CTNF 18/338,732 CTNF 101910 DETAILED ACTION This communication is in response to the Application No. 18/338,732 filed on June 21, 2023 in which Claims 1-15 are presented for examination. Notice of Pre-AIA or AIA Status 07-03-aia AIA 15-10-aia The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA. Information Disclosure Statement The information disclosure statement (IDS) submitted on 6/21/2023 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner. Claim Rejections - 35 USC § 112 07-30-02 AIA The following is a quotation of 35 U.S.C. 112(b): (b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention. The following is a quotation of 35 U.S.C. 112 (pre-AIA), second paragraph: The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention. 07-34-01 Claims 5, 6 and 10-15 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention. Claims 5, 6, 10, 14 and 15 recite the limitation "the student model" without any prior recitation of such a student model in Independent Claim 1 and Independent Claim 10. There is insufficient antecedent basis for this limitation in the claim. In reference to dependent claims 11-13, claims 11-13 do not cure the deficiencies noted in the rejection of independent claim 10. Therefore, these claims are rejected under the same rationale as claim 10. Claim Rejections - 35 USC § 101 07-04-01 AIA 07-04 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 1-15 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. Regarding Claim 1 : Step 1 : Claim 1 is a method claim. Therefore, Claims 1-8 are directed to either a process, machine, manufacture, or composition of matter. Step 2A Prong 1 : selecting classes from a database comprising a set of classes (mental process – selecting classes from a database comprising a set of classes may be performed manually by a user with the aid of pen and paper by observing/analyzing a set of classes in a database) generating a mean feature group comprising mean features extracted from the selected classes (mental process – generating a mean feature group comprising mean features extracted from the selected classes may be performed manually by a user with the aid of pen and paper by observing/analyzing mean features extracted from the selected classes) determining a first similarity between the extracted feature and a mean feature corresponding to the input data (mental process – determining a first similarity between the extracted feature and a mean feature corresponding to the input data may be performed manually by a user with the aid of pen and paper by observing/analyzing the extracted feature and a mean feature corresponding to the input data) determining a second similarity comprising a self-similarity of the mean feature (mental process – determining a second similarity comprising a self-similarity of the mean feature may be performed manually by a user with the aid of pen and paper by observing/analyzing the mean feature) Step 2A Prong 2 : This judicial exception is not integrated into a practical application. Additional Elements: a method of training a neural network model (recited at a high-level of generality (i.e., as generic a neural network) such that it amounts to no more than mere instructions to apply the exception using generic computer components) receiving a batch comprising input data and extracting, by the neural network model, a feature from the input data (Adding insignificant extra-solution activity to the judicial exception – see MPEP 2106.05(g)) wherein the neural network model is to be trained according to a mean feature set (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 – see MPEP 2106.05(f)) updating a parameter of the neural network model based on the first similarity and the second similarity (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 – see MPEP 2106.05(f)) Step 2B : The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. Additional Elements: a method of training a neural network model (mere instructions to apply the exception using generic computer components cannot provide an inventive concept) receiving a batch comprising input data and extracting, by the neural network model, a feature from the input data (MPEP 2106.05(d)(II) indicates that merely “Receiving or transmitting data over a network” is a well-understood, routine, conventional function when it is claimed in a merely generic manner (as it is in the present claim). Thereby, a conclusion that the claimed limitation is well-understood, routine, conventional activity is supported under Berkheimer) wherein the neural network model is to be trained according to a mean feature set (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 – see MPEP 2106.05(f) – Examiner’s note: high level recitation of applying a neural network model to previously determined data without significantly more. This cannot provide an inventive concept) updating a parameter of the neural network model based on the first similarity and the second similarity (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 – see MPEP 2106.05(f) – Examiner’s note: high level recitation of updating a neural network model based on the determined data without significantly more. This cannot provide an inventive concept) For the reasons above, Claim 1 is rejected as being directed to an abstract idea without significantly more. This rejection applies equally to dependent claims 2-8. The additional limitations of the dependent claims are addressed below. Regarding Claim 2 : Step 2A Prong 1 : selecting a first number of classes in ascending order of a variance feature from among classes in the database (mental process – selecting a first number of classes in ascending order of a variance feature from among classes in the database may be performed manually by a user with the aid of pen and paper by observing/analyzing a set of classes in a database) selecting the classes by selecting a second number of classes having a farthest distance between mean features from among the first number of classes (mental process – selecting the classes by selecting a second number of classes having a farthest distance between mean features from among the first number of classes may be performed manually by a user with the aid of pen and paper by observing/analyzing mean features from among the first number of classes) Step 2A Prong 2 & Step 2B : There are no additional elements. Regarding Claim 3 : Step 2A Prong 1 : wherein the first similarity is determined based on a cosine similarity of a matrix for the extracted feature and a transposed matrix of a matrix for the mean feature (mathematical process – determining the first similarity based on a cosine similarity of a matrix for the extracted feature and a transposed matrix of a matrix for the mean feature may be performed by mathematical process, utilizing a cosine similarity of a matrix for the extracted feature and a transposed matrix of a matrix for the mean feature) Step 2A Prong 2 & Step 2B : There are no additional elements. Regarding Claim 4 : Step 2A Prong 1 : wherein the determining of the second similarity is based on a cosine similarity of a matrix for the mean feature and a transposed matrix of a matrix for the mean feature (mathematical process – determining the second similarity based on a cosine similarity of a matrix for the mean feature and a transposed matrix of a matrix for the mean feature may be performed by mathematical process, utilizing a cosine similarity of a matrix for the mean feature and a transposed matrix of a matrix for the mean feature) Step 2A Prong 2 & Step 2B : There are no additional elements. Regarding Claim 5 : Step 2A Prong 1 : See the rejection of Claim 1 above, which Claim 5 depends on. Step 2A Prong 2 & Step 2B : Additional Elements: wherein the parameter of the student model is updated based on a cosine similarity of a matrix for the first similarity and a matrix for the second similarity (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 – see MPEP 2106.05(f) – Examiner’s note: high level recitation of updating the student model based on a cosine similarity without significantly more. This cannot provide an inventive concept) Accordingly, under Step 2A Prong 2 and Step 2B, these additional elements do not integrate the abstract idea into practical application because they do not impose any meaningful limits on practicing the abstract idea, as discussed above in the rejection of claim 1. Regarding Claim 6 : Step 2A Prong 1 : See the rejection of Claim 1 above, which Claim 6 depends on. Step 2A Prong 2 & Step 2B : Additional Elements: wherein the parameter of the student model is updated such that a loss function based on a matrix for the first similarity and a matrix for the second similarity is minimized (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 – see MPEP 2106.05(f) – Examiner’s note: high level recitation of updating the student model based on a loss function without significantly more. This cannot provide an inventive concept) Accordingly, under Step 2A Prong 2 and Step 2B, these additional elements do not integrate the abstract idea into practical application because they do not impose any meaningful limits on practicing the abstract idea, as discussed above in the rejection of claim 1. Regarding Claim 7 : Step 2A Prong 1 : wherein the mean feature is determined based on the number of classes and a channel size of the mean feature set (mental process – determining the mean feature based on the number of classes and a channel size of the mean feature set may be performed manually by a user with the aid of pen and paper by observing/analyzing the number of classes and a channel size of the mean feature set) Step 2A Prong 2 & Step 2B : There are no additional elements. Regarding Claim 8 : Step 2A Prong 1 : wherein the extracted feature is determined based on a batch size of batches comprising the input data and a channel size of the mean feature set (mental process – determining the extracted feature based on a batch size of batches comprising the input data and a channel size of the mean feature set may be performed manually by a user with the aid of pen and paper by observing/analyzing the input data and a channel size of the mean feature set) Step 2A Prong 2 & Step 2B : There are no additional elements. Regarding Claim 9 : Claim 9 recites substantially the same limitations as Claim 1 , in the form of a machine which performs the method of Claim 1 . Step 2A Prong 1 : See the rejection of Claim 1 above. Step 2A Prong 2 & Step 2B : Additional Elements: A non-transitory computer-readable storage medium storing instructions that, when executed by a processor, cause the processor to perform the method of claim 1 (recited at a high-level of generality (i.e., as generic a non-transitory computer-readable storage medium and a processor) such that it amounts to no more than mere instructions to apply the exception using generic computer components) For the reasons above, Claim 9 is rejected as being directed to an abstract idea without significantly more. Regarding Claim 10: Step 1 : Claim 10 is an apparatus. Therefore, Claims 10-15 are directed to either a process, machine, manufacture, or composition of matter. Step 2A Prong 1 : select classes to be used for training from a database of classes (mental process – selecting classes to be used for training from a database of classes may be performed manually by a user with the aid of pen and paper by observing/analyzing a database of classes) generate a mean feature group comprising the mean features by extracting the mean features from the selected classes (mental process – generating a mean feature group comprising the mean features by extracting the mean features from the selected classes may be performed manually by a user with the aid of pen and paper by observing/analyzing the mean features extracted from the selected classes) determine a first similarity between the extracted feature and a mean feature corresponding to the input data among the mean features (mental process – determining a first similarity between the extracted feature and a mean feature corresponding to the input data among the mean features may be performed manually by a user with the aid of pen and paper by observing/analyzing the extracted feature and a mean feature corresponding to the input data among the mean features) determine a second similarity comprising a self-similarity of the mean feature (mental process – determining a second similarity comprising a self-similarity of the mean feature may be performed manually by a user with the aid of pen and paper by observing/analyzing the mean feature) Step 2A Prong 2 : This judicial exception is not integrated into a practical application. Additional Elements: one or more processors; and a memory storing that when executed by the one or more processors cause the one or more processors to (recited at a high-level of generality (i.e., as generic one or more processors and a memory) such that it amounts to no more than mere instructions to apply the exception using generic computer components) receive a batch comprising input data and extract a feature from the input data by the neural network model (Adding insignificant extra-solution activity to the judicial exception – see MPEP 2106.05(g)) wherein the neural network model is to be trained based on a mean feature set (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 – see MPEP 2106.05(f)) update a parameter of the student model based on the first similarity and the second similarity (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 – see MPEP 2106.05(f)) Step 2B : The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. Additional Elements: one or more processors; and a memory storing that when executed by the one or more processors cause the one or more processors to (mere instructions to apply the exception using generic computer components cannot provide an inventive concept) receive a batch comprising input data and extract a feature from the input data by the neural network model (MPEP 2106.05(d)(II) indicates that merely “Receiving or transmitting data over a network” is a well-understood, routine, conventional function when it is claimed in a merely generic manner (as it is in the present claim). Thereby, a conclusion that the claimed limitation is well-understood, routine, conventional activity is supported under Berkheimer) wherein the neural network model is to be trained based on a mean feature set (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 – see MPEP 2106.05(f) – Examiner’s note: high level recitation of applying a neural network model to previously determined data without significantly more. This cannot provide an inventive concept) update a parameter of the student model based on the first similarity and the second similarity (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 – see MPEP 2106.05(f) – Examiner’s note: high level recitation of updating a neural network model based on the determined data without significantly more. This cannot provide an inventive concept) For the reasons above, Claim 10 is rejected as being directed to an abstract idea without significantly more. This rejection applies equally to dependent claims 11-15. The additional limitations of the dependent claims are addressed below. Regarding Claim 11 : Step 2A Prong 1 : select a first number of classes predetermined in ascending order of a variance feature from among classes in the database (mental process – selecting a first number of classes predetermined in ascending order of a variance feature from among classes in the database may be performed manually by a user with the aid of pen and paper by observing/analyzing a set of classes in the database) select the classes by selecting a second number of classes having a farthest distance between mean features from among the first number of classes (mental process – selecting the classes by selecting a second number of classes having a farthest distance between mean features from among the first number of classes may be performed manually by a user with the aid of pen and paper by observing/analyzing mean features from among the first number of classes) Step 2A Prong 2 & Step 2B : one or more processors (recited at a high-level of generality (i.e., as generic one or more processors) such that it amounts to no more than mere instructions to apply the exception using generic computer components) Accordingly, under Step 2A Prong 2 and Step 2B, these additional elements do not integrate the abstract idea into practical application because they do not impose any meaningful limits on practicing the abstract idea, as discussed above in the rejection of claim 10. Regarding Claim 12 : Step 2A Prong 1 : determine the first similarity based on a cosine similarity of a matrix for the extracted feature and a transposed matrix of a matrix for the mean feature (mathematical process – determining the first similarity based on a cosine similarity of a matrix for the extracted feature and a transposed matrix of a matrix for the mean feature may be performed by mathematical process, utilizing a cosine similarity of a matrix for the extracted feature and a transposed matrix of a matrix for the mean feature) Step 2A Prong 2 & Step 2B : one or more processors (recited at a high-level of generality (i.e., as generic one or more processors) such that it amounts to no more than mere instructions to apply the exception using generic computer components) Accordingly, under Step 2A Prong 2 and Step 2B, these additional elements do not integrate the abstract idea into practical application because they do not impose any meaningful limits on practicing the abstract idea, as discussed above in the rejection of claim 10. Regarding Claim 13 : Step 2A Prong 1 : determine the second similarity based on a cosine similarity of a matrix for the mean feature and a transposed matrix of a matrix for the mean feature (mathematical process – determining the second similarity based on a cosine similarity of a matrix for the mean feature and a transposed matrix of a matrix for the mean feature may be performed by mathematical process, utilizing a cosine similarity of a matrix for the mean feature and a transposed matrix of a matrix for the mean feature) Step 2A Prong 2 & Step 2B : one or more processors (recited at a high-level of generality (i.e., as generic one or more processors) such that it amounts to no more than mere instructions to apply the exception using generic computer components) Accordingly, under Step 2A Prong 2 and Step 2B, these additional elements do not integrate the abstract idea into practical application because they do not impose any meaningful limits on practicing the abstract idea, as discussed above in the rejection of claim 10. Regarding Claim 14 : Step 2A Prong 1 : See the rejection of Claim 10 above, which Claim 14 depends on. Step 2A Prong 2 & Step 2B : Additional Elements: one or more processors (recited at a high-level of generality (i.e., as generic one or more processors) such that it amounts to no more than mere instructions to apply the exception using generic computer components) update the parameter of the student model based on a cosine similarity of a matrix for the first similarity and a matrix for the second similarity (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 – see MPEP 2106.05(f) – Examiner’s note: high level recitation of updating the student model based on a cosine similarity without significantly more. This cannot provide an inventive concept) Accordingly, under Step 2A Prong 2 and Step 2B, these additional elements do not integrate the abstract idea into practical application because they do not impose any meaningful limits on practicing the abstract idea, as discussed above in the rejection of claim 10. Regarding Claim 15 : Step 2A Prong 1 : See the rejection of Claim 10 above, which Claim 15 depends on. Step 2A Prong 2 & Step 2B : Additional Elements: one or more processors (recited at a high-level of generality (i.e., as generic one or more processors) such that it amounts to no more than mere instructions to apply the exception using generic computer components) update the parameter of the student model so that a loss function based on a matrix for the first similarity and a matrix for the second similarity is minimized (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 – see MPEP 2106.05(f) – Examiner’s note: high level recitation of updating the student model based on a loss function without significantly more. This cannot provide an inventive concept) Accordingly, under Step 2A Prong 2 and Step 2B, these additional elements do not integrate the abstract idea into practical application because they do not impose any meaningful limits on practicing the abstract idea, as discussed above in the rejection of claim 10. Claim Rejections - 35 USC § 103 07-20-aia AIA 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. 07-21-aia AIA Claim s 1, 6, 9, 10 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Ravichandran et al. (U.S. Patent No 10963754) (hereinafter Ravichandran), in view of Kim et al. (S. Kim, D. Min, B. Ham, S. Lin and K. Sohn, "FCSS: Fully Convolutional Self-Similarity for Dense Semantic Correspondence," in IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 41, no. 3, pp. 581-595, 1 March 2019) (hereinafter Kim) . Regarding Claim 1 : Ravichandran teaches: “A method of training a neural network model, the method comprising:” (preamble) “selecting classes from a database comprising a set of classes” (Ravichandran, Col. 2, Line 6, “using a fixed set of classes” and Col. 17 Lines 58-59, “data store can include several separate data tables, databases”; Examiner’s note: using a fixed set of classes implies selecting classes and using them afterwards. ) “generating a mean feature group comprising mean features from the selected classes” (Ravichandran, Col. 2, Lines 13-17, “embedding in which points cluster around a single prototype representation for each class. This prototype is a mean of embedded support samples for each class”) “receiving a batch comprising input data and extracting, by the neural network model, a feature from the input data” (Ravichandran, 401 and 403 in FIG. 4; Col 3, Lines 1-5, “embedding network (any network architecture such as CNNs, DNNs, RNNs, etc.) takes the samples and generate a feature vector per sample”; Examiner’s note: embedding network is a neural network; taking samples teaches receiving a batch comprising input data; generating a feature vector per sample teaches extracting a feature from the input data. ) “wherein the neural network model is to be trained according to a mean feature set” (Ravichandran, 403 in Fig. 4, “generating a set of vectors”; Col. 5, Lines 58-59, “train the embedding (adjust weights) at 411”; Examiner’s note: a set of vectors in 403 in FIG. 4 teaches the mean feature set. ) “determining a first similarity between the extracted feature and a mean feature corresponding to the input data” (Ravichandran, Col 5. Lines 41-44, “distances calculated per class, from the center variable for the class to the query samples”; Examiner’s note: determining a similarity encompasses distance calculation 1 . ) “updating a parameter of the neural network model based on the first similarity” (Ravichandran, 411 in Fig. 4, “backpropagate to the embedding”; Col. 5, Lines 58-59, “train the embedding (adjust weights) at 411”; Examiner’s note: backpropagating and training the embedding teach updating a parameter of the neural network. ) Ravichandran fails to teach determining a second similarity comprising a self-similarity of the mean feature and updating a parameter of the neural network model based on the second similarity. Kim teaches “Fully Convolutional Self-Similarity for Dense Semantic Correspondence (title)” comprising: “determining a second similarity comprising a self-similarity of a certain feature” (Kim, Section 5, “We presented the FCSS descriptor, which formulates local self-similarity within a fully convolutional network.”; Examiner’s note: the mean feature is taught by Ravichandran – see a limitation about generating a mean feature group above in claim 1. ) “updating a parameter of the neural network model based on the second similarity” (Kim, Section 5, “[…] the sampling patterns and the self-similarity measure were jointly learned within the proposed network […] The network was additionally trained in a weakly-supervised manner […]”; Examiner’s note: the second similarity comprising a self-similarity is learned and trained. Training teaches updating a parameter of the neural network model. ) It would have been obvious to one having ordinary skill in the art before the effective filing date of the invention was made to modify the invention in Ravichandran by applying the self-similarity features as taught in Kim to the mean features in Ravichandran in order to “allow[] for end-to-end training of the proposed network” (Kim, Page 582, Col. 1, Lines 12-13). Regarding Claim 6 : The combination of Ravichandran and Kim teaches: “wherein the parameter of the student model is updated such that a loss function based on a matrix for the first similarity and a matrix for the second similarity is minimized” (Ravichandran, 411 in Fig. 4, “calculate a loss function and backpropagate to the embedding, the loss function over a fixed geometry”; Examiner’s note: backpropagating teaches updating the parameter of the student model to minimize error; the collection of embedding teaches a matrix; the first similarity is taught by Ravichandran – see supra claim 1; the second similarity is taught by Kim – see supra claim 1. ) The reasons of obviousness have been noted in the rejection of Claim 1 above and applicable herein. Claim 9 is a system to perform the method of Claim 1 , therefore it is rejected under the same rationale. Claim 10 recites substantially the same limitations as Claim 1 , in the form of a system, therefore it is rejected under the same rationale. Claim 15 recites substantially the same limitations as Claim 6 , in the form of a system, therefore it is rejected under the same rationale . 07-21-aia AIA Claim s 2 and 11 are rejected under 35 U.S.C. 103 as being unpatentable over Ravichandran in view of Kim as applied in claim 1, in view of DexLab (“Rudiments of Hierarchical Clustering: Ward’s Method and Divisive Clustering”, published on 7/10/2018, available at <https://www.dexlabanalytics.com/blog/rudiments-of-hierarchical-clustering-wards-method-and-divisive-clustering>), and further in view of Sandhya et al. (Sandhya, N., et al. "Farthest Neighbor Approach for Finding Initial Centroids in K-Means." International Journal of Data Engineering (IJDE) 5.1 (2014): 1-13.) (hereinafter Sandhya) Regarding Claim 2 : Ravichandran teaches: “The method of claim 1, wherein the selecting of the classes comprises:” (preamble) Ravichandran fails to teach selecting a first number of classes in ascending order of a variance feature from among classes in the database and selecting the classes by selecting a second number of classes having a farthest distance between mean features from among the first number of classes. DexLab teaches “Rudiments of Hierarchical Clustering: Ward’s Method and Divisive Clustering (title)” comprising: “selecting a first number of classes in ascending order of a variance feature from among classes in the database” (DexLab, Paragraphs 2-3, “Hierarchical clustering, one the most common methods of clustering, builds a hierarchy of clusters either by a ‘’bottom up’’ approach (Agglomerative clustering) […] a special type of agglomerative hierarchical clustering technique that was introduced by Ward in 1963 […] and is used to generate clusters that have minimum within-cluster variance”; Examiner’s note: agglomerative hierarchical clustering teaches selecting a first number of classes in ascending order of a variance feature from among classes. ) It would have been obvious to one having ordinary skill in the art before the effective filing date of the invention was made to modify the teachings in Ravichandran and Kim by applying the hierarchical clustering techniques as taught in DexLab in order to “minimize[] the distance between the observations and the centers of the clusters,” resulting in a dendrogram that is often easier to interpret than dendrograms from other linkage methods (DexLab, Section “Ward’s method”). Sandhya teaches “Farthest Neighbor Approach for Finding Initial Centroids in K-Means (title)” comprising: “selecting the classes by selecting a second number of classes having a farthest distance between mean features from among the first number of classes” (Sandhya, Section 9, “accuracy and efficiency of the k-means algorithm is improved when the initial centroids are chosen using farthest neighbors than random selection of initial centroids”; Examiner’s note: choosing centroids using farthest neighbors teaches selecting the classes by selecting a second number of classes having a farthest distance between mean features. ) It would have been obvious to one having ordinary skill in the art before the effective filing date of the invention was made to modify the teachings in Ravichandran, Kim and DexLab by applying the farthest neighbor approach as taught in Sandhya in order to improve accuracy and efficiency of the k-means algorithm (Sandhya, Section 9, Lines 2-3). Claim 11 recites substantially the same limitations as Claim 2 , in the form of a system, therefore it is rejected under the same rationale . 07-21-aia AIA Claim s 3-5 and 12-14 are rejected under 35 U.S.C. 103 as being unpatentable over Ravichandran in view of Kim as applied in claim 1, and further in view of Li et al. (Li Y, Zhong N, Taniar D, Zhang H. "MCGNet+: an improved motor imagery classification based on cosine similarity." Brain Inform. 2022 Feb 1) (hereinafter Li) . Regarding Claim 3 : Ravichandran teaches : determining the first similarity based on the mean feature and extracted feature (see supra claim 1). Ravichandran fails to teach that the first similarity is determined based on a cosine similarity of a matrix for the extracted feature and a transposed matrix of a matrix for the mean feature. Li teaches “MCGNet+: an improved motor imagery classification based on cosine similarity (title) comprising: wherein the first similarity is determined based on a cosine similarity of a matrix for a feature and a transposed matrix of a matrix for another feature (Li, Section 1, “we use mutual information to generate the initial adjacency matrix and use cosine similarity to update the adjacency matrix dynamically, and achieve better performance”; Examiner’s note: using cosine similarity of a matrix and a transposed matrix achieves better performance; an adjacency matrix can be equal to its transposed matrix; the extracted feature and the mean feature are taught by Ravichandran - see supra claim 1. ) It would have been obvious to one having ordinary skill in the art before the effective filing date of the invention was made to modify the teachings in Ravichandran and Kim by applying the cosine similarity of a matrix for a feature and a transposed matrix of a matrix of another feature as taught in Li in order to “achieve better performance” (Li, Section 1 “Introduction”). Regarding Claim 4 : The combination of Ravichandran, Kim and Li teaches: “wherein the determining of the second similarity is based on a cosine similarity of a matrix for the mean feature and a transposed matrix of a matrix for the mean feature” ( Examiner’s note: the cosine similarity of a matrix and a transposed matrix is taught by Li – see supra claim 3; the mean feature is taught by Ravichandran - see supra claim 1. ) The reasons of obviousness have been noted in the rejection of Claim 3 above and applicable herein. Regarding Claim 5 : The combination of Ravichandran, Kim and Li teach: “wherein the parameter of the student model is updated based on a cosine similarity of a matrix for the first similarity and a matrix for the second similarity” ( Examiner’s note: using cosine similarity of a matrix is taught by Li – see supra claim 3; updating the parameter of the student model and the first similarity are taught by Ravichandran – see supra claim 1; the second similarity is taught by Kim – see supra claim 1. ) The reasons of obviousness have been noted in the rejection of Claim 3 above and applicable herein. Claim 12 recites substantially the same limitations as Claim 3 , in the form of a system, therefore it is rejected under the same rationale. Claim 13 recites substantially the same limitations as Claim 4 , in the form of a system, therefore it is rejected under the same rationale. Claim 14 recites substantially the same limitations as Claim 5 , in the form of a system, therefore it is rejected under the same rationale . 07-21-aia AIA Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Ravichandran in view of Kim as applied in claim 1, and further in view of Li et al. (Li, Cheng, and Bingyu Wang. "Fisher linear discriminant analysis." CCIS Northeastern University 6 (2014)) (hereinafter Li2) . Regarding Claim 7 : Ravichandran teaches : the mean feature set (see supra claim 1). Ravichandran fails to teach that the mean feature is determined based on the number of classes and a channel size of the mean feature set. Li teaches “Fisher linear discriminant analysis (title)” comprising: wherein the mean feature is determined based on the number of classes and a channel size of a feature set (Li, Section 2.2, Lines 2-3, “there are totally C classes […] we now will seek (C − 1) projections [y 1 , y 2 , ...y C−1 ] by means of (C −1) […] we define the mean vector […] as: PNG media_image1.png 82 366 media_image1.png Greyscale “; Examiner’s note: C classes represent the number of classes and C – 1 teaches the channel size of the mean feature set; the mean vector is determined based on the number of classes and the channel size of the mean feature set. ) It would have been obvious to one having ordinary skill in the art before the effective filing date of the invention was made to modify the teachings in Ravichandran and Kim by applying linear discriminant analysis as taught in Li2 in order to “represent a massive reduction in the dimensionality of the problem” (Li2, Section 2.2 “MultiClasses Problem”) . 07-21-aia AIA Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Ravichandran in view of Kim as applied in claim 1, and further in view of Ioffe et al. (Ioffe, S. & Szegedy, C. (2015). “Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift.” Proceedings of the 32nd International Conference on Machine Learning , in Proceedings of Machine Learning Research 37:448-456 Available from https://proceedings.mlr.press/v37/ioffe15.html.) (hereinafter Ioffe) . Regarding Claim 8 : Ravichandran teaches : the mean feature set (see supra claim 1). Ravichandran fails to teach that the extracted feature is determined based on a batch size of batches comprising the input data and a channel size of the mean feature set. Ioffe teaches “Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift (title)” comprising: wherein the extracted feature is determined based on a batch size of batches comprising the input data and a channel size of a feature set (Ioffe, Section 3, Algorithm 1, “Batch Normalizing Transform, applied to activation x over a mini-batch.” PNG media_image2.png 213 316 media_image2.png Greyscale ; Examiner’s note: batch normalization teaches to determine extracted features by normalizing them based on the mean and variance of the current mini-batch (batch size) for each feature set”) It would have been obvious to one having ordinary skill in the art before the effective filing date of the invention was made to modify the teachings in Ravichandran and Kim by applying batch normalization as taught in Ioffe so that “computation over a mini-batch can be more efficient than m computations for individual examples on modern computing platforms.” (Ioffe, Section 1 “Introduction”) . Conclusion 07-96 AIA The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Tung et al. teaches similarity-preserving knowledge distillation. Wen et al. teaches maintaining class mean features, minimizing intra-class variance and structuring feature space. Khosla et al. teaches computing similarity across multiple representations including pairwise similarity relationships . Any inquiry concerning this communication or earlier communications from the examiner should be directed to YONG D RHO whose telephone number is (571)270-0194. The examiner can normally be reached 8am-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, Viker Lamardo can be reached at 571-270-5871. 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. /YONG DOO RHO/Examiner, Art Unit 2147 /VIKER A LAMARDO/Supervisory Patent Examiner, Art Unit 2147 Application/Control Number: 18/338,732 Page 2 Art Unit: 2147 Application/Control Number: 18/338,732 Page 3 Art Unit: 2147 Application/Control Number: 18/338,732 Page 4 Art Unit: 2147 Application/Control Number: 18/338,732 Page 5 Art Unit: 2147 Application/Control Number: 18/338,732 Page 6 Art Unit: 2147 Application/Control Number: 18/338,732 Page 7 Art Unit: 2147 Application/Control Number: 18/338,732 Page 8 Art Unit: 2147 Application/Control Number: 18/338,732 Page 9 Art Unit: 2147 Application/Control Number: 18/338,732 Page 10 Art Unit: 2147 Application/Control Number: 18/338,732 Page 11 Art Unit: 2147 Application/Control Number: 18/338,732 Page 12 Art Unit: 2147 Application/Control Number: 18/338,732 Page 13 Art Unit: 2147 Application/Control Number: 18/338,732 Page 14 Art Unit: 2147 Application/Control Number: 18/338,732 Page 15 Art Unit: 2147 Application/Control Number: 18/338,732 Page 16 Art Unit: 2147 Application/Control Number: 18/338,732 Page 17 Art Unit: 2147 Application/Control Number: 18/338,732 Page 18 Art Unit: 2147 Application/Control Number: 18/338,732 Page 19 Art Unit: 2147 Application/Control Number: 18/338,732 Page 20 Art Unit: 2147 Application/Control Number: 18/338,732 Page 21 Art Unit: 2147 Application/Control Number: 18/338,732 Page 22 Art Unit: 2147 Application/Control Number: 18/338,732 Page 23 Art Unit: 2147 Application/Control Number: 18/338,732 Page 24 Art Unit: 2147 Application/Control Number: 18/338,732 Page 25 Art Unit: 2147 Application/Control Number: 18/338,732 Page 26 Art Unit: 2147 Application/Control Number: 18/338,732 Page 27 Art Unit: 2147 1 Similarity measure can “depend on the nature of the data and the specific application at hand. There are three commonly used similarity measures” including Euclidean distance. <What are Methods of Similarity Search for AI?>
Prosecution Insights