentable over Ariyoshi in view of Takimoto and further in view of Goa.
Regarding claim 7, Ariyoshi in view of Takimoto teach the information processing apparatus according to claim 1.
Ariyoshi and Takimoto are combined with the same rationale used in claim 1 and analogous claims 9 and 10.
Ariyoshi does not explicitly teach wherein the at least one processor is configured to execute the instructions to rank and output the selected K classes in descending order of a possibility that the serial data belong to the class.
Goa teaches wherein the at least one processor is configured to execute the instructions to rank and output the selected K classes in descending order of a possibility that the serial data belong to the class (Para 0054, Reference should now be had to FIG. 5, which depicts a partial flow chart showing optional sub-steps which can be performed during the combination of the feature vectors in block 110 of FIG. 1. Specifically, all classes can be ranked for each of the feature vectors, per block 136, and then a merged rank list can be generated by picking that class from among each of the feature vectors which yields the highest rank. This permits discriminating among the correct and incorrect ones of the classes. As shown in block 136, we can rank the likelihoods obtained from each model in a descending order such that all the classes are ranked for each feature vector. Per block 138, one can select the state with the highest rank from the rank ordered list R.sub.1, R.sub.2, . . . of each model corresponding to each feature vector so as to obtain the merged rank list. [rank and output the selected K classes in descending order of a possibility that the serial data belong to the class]).
Ariyoshi and Goa are considered to be analogous to the claim invention because they are in the same field of machine learning. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filling date of the claimed invention to have modified Ariyoshi to incorporate the teachings of Goa and rank the output in deciding order. Doing so to that one can select the state with the highest ran from the rank order list (Goa Para 0054, Reference should now be had to FIG. 5, which depicts a partial flow chart showing optional sub-steps which can be performed during the combination of the feature vectors in block 110 of FIG. 1. Specifically, all classes can be ranked for each of the feature vectors, per block 136, and then a merged rank list can be generated by picking that class from among each of the feature vectors which yields the highest rank. This permits discriminating among the correct and incorrect ones of the classes. As shown in block 136, we can rank the likelihoods obtained from each model in a descending order such that all the classes are ranked for each feature vector. Per block 138, one can select the state with the highest rank from the rank ordered list R.sub.1, R.sub.2, . . . of each model corresponding to each feature vector so as to obtain the merged rank list.).
Claim(s) 8 is rejected under 35 U.S.C. 103 as being unpatentable over Ariyoshi in view of Takimoto and further in view of M. Zhu et al., "Class Weights Random Forest Algorithm for Processing Class Imbalanced Medical Data," in IEEE Access, vol. 6, pp. 4641-4652, 2018, (“Zhu”).
Regarding claim 8, Ariyoshi in view of Takimoto teach the information processing apparatus according to claim 1.
Ariyoshi does not explicitly teach wherein the at least one processor is configured to execute the instructions to perform weighting about ease of selection, on each of the N classes, and then selects the K classes.
Zhu teaches wherein the at least one processor is configured to execute the instructions to perform weighting about ease of selection, on each of the N classes, and then selects the K classes (Zhu page 4642 I. Introduction para 4, Therefore, we proposed a class weights voting (CWsRF) algorithm based on random forest algorithm (RF), which contains an approach (CWsV) and trains a collection of classifiers in different weights per class to combine the output of each classifier into an ultimate prediction [87]. Different weights per class are obtained from the empirical error of different classifiers. The algorithm assigns individual weights for each class instead of a single weight and focuses on the problem of effective identification for the minority class. It can improve the recognition performance for minority class while maintaining those for majority class [weighting about ease of selection].
Page 4643,
PNG
media_image6.png
507
496
media_image6.png
Greyscale
The classification capability of each classifier is often used to evaluate the weight of a classifier; therefore the classifier's prior accuracy (ACC) is used to measure the different weight per class (W / ACC); and the weights were used to calculate votes.
Page 4644,
PNG
media_image7.png
633
473
media_image7.png
Greyscale
[on each of the N classes, and then selects the K classes.]).
Ariyoshi and Zhu are considered to be analogous to the claim invention because they are in the same field of machine learning. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filling date of the claimed invention to have modified Ariyoshi to incorporate the teachings of Zhu and weight each class. Doing so to weight the decision of combining classifiers and effectively identify the minority class (Zhu Page 4642 I. Introduction para 3 line 1-17, based on random forest (RF) algorithm. Random forest classifiers can achieve high accuracy in data classification compared with many standard classification methods [65]_[68], and it can minimize the overall classification error rate and has the ability to handle class imbalanced data [4], [56], [69]. However, when the imbalanced rate increases (e.g., 15%), the classification ability is weakened [56], [70], [71], it is because that each classifier has the same weight when the classifiers are combined [70], [72]. Therefore, to solve this problem, a method of combining classifiers with different weights is proposed. Weight voting random forest (WRF), an algorithm of getting the decisions of each classifier is multiplied by a weight to reflect the individual confidence of these decisions [73]_[75]. A weighted majority voting method which is based on class-conditional independence of the classifier outputs is proposed [76].
page 4642 I. Introduction para 4, Therefore, we proposed a class weights voting (CWsRF) algorithm based on random forest algorithm (RF), which contains an approach (CWsV) and trains a collection of classifiers in different weights per class to combine the output of each classifier into an ultimate prediction [87]. Different weights per class are obtained from the empirical error of different classifiers. The algorithm assigns individual weights for each class instead of a single weight and focuses on the problem of effective identification for the minority class. It can improve the recognition performance for minority class while maintaining those for majority class).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ALFREDO CAMPOS whose telephone number is (571)272-4504. The examiner can normally be reached 7:00 - 4:00 pm M - F.
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, Michael J. Huntley can be reached at (303) 297-4307. 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.
/ALFREDO CAMPOS/Examiner, Art Unit 2129
/MICHAEL J HUNTLEY/Supervisory Patent Examiner, Art Unit 2129