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. DETAILED ACTION This action is in response to the amendments filed 25 July 2023 . Claims 1-8 are pending and have been examined. Information Disclosure Statement The information disclosure statement (IDS) submitted on 25 July 2023 is being considered by the examiner. Claim Rejections - 35 USC § 112 (b) 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 appl icant regards as his invention. Claims 1, 2 , 4 , and 5 , and dependent Claims 3 and 6, 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. Claim 1 recites the limitation "the selected input data" in "on the basis of the selected input data " (emphasis added). There is insufficient antecedent basis for this limitation in the claim. For the purposes of examination, the limitation of Claim 1 has been interpreted to read: "extracting a feature quantity on the basis of the selected input data the selection of input data selected by the feature selection layer " (emphasis added). Similarly, f or the purposes of examination, the later limitation of Claim 1 has been interpreted to read: " reconstructing the selected input data input data selected by the feature selection layer " (emphasis added). Claim 2 recites the limitation "each sample" in " a domain of each sample " (emphasis added). There is insufficient antecedent basis for this limitation in the claim. For the purposes of examination, the limitation of Claim 2 has been interpreted to read: "input data comprising samples including information about a domain of each sample of the input data ." Claim 4 recites the limitation "between the domains" in " a degree of similarity between the domains " (emphasis added). There is insufficient antecedent basis for this limitation in the claim. For the purposes of examination, the limitation of Claim 4 has been interpreted to read: " the input data includ e information about a domain of each sample of the input data [as interpreted from Claim 2 ] ... wherein the neural network further includes an interdomain distance calculation layer for calculating a degree of similarity between the domains of each sample of the input data [as interpreted from Claim 4 ]." Claim 5 recites the limitation "between the domain" in "c alculating a degree of similarity between the domain " (emphasis added). There is insufficient antecedent basis for this limitation in the claim. For the purposes of examination, the limitation of Claim 5 has been interpreted to read: "the input data includ e information about a domain of each sample of the input data [as interpreted from Claim 2] ... wherein the neural network further includes a domain identification layer for identifying the domain of a sample of the input data and an interdomain distance calculation layer for calculating a degree of similarity between the domain s identified in the interdomain identification layer . " 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 . Claim 1 of the instant application is provisionally rejected on the ground of non - statutory double patenting as being unpatentable over Claim 5 of co-pending Application No. 18/227,261 (reference application). Although the claims at issue are not identical, they are not patentably distinct from each other because Claim 1 of the instant application is fully anticipated by Claim 5 in the co-pending application This is a provisional non-statutory double patenting rejection because the patentably indistinct claims have not in fact been patented. The following listing of independent Claim 1 (instant) and dependent Claim 5 (reference) indicates corresponding limitations forming the basis of the provisional rejection: 18/226,059 (instant) 18/227,261 (reference) 1 A method of learning a neural network, wherein the neural network includes: 1 A method of learning a neural network, wherein the neural network includes: 1 a feature selection layer for selecting a part of input data; 1 a feature selection layer for selecting a part of input data including information about a domain of each sample; 1 a feature extraction layer for extracting a feature quantity on the basis of the selected input data; 1 a feature extraction layer for extracting a feature quantity on the basis of the selected input data; and 1 a prediction layer for performing a prediction on the basis of the feature quantity; and 1 a prediction layer for performing a prediction on the basis of the feature quantity, and 1 a partial reconstruction layer for reconstructing the selected input data on the basis of the feature quantity, and 5 the neural network further includes a partial reconstruction layer for reconstructing the selected input data on the basis of the feature quantity 1 the method comprises adjusting a weight parameter of the neural network on the basis of a prediction accuracy by the prediction layer 1 the method comprises adjusting a weight parameter of the neural network to increase a prediction accuracy by the prediction layer and layer and to reduce a contribution to a prediction result of the prediction layer by the domain of the input data. 1 and a reconstruction error in the partial reconstruction layer. 5 the weight parameter of the neural network is adjusted on the basis of a reconstruction error in the partial reconstruction layer. 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 1-8 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 recites a method of learning , and thus the claimed process falls within a statutory category of invention. Step 2A Prong 1 The claim recites selecting a part of input data , which is a mental process. The claim recites extracting a feature quantity on the basis of the selected input data , which is a mental process. The claim recites performing a prediction on the basis of the feature quantity , which is a mental process. The claim recites reconstructing the selected input data on the basis of the feature quantity , which is a mental process. The claim recites adjusting a weight parameter ... on the basis of a prediction accuracy by the prediction layer and a reconstruction error in the partial reconstruction layer , which is a mental process. Thus, the claim recites an abstract idea. Step 2A Prong 2 , Step 2B The additional element a neural network does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h), "limit the use of the abstract idea to a particular technological environment"). The additional elements the neural network includes: a feature selection layer ... , a feature extraction layer ... , a prediction layer ..., and a partial reconstruction layer invoke a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible. Regarding Claim 2 Step 1 Regarding Claim 2, the rejection of Claim 1 is incorporated. Step 2A Prong 1 The claim recites selecting a part of input data (as recited by Claim 1) , wherein the input data include information about a domain of each sample , which is a mental process. The claim recites the weight parameter of the neural network is adjusted to reduce a contribution to a prediction result of the prediction layer by the information about the domain , which is a mental process. Thus, the claim recites an abstract idea. Step 2A Prong 2 , Step 2B The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible. Regarding Claim 3 Step 1 Regarding Claim 3 , the rejection of Claim 2 is incorporated. Step 2A Prong 1 The claim recites identifying the domain , which is a mental process. The claim recites a weight parameter of the domain identification layer is adjusted to increase an identification accuracy in the domain identification layer , which is a mental process. The claim recites weight parameters of the feature selection layer and the feature extraction layer are adjusted to reduce the identification accuracy in the domain identification layer , which is a mental process. Thus, the claim recites an abstract idea. Step 2A Prong 2 , Step 2B The additional element the neural network further includes a domain identification layer invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible. Regarding Claim 4 Step 1 Regarding Claim 4, the rejection of Claim 2 is incorporated. Step 2A Prong 1 The claim recites calculating a degree of similarity between the domains , which is a mental process. The claim recites weight parameters of the feature selection layer and the feature extraction layer are adjusted to increase the degree of similarity between the domains calculated in the interdomain distance calculation layer , which is a mental process. Thus, the claim recites an abstract idea. Step 2A Prong 2 , Step 2B The additional element the neural network further includes an interdomain distance calculation layer invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible. Regarding Claim 5 Step 1 Regarding Claim 5, the rejection of Claim 2 is incorporated. Step 2A Prong 1 The claim recites identifying the domain , which is a mental process. The claim recites calculating a degree of similarity between the domain , which is a mental process. The claim recites a weight parameter of the domain identification layer is adjusted to increase an identification accuracy in the domain identification layer , which is a mental process. The claim recites weight parameters of the feature selection layer and the feature extraction layer are adjusted to reduce the identification accuracy in the domain identification layer , which is a mental process. The claim recites weight parameters of the feature selection layer and the feature extraction layer are adjusted ... to increase the degree of similarity between the domains calculated in the interdomain distance calculation layer , which is a mental process. Thus, the claim recites an abstract idea. Step 2A Prong 2 , Step 2B The additional element wherein the neural network further includes a domain identification layer ... and an interdomain distance calculation layer invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible. Regarding Claim 6 Step 1 Regarding Claim 6, the rejection of Claim 1 is incorporated. Step 2A Prong 1 The claim recites the weight parameter ... is adjusted to predict ... by using the data , which is a mental process. Thus, the claim recites an abstract idea. Step 2A Prong 2 The additional element the input data include data obtained from a device amounts to insignificant extra-solution activity (see MPEP 2106.05(g), "mere data gathering"). The additional element an attribute information about a failure that may occur in the device and a failure that has occurred in the device does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h), "limit the use of the abstract idea to a particular technological environment"). The additional element the neural network does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h), "limit the use of the abstract idea to a particular technological environment"). The additional element an unexperienced failure that has not occurred in the device does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h), "limit the use of the abstract idea to a particular technological environment"). Step 2B The additional element the input data include data obtained from a device is well-understood, routine, conventional activity (see MPEP 2106.05(d), "receiving or transmitting data over a network"). The additional element an attribute information about a failure that may occur in the device and a failure that has occurred in the device does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h), "limit the use of the abstract idea to a particular technological environment"). The additional element the neural network does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h), "limit the use of the abstract idea to a particular technological environment"). The additional element an unexperienced failure that has not occurred in the device does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h), "limit the use of the abstract idea to a particular technological environment"). The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible. Regarding Claim 7 Step 1 Claim 7 recites a feature selection apparatus , and thus the claimed machine falls within a statutory category of invention. Step 2A Prong 1 The claim recites to adjust a weight parameter of a neural network on the basis of a prediction accuracy ... and a reconstruction error , which is a mental process. The claim recites selects a part of input data , which is a mental process. The claim recites selecting a part of input data , which is a mental process. The claim recites extracting a feature quantity on the basis of the selected input data , which is a mental process. The claim recites performing a prediction on the basis of the feature quantity , which is a mental process. The claim recites reconstructing the selected input data on the basis of the feature quantity , which is a mental process. Thus, the claim recites an abstract idea. Step 2A Prong 2 , Step 2B The additional element performs learning invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element using the learned neural network invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element the neural network includes: a feature selection layer ... a feature extraction layer ... the prediction layer ... the partial reconstruction layer invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible. Regarding Claim 8 Step 1 Claim 8 recites a feature selection method , and thus the claimed process falls within a statutory category of invention. Step 2A Prong 1 The claim recites adjust a weight parameter ... on the basis of a prediction accuracy ... and a reconstruction error , which is a mental process. The claim recites selecting a part of the input data , which is a mental process. The claim recites extracting a feature quantity on the basis of the selected input data , which is a mental process. The claim recites performing a prediction on the basis of the feature quantity , which is a mental process. The claim recites reconstructing the selected input data on the basis of the feature quantity , which is a mental process. Thus, the claim recites an abstract idea. Step 2A Prong 2 , Step 2B The additional element performing learning invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element of a neural network does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h), "limit the use of the abstract idea to a particular technological environment"). The additional element using the learned neural network invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element the neural network includes: a feature selection layer ... a feature extraction layer ... the prediction layer ... the partial reconstruction layer invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible. Claim Rejections - 35 USC § 103 The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action: A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made. This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention. 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 . Claim s 1-5, 7, and 8 is r ejected under 35 U.S.C. 103 as being unpatentable over Abid, et al., "Concrete Autoencoders: Differentiable Feature Selection and Reconstruction" (hereinafter "Abid") in view of Sun, et al. (US 2023/0307908 A1, hereinafter "Sun") . Regarding Claim 1 , Abid teaches: A method of learning a neural network (Abid, p. 1, Abstract "We introduce the concrete autoencoder, an end-to-end differentiable method for global feature selection") , wherein the neural network includes: a feature selection layer for selecting a part of input data (Abid, p. 3, 3. Proposed Method "The concrete autoencoder is an adaption of the standard autoencoder ... for discrete feature selection. Instead of using series of fully-connected layers for the encoder, we propose a concrete selector layer with a user-specified number of nodes, k. This layer selects stochastic linear combinations of input features during training") ; a feature extraction layer for extracting a feature quantity on the basis of the selected input data (Abid, p. 4, Figure 2 , " Concrete autoencoder architecture and pseudocode ," depicting an encoder corresponding to the instant extraction layer , as in p. 5, 4. Experiments , Reconstruction error: "We extract the k selected features. We pass the resulting matrix X S through the reconstruction function f θ that we have trained. We measure the Frobenius norm between the original and reconstructed test matrices || f θ X S -X | | F , normalized by the number of features d ") ; ... ; a partial reconstruction layer for reconstructing the selected input data on the basis of the feature quantity (Abid, p. 4, Figure 2 : " Concrete autoencoder architecture and pseudocode . ... The architecture of the decoder remains the same during train and test time, namely that x = f θ u , where u is the vector consisting of each u i ," where Abid's decoder output layer corresponds to the instant partial reconstruction layer , as in p. 3, 3. Proposed Method : "The decoder of a concrete autoencoder, which serves as the reconstruction function .... In effect, then, the concrete autoencoder is a method for selecting a discrete set of features that are optimized for an arbitrarily-complex reconstruction function") , and the method comprises adjusting a weight parameter of the neural network on the basis of ... a reconstruction error in the partial reconstruction layer (Abid , p. 4, Figure 2: "Concrete autoencoder architecture and pseudocode. (a) The architecture of a concrete autoencoder consists of ... arbitrary decoding layers (e.g. a deep feedforward neural network).... The architecture of the decoder remains the same during train and test time, namely that x = f θ u , where u is the vector consisting of each u i ," where Abid's decoder parameters θ reasonably suggest the instant weight parameter ), as updated during training of p. 4, Algorithm I. Training a Concrete Auto e ncod e r ). Abid teaches a method of learning a neural network, wherein the neural network includes: a feature selection layer for selecting a part of input data; and a feature extraction layer for extracting a feature quantity on the basis of the selected input data ; a partial reconstruction layer for reconstructing the selected input data on the basis of the feature quantity, and adjusting a weight parameter of the neural network on the basis of a reconstruction error in the partial reconstruction layer . Abid does not explicitly teach a prediction layer for performing a prediction and adjusting a weight parameter of the neural network on the basis of a prediction accuracy by the prediction layer . However, Sun teaches: a prediction layer for performing a prediction (Sun, Fig. 8, layers of generator G responsible for predicting input sequences, and [0055] : " Adversarial training means that the G is trained to trick the D into believing that the sequences predicted by G are the actual measurements ") on the basis of the feature quantity (Sun, [0064]: "the feature extractor network is provided with 3 kHz data without down-sampling to ensure that the harmonic features are accurately extracted"); ... and adjusting a weight parameter of the neural network (Sun, [0060]: "All the Conv- EDNet model blocks are jointly optimized. The model is trained for a pre-determined number of epochs (E), and the weights of all the different components are updated simultaneously to minimize the loss function (5)") on the basis of a prediction accuracy by the prediction layer (Sun, [0066] : "For PV predictions, we observe that the Conv- EDNet and Conv- EDNet + models are accurately able to detect the onset of PV generation. This can be attributed to the novel feature extraction module, which has enhanced PV generation detection capability owing to its dependence on irradiance and STFT-based harmonic features") . It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Abid regarding a method of learning a neural network, wherein the neural network includes: a feature selection layer for selecting a part of input data; and a feature extraction layer for extracting a feature quantity on the basis of the selected input data; a partial reconstruction layer for reconstructing the selected input data on the basis of the feature quantity, and adjusting a weight parameter of the neural network on the basis of a reconstruction error in the partial reconstruction layer with those of Sun regarding a prediction layer for performing a prediction and adjusting a weight parameter of the neural network on the basis of a prediction accuracy by the prediction layer. The motivation to do so would be to facilitate training a model with improved disaggregation of model input data and improved model robustness (Sun, [0054]: "To further enhance the disaggregation quality and robustness of Cony- EDNet , we extend the proposed model with an adversarial learning component. ... A new component called the discriminator ('D ) is added to facilitate adversarial learning"). Regarding Claim 7 , Abid teaches: A feature selection apparatus (Abid, p. 4, 4. Experiments: "In this section, we carry out experiments to compare the performance of concrete autoencoders to other feature subset selections on standard public datasets" with Figs. 3-6 and Table 1, where a computing apparatus is inherent in Abid's experiments using datasets ) that performs precisely those steps recited by the method of Claim 1. Claim 7 is rejected under the same rationale as Claim 1. Regarding Claim 8 , Abid teaches: A feature selection method (Abid, p. 4, 4. Experiments: "In this section, we carry out experiments to compare the performance of concrete autoencoders to other feature subset selections on standard public datasets" ) that performs precisely those steps recited by the method of Claim 1. Claim 8 is rejected under the same rationale as Claim 1. Regarding Claim 2 , the rejection of Claim 1 is incorporated. Sun further teaches: wherein the input data include information about a domain of each sample (Sun, Fig. 9, Algorithm 2, ConvED -DAN training algorithm, line 4, "Source domain data," and line 11, "Target domain data," where Sun's source and target domain feature data corresponds to the instant input data domain ) , and the weight parameter of the neural network is adjusted to reduce a contribution to a prediction result of the prediction layer by the information about the domain (Sun, [0071] "To incorporate domain adaptation, we repurpose the discriminator network to perform the task of domain adaptation. The underlying idea is that the discriminator adversarially trains the model using the feature space instead of the disaggregated time-series used in Conv- EDNet +. The goal is to minimize the deviation between the source domain and the target domain feature space. ... The feature space MSE value aims at minimizing the gap between the source domain data 810 and the target domain data 820 feature space. Lastly, the discriminator 830 is updated using the binary cross-entropy loss, which computes the probability of the input feature space being obtained from the source or target domain data," where Sun's discriminator loss corresponds to the instant contribution to a prediction result) . It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the prior teachings of the Abid/Sun combination with the further teachings of Sun regarding the input data include information about a domain of each sample, and the weight parameter of the neural network is adjusted to reduce a contribution to a prediction result of the prediction layer by the information about the domain. The motivation to do so would be to facilitate training a model with improved disaggregation of model input data and improved model robustness (Sun, [0054]: "To further enhance the disaggregation quality and robustness of Cony- EDNet , we extend the proposed model with an adversarial learning component. ... A new component called the discriminator ('D ) is added to facilitate adversarial learning"). Regarding Claim 3 , the rejection of Claim 2 is incorporated. Sun further teaches: wherein the neural network further includes a domain identification layer for identifying the domain (Sun, [0071]: "In Conv- EDNet +, the network consists of the Conv- EDNet energy disaggregation model (a generator 800) combined with a discriminator network 805 of a discriminator 830 to perform adversarial training. ... To incorporate domain adaptation, we repurpose the discriminator network to perform the task of domain adaptation. ... The goal is to minimize the deviation between the source domain and the target domain feature space," where Sun's discriminator corresponds to the instant domain identification layer ) , a weight parameter of the domain identification layer is adjusted (Sun, Fig. 5A, Algorithm 1 Conv- EDNet + training algorithm , line 19, Θ d = Adam ∇ Θ L D , Θ d , where Θ d represents discriminator weight parameters, as in [0060] "All the Conv- EDNet model blocks are jointly optimized. The model is trained for a pre-determined number of epochs (E), and the weights of all the different components are updated simultaneously to minimize the loss function (5)") to increase an identification accuracy in the domain identification layer (Sun, [0055]: "The D is trained to maximize the probability of correctly classifying the real samples (i.e., the measurements) and the generated samples (produced by D ). In contrast, G is trained to produce output samples that are hard correctly distinguish by D ," where Sun's maximized probability of correctly classifying corresponds to the instant increase accuracy ) , and weight parameters of the feature selection layer and the feature extraction layer are adjusted (Sun, Fig. 5A, Algorithm 1 Conv- EDNet + training algorithm , lines 8 and 11, where Θ ae represents the encoder weight matrix, and [0060] "All the Conv-EDNet model blocks are jointly optimized. The model is trained for a pre-determined number of epochs (E), and the weights of all the different components are updated simultaneously to minimize the loss function (5)") to reduce the identification accuracy in the domain identification layer (Sun, [0055]: "The D is trained to maximize the probability of correctly classifying the real samples (i.e., the measurements) and the generated samples (produced by D ). In contrast, G is trained to produce output samples that are hard correctly distinguish by D ," where Sun's adversarial generator corresponds to the instant accuracy reduction ) . It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the prior teachings of the Abid /Sun combination with the further teachings of Sun regarding wherein the neural network further includes a domain identification layer for identifying the domain, a weight parameter of the domain identification layer is adjusted to increase an identification accuracy in the domain identification layer, and weight parameters of the feature selection layer and the feature extraction layer are adjusted to reduce the identification accuracy in the domain identification layer . The motivation to do so would be to facilitate training a model when faced with limited labeled data from the model's target domain (Sun, [0011]: "We combine adversarial learning and joint adaptation concepts and modify the Conv- EDNet + to perform the task of distribution gap minimization of the feature space and label space between the source domain and target domain. This approach enables training the model on synthetically obtained data and its application to real-world measurement data. By using limited labeled data in the source domain and unlabeled target domain data, the proposed network can generate satisfactory results for energy disaggregation using CPOW measurements"). Regarding Claim 4 , the rejection of Claim 2 is incorporated. Sun teaches: wherein the neural network further includes an interdomain distance calculation layer (Sun, [0071]: "To incorporate domain adaptation, we repurpose the discriminator network to perform the task of domain adaptation. The underlying idea is that the discriminator adversarially trains the model using the feature space instead of the disaggregated time-series used in Conv- EDNet +. The goal is to minimize the deviation between the source domain and the target domain feature space," where Sun's discriminator corresponds to the instant distance layer , and Sun's source/target feature space deviation corresponds to the instant interdomain distance ) for calculating a degree of similarity between the domains (Sun, [0071]: "the generator 800 is updated using a weighted function corresponding to the MSE loss value on source domain data and the MSE loss value on the feature space. The feature space MSE value aims at minimizing the gap between the source domain data 810 and the target domain data 820 feature space. Lastly, the discriminator 830 is updated using the binary cross-entropy loss, which computes the probability of the input feature space being obtained from the source or target domain data," where Sun's binary cross-entropy loss corresponds to the instant degree of similarity , which is used for training, as in Fig. 5A, Algorithm 1) , and weight parameters of the feature selection layer and the feature extraction layer are adjusted (Sun, Fig. 5A, Algorithm 1 Conv- EDNet + training algorithm, lines 8 and 11, where Θ ae represents the encoder weight matrix, and [0060] "All the Conv-EDNet model blocks are jointly optimized. The model is trained for a pre-determined number of epochs (E), and the weights of all the different components are updated simultaneously to minimize the loss function (5)") to increase the degree of similarity between the domains calculated in the interdomain distance calculation layer (Sun, [0055]: "The D is trained to maximize the probability of correctly classifying the real samples (i.e., the measurements) and the generated samples (produced by D ). In contrast, G is trained to produce output samples that are hard correctly distinguish by D ," where Sun's adversarial ly hard to distinguish corresponds to the instant degree of similarity ) . It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the prior teachings of the Abid/Sun combination with the further teachings of Sun regarding the neural network further includes an interdomain distance calculation layer for calculating a degree of similarity between the domains, and weight parameters of the feature selection layer and the feature extraction layer are adjusted to increase the degree of similarity between the domains calculated in the interdomain distance calculation layer. The motivation to do so would be to facilitate training a model when faced with limited labeled data from the model's target domain (Sun, [0011]: "We combine adversarial learning and joint adaptation concepts and modify the Conv- EDNet + to perform the task of distribution gap minimization of the feature space and label space between the source domain and target domain. This approach enables training the model on synthetically obtained data and its application to real-world measurement data. By using limited labeled data in the source domain and unlabeled target domain data, the proposed network can generate satisfactory results for energy disaggregation using CPOW measurements"). Regarding Claim 5 , the rejection of Claim 2 is incorporated. Sun further teaches: the neural network further includes a domain identification layer for identifying the domain and an interdomain distance calculation layer (Sun, [0071]: "To incorporate domain adaptation, we repurpose the discriminator network to perform the task of domain adaptation. The underlying idea is that the discriminator adversarially trains the model using the feature space instead of the disaggregated time-series used in Conv- EDNet +. The goal is to minimize the deviation between the source domain and the target domain feature space," where Sun's discriminator corresponds to the instant domain identification layer and to the distance layer ) for calculating a degree of similarity between the domain (Sun, [0071]: "the generator 800 is updated using a weighted function corresponding to the MSE loss value on source domain data and the MSE loss value on the feature space. The feature space MSE value aims at minimizing the gap between the source domain data 810 and the target domain data 820 feature space. Lastly, the discriminator 830 is updated using the binary cross-entropy loss, which computes the probability of the input feature space being obtained from the source or target domain data," where Sun's binary cross-entropy loss corresponds to the instant degree of similarity , which is used for training, as in Fig. 5A, Algorithm 1) , a weight parameter of the domain identification layer is adjusted (Sun, Fig. 5A, Algorithm 1 Conv- EDNet + training algorithm , line 19, Θ d = Adam ∇ Θ L D , Θ d , where Θ d represents discriminator weight parameters, as in [0060] "All the Conv-EDNet model blocks are jointly optimized. The model is trained for a pre-determined number of epochs (E), and the weights of all the different components are updated simultaneously to minimize the loss function (5)") to increase an identification accuracy in the domain identification layer (Sun, [0055]: "The D is trained to maximize the probability of correctly classifying the real samples (i.e., the measurements) and the generated samples (produced by D ). In contrast, G is trained to produce output samples that are hard correctly distinguish by D ," where Sun's maximized probability of correctly classifying corresponds to the instant increase accuracy ) , and weight parameters of the feature selection layer and the feature extraction layer are adjusted (Sun, Fig. 5A, Algorithm 1 Conv- EDNet + training algorithm , and [0060] "All the Conv- EDNet model blocks are jointly optimized. The model is trained for a pre-determined number of epochs (E), and the weights of all the different components are updated simultaneously to minimize the loss function (5)") to reduce the identification accuracy in the domain identification layer, and to increase the degree of similarity between the domains calculated in the interdomain distance calculation layer (Sun, [0055]: "The D is trained to maximize the probability of correctly classifying the real samples (i.e., the measurements) and the generated samples (produced by D ). In contrast, G is trained to produce output samples that are hard correctly distinguish by D ," where Sun's adversarially hard to distinguish corresponds to the instant degree of similarity ) . It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the prior teachings of the Abid/Sun combination with the further teachings of Sun regarding the neural network further includes a domain identification layer for identifying the domain and an interdomain distance calculation layer for calculating a degree of similarity between the domain, a weight parameter of the domain identification layer 1s adjusted to in crease an identification accuracy in the domain identification layer, and weight parameters of the feature selection layer and the feature extraction layer are adjusted to reduce the identification accuracy in the domain identification layer, and to increase the degree of similarity between the domains calculated in the interdomain distance calculation layer . The motivation to do so would be to facilitate training a model when faced with limited labeled data from the model's target domain (Sun, [0011]: "We combine adversarial learning and joint adaptation concepts and modify the Conv- EDNet + to perform the task of distribution gap minimization of the feature space and label space between the source domain and target domain. This approach enables training the model on synthetically obtained data and its application to real-world measurement data. By using limited labeled data in the source domain and unlabeled target domain data, the proposed network can generate satisfactory results for energy disaggregation using CPOW measurements"). Claim 6 is r ejected under 35 U.S.C. 103 as being unpatentable over Abid, et al., "Concrete Autoencoders: Differentiable Feature Selection and Reconstruction" (hereinafter "Abid") in view of Sun, et al. (US 2023/0307908 A1, hereinafter "Sun") in view of Mansouri, et al., "A Deep Explainable Model for Fault Prediction Using IoT Sensors" (hereinafter "Mansouri") . Regarding Claim 6 , the rejection of Claim 1 is incorporated. The Abid /Sun combination teaches learning a neural network that includes a feature selection layer, a feature extraction layer, and a prediction layer The Abid/ Sun combination does not explicitly teach wherein the input data include data obtained from a device and an attribute information about a failure that may occur in the device and a failure that has occurred in the device, the weight parameter of the neural network is adjusted to predict an unexperienced failure that has not occurred in the device, by using the data obtained from the device . However, Mansouri teaches: However, Mansouri teaches: wherein the input data include data obtained from a device ( Mansouri , p. 66933 , Abstract: " IoT sensors and deep learning models can widely be applied for fault prediction. ... This paper first examines different deep learning techniques to carry out univariate time series analysis based on vibration sensors installed on four industrial bearings to predict a fault occurring in a predefined time window") and an attribute information about a failure that may occur in the device and a failure that has occurred in the device ( Mansouri , p. 66935 , I. Introduction : "To this end the current work evaluates several recurrent deep learning models for fault prediction in disintegrated bearings using vibration sensor readings" and p. 66935 , II. Framework : "The first goal of this research is to predict a machine fault within the time window ahead. This binary classification task predicts whether a sequence of vibration readings leads to a fault or a normal condition," where Mansouri's predicted future fault and disintegrated bearings corresponds to the possible failure and failure that has occurred , respectively ) , the weight parameter of the neural network is adjusted (Mansouri, p. 66936 , II. Framework : "The selector model (2), parametrized by θ , accepts an input and generates m independent probabilities with the same size as the input features" and p. 66937 , Figure 2: "The GSX workflow. During the training, data passes through GSX and the baseline model simultaneously to calculate the error," depicting the selector model as a component of GSX) to predict an unexperienced failure that has not occurred in the device, by using the data obtained from the device ( Mansouri , p. 66940 , IV. Conclusion : " This paper develops an explainable deep learning model for the predictive preventive maintenance task, aiming to predict a fault happening in the near future by processing the preceding vibration signals. The primary dataset consists of chronological sequences of vibration readings and their related labels denoting whether they were faulty or not ") . It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of the Abid /Sun combination regarding learning a neural network that includes a feature selection layer, a feature extraction layer, and a prediction layer with those of Mansouri regarding wherein the input data include data obtained from a device and an attribute information about a failure that may occur in the device and a failure that has occurred in the device, the weight parameter of the neural network is adjusted to predict an unexperienced failure that has not occurred in the device, by using the data obtained from the device . The motivation to do so would be to facilitate training a model with improved explainability over a longer sampling time window ( Mansouri, p. 66933 , Abstract : " hybrid models outperform other models when the time window increases. Then, instance-wise feature selection has been considered to highlight the most contributing features for its outputs regarding any input. In this problem, the main challenge is to propose a trainable feature selection model with the minimum number of selected features whilst its performance is close to the baseline model. This paper develops a novel explainable method called the Gumbel-Sigmoid eXplanator (GSX) to tackle these problems "). Conclusion The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Acharya, et al., "Feature Selection and Extraction for Graph Neural Networks , " teach a method for extending the use of feature selection and feature extraction matrices parameterized by a temperature value to perform classification using a graph convolutional network. Any inquiry concerning this communication or earlier communications from the examiner should be directed to FILLIN "Examiner name" \* MERGEFORMAT ROBERT N DAY whose telephone number is FILLIN "Phone number" \* MERGEFORMAT (703)756-1519 . The examiner can normally be reached FILLIN "Work Schedule?" \* MERGEFORMAT M-F 9-5 . Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an intervie