CTNF 17/211,681 CTNF 98837 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. Response to Arguments Applicant’s arguments filed 04/17/2026 on pages 8-11 regarding the rejection under 35 U.S.C. 103 with respect to claims 1-23 have been fully considered but are moot. New reference Wang has been incorporated below. Specification The formulas presented in the specification are not clearly legible. One example is formula (2) in paragraph [0044]. Please review all of the formulas. Claim Rejections - 35 USC § 102 07-07-aia AIA 07-07 The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action: A person shall be entitled to a patent unless – 07-08-aia AIA (a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention. 07-15 AIA Claim 1 is being rejected under 35 U.S.C. 102( a)(1 ) as being anticipated by Wang et al. (Learning Priors for Adversarial Autoencoders); hereinafter Wang Regarding claim 1, Wang teaches a computer-implemented method for creating a generative model, the method comprising: (Wang [page 1, 1. Introduction]: “These models involve specifying a prior distribution over latent variables and defining a deep generative network (i.e., the decoder) that maps latent variables to data space in stochastic or deterministic fashion.) performing one or more operations based on a plurality of training images to generate a trained encoder network and (Wang [page 5]: “Table 1 compares their Inception Score for image generation on CIFAR-10 with a latent code size of 64.”; Note: See Algorithm 1 on page 5 to see that the encoder is generated at “// Update network parameters”) a trained prior network, (Wang [page 1, 1. Introduction]: “We replace the simple prior with a learned prior by training the code generator to output latent variables that will minimize an adversarial loss in data space.”) wherein the trained encoder network converts each image included in the plurality of training images into a set of visual attributes (Wang [page 2]: “They thus replace the KL-divergence with an adversarial loss imposed on the encoder output, requiring that the latent code z produced by the encoder should have an aggregated posterior distribution the same as the prior p(z).”; Note: The specification of the instant application addresses on paragraph [0005] that “The trained VAE would include an encoder network that converts each image into hundreds or thousands of numeric latent variable values. Each latent variable would represent a corresponding visual attribute found in one or more of the images used to train the VAE”;) and the trained prior network learns a distribution of the set of visual attributes across the plurality of training images; (Wang [page 1, 1. Introduction]: “We replace the simple prior with a learned prior by training the code generator to output latent variables that will minimize an adversarial loss in data space.”; and [page 3, III. Learning the Prior]: “to find a prior that, together with the decoder in Fig. 3, would lead to a distribution that maximizes the data likelihood”) performing one or more operations to train one or more classifiers to distinguish between values for the set of visual attributes generated by the trained encoder network and values for the set of visual attributes selected from the distribution learned by the trained prior network; and (See Figure 3 of Wang to see that the discriminator on the bottom is the classifier that distinguishes the input from the gaussian distribution and the latent codes (visual attributes) from the encoder.) combining the trained prior network to produce a trained prior component that is included in the generative model, and the one or more classifiers (See Figure 3 of Wang to see that the code generator (trained prior network) and the discriminators (classifiers) are structurally combined in the same architecture. The entire AAE architecture is the trained prior component.) wherein, in operation, the trained prior component produces one or more values for the set of visual attributes in order to generate a new image that is not included in the plurality of training images. (Wang [page 5]: “Fig. 6 further visualizes sample images generated with these models by driving the decoder with latent codes drawn from the prior or the code generator in our case. It is observed that our model produces much sharper images than the others . This confirms that a learned and flexible prior is beneficial to the characterization and generation of data.”) 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 Claims 2 and 4 are reject ed under 35 U.S.C. 103 as being unpatentable over Wang et al. (Learning Priors for Adversarial Autoencoders); hereinafter Wang in view of Bauer et al. (Resampled Priors for Variational Autoencoders); hereinafter Bauer Claim 2 is rejected over Wang and Bauer with the incorporation of claim 1. Regarding claim 2, Wang does not appear to explicitly teach wherein combining the trained prior network and the one or more classifiers comprises combining one or more first values selected from the distribution learned by the trained prior network with a reweighting factor that is based on the one or more first values. However, Bauer teaches wherein combining the trained prior network and the one or more classifiers comprises combining one or more first values selected from the distribution learned by the trained prior network with a reweighting factor that is based on the one or more first values. (Bauer [page 4, 4 VAEs with resampled priors]: “we use the original VAE prior as a proposal distribution π(z) and define a resampled prior p(z) through a rejection sampler with learned acceptance probability a(z).”; [page 2, 3 Learned Accept/Reject Sampling]: “We then transform π(z) into a more expressive distribution p∞(z) by multiplying it by an acceptance function aλ(z) ∈ [0; 1]; and [page 3]: “Thus, we can sample from p ∞ as follows: draw a candidate sample z from the proposal π and accept it with probability a(z ); if accepted, return z as the sample ; otherwise repeat the process. This means we can think of a(z) as a filtering function which — by rejecting different fractions of samples at different locations — can transform π(z) into a distribution closer to q(z).”; The acceptance function is the reweighting factor.) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the acceptance function of Bauer to improve VAE performance (Bauer, Abstract). Wang and Bauer are analogous art because they both concern priors in variational autoencoders. Claim 4 is rejected over Wang and Bauer. Regarding claim 4, Wang teaches a computer-implemented method for creating a generative model, the method comprising: (Wang [page 1, 1. Introduction]: “These models involve specifying a prior distribution over latent variables and defining a deep generative network (i.e., the decoder) that maps latent variables to data space in stochastic or deterministic fashion.”) performing one or more operations based on a training dataset to generate a trained encoder network (Wang [page 5]: “Table 1 compares their Inception Score for image generation on CIFAR-10 with a latent code size of 64.”; Note: See Algorithm 1 on page 5 to see that the encoder is generated at “// Update network parameters”) and a trained prior network, (Wang [page 1, 1. Introduction]: “We replace the simple prior with a learned prior by training the code generator to output latent variables that will minimize an adversarial loss in data space.”) wherein the trained encoder network converts a plurality of data points included in the training dataset into a set of latent variables, (Wang [page 2]: “They thus replace the KL-divergence with an adversarial loss imposed on the encoder output, requiring that the latent code z produced by the encoder should have an aggregated posterior distribution the same as the prior p(z).”; Note: The specification of the instant application addresses on paragraph [0005] that “The trained VAE would include an encoder network that converts each image into hundreds or thousands of numeric latent variable values. Each latent variable would represent a corresponding visual attribute found in one or more of the images used to train the VAE”;) and the trained prior network learns a distribution of the set of latent variables across the training dataset; (Wang [page 1, 1. Introduction]: “We replace the simple prior with a learned prior by training the code generator to output latent variables that will minimize an adversarial loss in data space.”; and [page 3, III. Learning the Prior]: “to find a prior that, together with the decoder in Fig. 3, would lead to a distribution that maximizes the data likelihood”) performing one or more operations to train one or more classifiers to distinguish between values for the set of latent variables generated via the trained encoder network and values sampled from the distribution learned by the trained prior network; and (See Figure 3 of Wang to see that the discriminator on the bottom is the classifier that distinguishes the input from the gaussian distribution and the latent codes (visual attributes) from the encoder.) creating a trained prior component based on the trained prior network and one or more classifiers, (See Figure 3 of Wang to see that the code generator (trained prior network) and the discriminators (classifiers) are structurally combined in the same architecture. The entire AAE architecture is the trained prior component.) wherein, in operation, the trained prior component produces the one or more second values in order to generate a new data point that is not included in the training dataset. (Wang [page 5]: “Fig. 6 further visualizes sample images generated with these models by driving the decoder with latent codes drawn from the prior or the code generator in our case. It is observed that our model produces much sharper images than the others . This confirms that a learned and flexible prior is beneficial to the characterization and generation of data.”) Wang does not appear to explicitly teach wherein the trained prior component applies a reweighting factor to one or more first values sampled from the distribution learned by the trained prior network to generate one or more second values for the set of latent variables, However, Bauer teaches wherein the trained prior component (Bauer [page 4, 4 VAEs with resampled priors]: “we use the original VAE prior as a proposal distribution π(z) and define a resampled prior p(z) through a rejection sampler with learned acceptance probability a(z).”) applies a reweighting factor (Bauer [page 3]: “Thus, we can sample from p ∞ as follows: draw a candidate sample z from the proposal π and accept it with probability a(z); if accepted, return z as the sample ; otherwise repeat the process. This means we can think of a(z) as a filtering function which — by rejecting different fractions of samples at different locations — can transform π(z) into a distribution closer to q(z).”) to one or more first values sampled from the distribution learned by the trained prior network to generate one or more second values for the set of latent variables, (Bauer [page 3]: “Thus, we can sample from p ∞ as follows: draw a candidate sample z from the proposal π and accept it with probability a(z ); if accepted, return z as the sample ; otherwise repeat the process. This means we can think of a(z) as a filtering function which — by rejecting different fractions of samples at different locations — can transform π(z) into a distribution closer to q(z).”) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the reweighting factor of Bauer to improve VAE performance (Bauer, Abstract). Wang and Bauer are analogous art because they both concern priors in variational autoencoders . 07-21-aia AIA Claim 3 is rejected under 35 U.S.C. 103 as being unpatentable over Wang in view of Zhang et al. (US 20210358577 A1); hereinafter Zhang Claim 3 is rejected over Wang and Zhang with the incorporation of claim 1. Regarding claim 3, Wang does not appear to explicitly teach wherein the new image comprises at least one face. However, Zhang teaches wherein the new image comprises at least one face. (Zhang [0077]: “and the VAE has been trained to encode and decode human faces, then by inputting a random value of Z into the decoder 208 p it is possible to generate a new face that did not belong to any of the sampled subjects during training”) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the new face image generation of Zhang for efficient use of latent space (Zhang, [0093]). Wang and Zhang are analogous art because they both concern image generation using VAEs and latent space . 07-21-aia AIA Claim s 5 and 6 are rejected under 35 U.S.C. 103 as being unpatentable over Wang and Bauer in view of Zhang Claim 5 is rejected over Wang, Bauer and Zhang with the incorporation of claim 4. Regarding claim 5, Wang does not appear to explicitly teach wherein the distribution learned by the trained prior network comprises a hierarchy of latent variables, and wherein the one or more first values are sampled from the distribution learned by the trained prior network by: sampling one first value from a first group of latent variables included in the hierarchy of latent variables; and sampling another first value from a second group of latent variables included in the hierarchy of latent variables based on the first value and a feature map. However, Zhang teaches wherein the distribution learned by the trained prior network comprises a hierarchy of latent variables, and wherein the one or more first values are sampled from the distribution learned by the trained prior network by: sampling one first value from a first group of latent variables included in the hierarchy of latent variables; and (Zhang [0044]: “VAEM uses a hierarchy of latent variables which is fit in two stages . In the first stage, one type-specific VAE is learned for each dimension. These initial one-dimensional VAEs capture marginal distribution properties and provide a latent representation that is uniform across dimensions.”; Note: The first stage is the first group of latent variables) sampling another first value from a second group of latent variables included in the hierarchy of latent variables based on the first value and a feature map. (Zhang [0044]: “In the second stage, another VAE is used to capture dependencies among the one-dimensional latent representations from the first stage.”; Note: The second stage is the second group of latent variables and the use of the feature map at the second stage is referenced in paragraph [0115]) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the sampling of Zhang for efficient use of latent space (Zhang, [0093]). Wang and Zhang are analogous art because they both concern image generation using VAEs and latent space. Claim 6 is rejected over Wang, Bauer and Zhang with the incorporation of claim 4. Regarding claim 6, Wang teaches wherein the one or more classifiers comprise a first classifier that distinguishes between a third value sampled from the first group of latent variables using the trained prior network and a fourth value for the first group of latent variables generated by the trained encoder network, (Wang [page 4]: “Repeat (for each epochs E i ) Repeat (for each mini-batch x j )”; Note: See Algorithm 1 of Wang to see that each iteration will produce additional values. See Figure 3 of Wang to see that the discriminator on the bottom is the classifier that distinguishes the input from the gaussian distribution and the latent codes (visual attributes) from the encoder.)) and a second classifier that distinguishes between a fifth value sampled from the second group of latent variables using the trained prior network and a sixth value for the second group of latent variables generated by the trained encoder network. (See Figure 3 of Wang to see that the first classifier is the discriminator on the bottom and the second classifier is the discriminator on the top) 07-21-aia AIA Claims 7, 12, 13, 14, 15, 16, 17, 21, 22 and 23 are re jected under 35 U.S.C. 103 as being unpatentable over Wa ng and Bauer in view of Grover et al. (Bias Correction of Learned Generative Models using Likelihood-Free Importance Weighting); hereinafter Grover Cl aim 7 is rejected over Wang, Bauer and Grover with the incorporation of claim 4. Regarding claim 7, Wang does not appear to explicitly teach wherein the reweighting factor is applied to the one or more first values by resampling the one or more first values based on importance weights that are proportional to the reweighting factor. However, Grover teaches wherein the reweighting factor is applied to the one or more first values by resampling the one or more first values based on importance weights that are proportional to the reweighting factor. (Grover [page 5, 4 Importance Resampled Generative Modeling]: Sampling-Importance-Resampling. While exact sampling from p θ,Φ ;- is intractable, we can instead perform sample from a particle-based approximation to p θ,Φ ;- via sampling-importance-resampling [25, 26] (SIR). We define the SIR approximation to p θ,Φ ;- via the following density: (11); Note: See Algorithm 1 step 4 of Grover to see the proportionality.) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the importance weights of Grover to improve image generation (Grover, page 4, paragraph 3). Wang and Grover are analogous art because they both concern image generation. Claim 12 is rejected over Wang, Bauer and Grover with the incorporation of claim 4. Wang does not appear to explicitly teach further comprising calculating the reweighting factor based on output generated by the one or more classifiers from the one or more first values. However, Grover teaches further comprising calculating the reweighting factor based on output generated by the one or more classifiers from the one or more first values. (Grover [page 3, 3 Likelihood-Free Importance Weighting]: “To train the classifier, we only require datasets of samples from pθ(x) and p(x) and estimate γ to be the ratio of the size of two datasets. Let cφ : X → [0, 1] denote the probability assigned by the classifier with parameters φ to a sample x belonging to the positive class y = 1. As shown in prior work [9, 22], if cφ is Bayes optimal, then the importance weights can be obtained via this classifier as: PNG media_image1.png 43 226 media_image1.png Greyscale ) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the importance weights of Grover to improve image generation (Grover, page 4, paragraph 3). Wang and Grover are analogous art because they both concern image generation. Claim 13 is rejected over Wang, Bauer and Grover with the incorporation of claim 4. Regarding claim 13, Wang does not teach wherein performing the one or more operations to train the one or more classifiers comprises iteratively updating weights of the one or more classifiers based on a binary cross-entropy loss. However, Grover teaches wherein performing the one or more operations to train the one or more classifiers comprises iteratively updating weights of the one or more classifiers based on a binary cross-entropy loss. (Grover [page 16; C.2 Synthetic experiment]: “We believe the default calibration behavior is largely due to the fact that our binary classifiers distinguishing real and fake data do not require very complex neural networks architectures and training tricks that lead to miscalibration for multi-class classification. As shown in [61], shallow networks are well-calibrated and [62] further argue that a major reason for miscalibration is the use of a softmax loss typical for multi-class problems.”; page 16, C.1 Calibration; and “The classifier used in this case is a multi-layer perceptron with a single hidden layer of 100 units and has been trained to minimize the cross-entropy loss by first order optimization methods.”;) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the importance weights of Grover to improve image generation (Grover, page 4, paragraph 3). Wang and Grover are analogous art because they both concern image generation. Claim 14 is rejected over Wang, Bauer and Grover. Regarding claim 14, Wang teaches a non-transitory computer readable medium storing instructions that, when executed by a processor, cause the processor to perform the steps of: performing one or more operations based on a training dataset to train a generative model, (Wang [page 1, 1. Introduction]: “These models involve specifying a prior distribution over latent variables and defining a deep generative network (i.e., the decoder) that maps latent variables to data space in stochastic or deterministic fashion.”) wherein the generative model includes a first component that converts a plurality of data points included in the training dataset into a set of latent variables (Wang [page 2]: “They thus replace the KL-divergence with an adversarial loss imposed on the encoder output, requiring that the latent code z produced by the encoder should have an aggregated posterior distribution the same as the prior p(z).”; Note: The specification of the instant application addresses on paragraph [0005] that “The trained VAE would include an encoder network that converts each image into hundreds or thousands of numeric latent variable values. Each latent variable would represent a corresponding visual attribute found in one or more of the images used to train the VAE”;) and a second component that generates a prior distribution of the set of latent variables across the training dataset; (Wang [page 1, 1. Introduction]: “We replace the simple prior with a learned prior by training the code generator to output latent variables that will minimize an adversarial loss in data space.”; and [page 3, III. Learning the Prior]: “to find a prior that, together with the decoder in Fig. 3, would lead to a distribution that maximizes the data likelihood”) performing one or more operations to train one or more classifiers to distinguish between values for the set of latent variables generated via the first component and values sampled from the prior distribution; and (See Figure 3 of Wang to see that the discriminator on the bottom is the classifier that distinguishes the input from the gaussian distribution and the latent codes (visual attributes) from the encoder.) creating a trained prior component based on the second component and one or more classifiers, (See Figure 3 of Wang to see that the code generator (trained prior network) and the discriminators (classifiers) are structurally combined in the same architecture. The entire AAE architecture is the trained prior component.) wherein, in operation, the trained prior component produces the one or more second values in order to generate a new data point that is not included in the training dataset. (Wang [page 5]: “Fig. 6 further visualizes sample images generated with these models by driving the decoder with latent codes drawn from the prior or the code generator in our case. It is observed that our model produces much sharper images than the others . This confirms that a learned and flexible prior is beneficial to the characterization and generation of data.”) Wang does not appear to explicitly teach wherein the trained prior component applies a reweighting factor to one or more first values sampled from the prior distribution to generate one or more second values for the set of latent variables, However, Bauer teaches wherein the trained prior component (Bauer [page 4, 4 VAEs with resampled priors]: “we use the original VAE prior as a proposal distribution π(z) and define a resampled prior p(z) through a rejection sampler with learned acceptance probability a(z).”) applies a reweighting factor to one or more first values sampled from the prior distribution to generate one or more second values for the set of latent variables, (Bauer [page 2, 3 Learned Accept/Reject Sampling]: “We then transform π(z) into a more expressive distribution p∞(z) by multiplying it by an acceptance function aλ(z) ∈ [0; 1]; and [page 3]: “Thus, we can sample from p ∞ as follows: draw a candidate sample z from the proposal π and accept it with probability a(z ); if accepted, return z as the sample ; otherwise repeat the process. This means we can think of a(z) as a filtering function which — by rejecting different fractions of samples at different locations — can transform π(z) into a distribution closer to q(z).”; The acceptance function is the reweighting factor.) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the acceptance function of Bauer to improve VAE performance (Bauer, Abstract). Wang and Bauer are analogous art because they both concern priors in variational autoencoders. Wang does not appear to explicitly teach wherein the reweighting factor is determined based on output generated by the one or more classifiers from the one or more first values, However, Grover teaches wherein the reweighting factor is determined based on output generated by the one or more classifiers from the one or more first values, (Grover [page 1, 1 Introduction]: “Our proposed solution to estimate the importance weights is to train a calibrated, probabilistic classifier to distinguish samples from the data distribution and the generative model.”; and [page 3, 3 Likelihood-Free Importance Weighting]: “if c Φ is Bayes optimal, then the importance weights can be obtained via this classifier as:”) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the importance weights of Grover to improve image generation (Grover, page 4, paragraph 3). Wang and Grover are analogous art because they both concern image generation. Claim 15 is rejected over Wang, Bauer and Grover with the incorporation of claim 14. Regarding claim 15, Wang teaches wherein the instructions further cause the processor to perform the steps of performing one or more decoding operations on the one or more second values via a decoder network included in the generative model to generate the new data point. (Wang [page 5]: “Fig. 6 further visualizes sample images generated with these models by driving the decoder with latent codes drawn from the prior or the code generator in our case. It is observed that our model produces much sharper images than the others . This confirms that a learned and flexible prior is beneficial to the characterization and generation of data.”) Claim 16 is rejected over Wang, Bauer and Grover with the incorporation of claim 14. Regarding claim 16, Wang teaches wherein the decoder network is implemented by at least one of a generator network included in a generative adversarial network, a decoder portion of a variational autoencoder, or an invertible decoder represented by one or more normalizing flows (Wang [page 1, 1. Introduction]: “in learning the VAE with a simple encoder and decoder , [7] conjecture that multimodal priors can achieve a higher variational lower bound on the data log-likelihood than is possible with the standard normal prior.”) Dependent claim 17 is claim 7 in the form of a non-transitory computer readable medium and is rejected for the same reasons as claim 7 stated above. For the rejections of the limitations specifically pertaining to the non-transitory computer readable medium of claim 14, please see the rejection of claim 14 above. Claim 21 is rejected over Wang, Bauer and Grover with the incorporation of claim 14. Regarding claim 21, Wang does not appear to explicitly teach wherein the instructions further cause the processor to perform the step of generating the reweighting factor by computing a quotient of a probability that is output by the one or more classifiers and a difference between the probability and one. However, Grover teaches wherein the instructions further cause the processor to perform the step of generating the reweighting factor by computing a quotient of a probability that is output by the one or more classifiers and a difference between the probability and one. (Grover [page 3, 3 Likelihood-Free Importance Weighting]: “To train the classifier, we only require datasets of samples from pθ(x) and p(x) and estimate γ to be the ratio of the size of two datasets. Let cφ : X → [0, 1] denote the probability assigned by the classifier with parameters φ to a sample x belonging to the positive class y = 1. As shown in prior work [9, 22], if cφ is Bayes PNG media_image1.png 43 226 media_image1.png Greyscale optimal, then the importance weights can be obtained via this classifier as: ) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the importance weights of Grover to improve image generation (Grover, page 4, paragraph 3). Wang and Grover are analogous art because they both concern image generation. Claim 22 is rejected over Wang, Bauer and Grover with the incorporation of claim 14. Regarding claim 22, Wang teaches wherein the second component is implemented by at least one of a prior network or a Gaussian distribution. (Note: See Figure 3 of Wang to see that the second component is also implemented by a gaussian distribution) Claim 23 is rejected over Wang, Bauer and Grover with the incorporation of claim 14. Regarding claim 23, Wang teaches wherein the first component is implemented by at least one of an encoder portion of a variational autoencoder, a numerical inversion applied to a generator network included in a generative adversarial network, or an inverse of a decoder included in a normalizing flow network. (Wang [page 1, 1. Introduction]: “in learning the VAE with a simple encoder and decoder, [7] conjecture that multimodal priors can achieve a higher variational lower bound on the data log-likelihood than is possible with the standard normal prior.”; Note: See Figure 3 of Wang to see the first component consists of an encoder) 07-21-aia AIA Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Wang and Bauer in view of Che et al. (Your GAN is Secretly an Energy-based Model and You Should Use Discriminator Driven Latent Sampling); hereinafter Che Claim 8 is rejected over Wang, Bauer and Che with the incorporation of claim 4. Regarding claim 8, Wang does not appear to explicitly teach wherein the reweighting factor is applied to the one or more first values by iteratively updating the one or more first values based on a gradient of an energy function associated with the distribution learned by the trained prior network and the reweighting factor. PNG media_image2.png 45 372 media_image2.png Greyscale However, Che teaches wherein the reweighting factor is applied to the one or more first values by iteratively updating the one or more first values based on a gradient of an energy function associated with the distribution learned by the trained prior network and the reweighting factor. (Che [2.2. Energy-Based Models and Langevin Dynamics]: “One common MCMC algorithm in continuous state spaces is Langevin dynamics, with an update equation PNG media_image3.png 23 114 media_image3.png Greyscale Langevin dynamics are guaranteed to exactly sample from the target distribution p(x) as .”; 2.2. Energy-Based Models and Langevin Dynamics) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the energy function of Che to improve sample quality (Che, page 6). Wang and Che are analogous art because they both concern image generation using adversarial networks . 07-21-aia AIA Claim s 9, 10 and 11 are rejected under 35 U.S.C. 103 as being unpatentable over Wang and Bauer in view of Liu et al. (SENet for Weakly-Supervised Relation Extraction); hereinafter Liu Claim 9 is rejected over Wang, Bauer and Liu with the incorporation of claim 4. Regarding claim 9, Wang does not appear to explicitly teach wherein at least one of the one or more classifiers comprises a residual neural network. However, Liu teaches wherein at least one of the one or more classifiers comprises a residual neural network. (See Figure 2 of Liu to see that the SE-ResNet-D is a squeeze-and excitation residual network.) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the residual network of Liu to improve the representation power of the network (Liu, page 1). Wang and Liu are analogous art because they both concern encoding data. Claim 10 is rejected over Wang, Bauer and Liu with the incorporation of claim 4. Regarding claim 10, Wang does not appear to explicitly teach wherein the residual neural network includes a first batch normalization layer having a first Swish activation function, a first convolutional layer, a second batch normalization layer having a second Swish activation function, a second convolutional layer, and a squeeze and excitation layer. However, Liu teaches wherein the residual neural network includes a first batch normalization layer having a first Swish activation function, a first convolutional layer, a second batch normalization layer having a second Swish activation function, a second convolutional layer, and a squeeze and excitation layer. (Liu [page 512]: “We use double pooling and Swish activation function in our model, achieving a better result“; Note: See Figure 2 of Liu to see that the SE-ResNet-D is a squeeze-and excitation residual network that consists of convolutional layers, batch normalization layers, swish activation function and squeeze and excitation.) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the residual network of Liu to improve the representation power of the network (Liu, page 1). Zhang and Liu are analogous art because they both concern encoding data. Claim 11 is rejected over Wang, Bauer and Liu with the incorporation of claim 4. Regarding claim 11, Wang does not appear to explicitly teach wherein the residual neural network includes a Swish activation function and a sequence of convolutional kernels However, Liu teaches wherein the residual neural network includes a Swish activation function and a sequence of convolutional kernels (Liu [page 512]: “We use double pooling and Swish activation function in our model, achieving a better result“; Note: See Figure 2 of Liu to see that the SE-ResNet-D is residual network that consists of a swish activation function and a sequence of convolutional kernels.) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the residual network of Liu to improve the representation power of the network (Liu, page 1). Zhang and Liu are analogous art because they both concern encoding data . 07-21-aia AIA Claim s 18 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Wang, Bauer, and Grover in view of Che Dependent claim 18 is claim 8 in the form of a non-transitory computer readable medium and is rejected for the same reasons as claim 8 stated above. For the rejections of the limitations specifically pertaining to the non-transitory computer readable medium of claim 14, please see the rejection of claim 14 above. Claim 19 is rejected over Wang, Bauer, Grover and Che with the incorporation of claim 14. Regarding claim 19, Wang does not appear to explicitly teach wherein the energy function comprises a difference between the prior distribution and the reweighting factor. However, Che teaches wherein the energy function comprises a difference between the prior distribution and the reweighting factor. (Che [page 3, 3.2. Rejection Sampling and MCMC in Latent Space]: “Interestingly, pt(z) has the form of an energy-based model, p t (z) = e −E(z) / Z’, with tractable energy function E(z) = − log p 0 (z) − d(G(z)).” Note: p0(z) represents the prior distribution and d(G(z) represents the reweighting factor as determined by the discriminator output score) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the energy function of Che to improve sample quality (Che, page 6). Wang and Che are analogous art because they both concern image generation using adversarial networks . 07-21-aia AIA Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Wang, Bauer, and Grover in view of Liu Claim 20 is rejected over Wang, Bauer, Grover and Che with the incorporation of claim 14. Regarding claim 20, Wang does not appear to explicitly teach wherein at least one of the one or more classifiers comprises a sequence of residual blocks, and at least one residual block in the sequence of residual blocks comprises a first batch normalization layer with a first Swish activation function, a first convolutional layer following the first batch normalization layer with the first Swish activation function, a second batch normalization layer with a second Swish activation function, a second convolutional layer following the second batch normalization layer with the second Swish activation function, and a squeeze and excitation layer. However, Liu teaches wherein at least one of the one or more classifiers comprises a sequence of residual blocks, and at least one residual block in the sequence of residual blocks comprises a first batch normalization layer with a first Swish activation function, a first convolutional layer following the first batch normalization layer with the first Swish activation function, a second batch normalization layer with a second Swish activation function, a second convolutional layer following the second batch normalization layer with the second Swish activation function, and a squeeze and excitation layer. (Liu [page 512]: “We use double pooling and Swish activation function in our model, achieving a better result“; Note: See Figure 2 of Liu to see that the SE-ResNet-D is a squeeze-and excitation residual network that consists of convolutional layers, batch normalization layers, swish activation function and squeeze and excitation.) It would have been obvious before the effective filing date to combine the trained prior component of Wang with the residual network of Liu to improve the representation power of the network (Liu, page 1). Zhang and Liu are analogous art because they both concern encoding data . Conclusion 07-96 AIA The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. (Pub. No.: US 20210397797 A1) – “Zekang Li” relates to Prior network latent distribution NPL: Wang, Jiayu et al. “Unregularized Auto-Encoder with Generative Adversarial Networks for Image Generation.” (2018). NPL: Gorijala, Stanislav et al. “Adversarial Latent Autoencoders.” (2020). Any inquiry concerning this communication or earlier communications from the examiner should be directed to DAVID H TRAN whose telephone number is (703)756-1525. The examiner can normally be reached M-F 9:30 am - 5:30 pm. 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. /DAVID H TRAN/Examiner, Art Unit 2147 /VIKER A LAMARDO/Supervisory Patent Examiner, Art Unit 2147 Application/Control Number: 17/211,681 Page 2 Art Unit: 2147 Application/Control Number: 17/211,681 Page 3 Art Unit: 2147 Application/Control Number: 17/211,681 Page 4 Art Unit: 2147 Application/Control Number: 17/211,681 Page 5 Art Unit: 2147 Application/Control Number: 17/211,681 Page 6 Art Unit: 2147 Application/Control Number: 17/211,681 Page 7 Art Unit: 2147 Application/Control Number: 17/211,681 Page 8 Art Unit: 2147 Application/Control Number: 17/211,681 Page 9 Art Unit: 2147 Application/Control Number: 17/211,681 Page 10 Art Unit: 2147 Application/Control Number: 17/211,681 Page 11 Art Unit: 2147 Application/Control Number: 17/211,681 Page 12 Art Unit: 2147 Application/Control Number: 17/211,681 Page 13 Art Unit: 2147 Application/Control Number: 17/211,681 Page 14 Art Unit: 2147 Application/Control Number: 17/211,681 Page 15 Art Unit: 2147 Application/Control Number: 17/211,681 Page 16 Art Unit: 2147 Application/Control Number: 17/211,681 Page 17 Art Unit: 2147 Application/Control Number: 17/211,681 Page 18 Art Unit: 2147 Application/Control Number: 17/211,681 Page 19 Art Unit: 2147 Application/Control Number: 17/211,681 Page 20 Art Unit: 2147 Application/Control Number: 17/211,681 Page 21 Art Unit: 2147 Application/Control Number: 17/211,681 Page 22 Art Unit: 2147 Application/Control Number: 17/211,681 Page 23 Art Unit: 2147 Application/Control Number: 17/211,681 Page 24 Art Unit: 2147 Application/Control Number: 17/211,681 Page 25 Art Unit: 2147 Application/Control Number: 17/211,681 Page 26 Art Unit: 2147
Prosecution Insights