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 .
Remarks
This Office Action is responsive to Applicants' Amendment filed on January 5, 2026, in which claims 1, 9, 11, 19, 21, 28, and 30 are currently amended. Claims 1-30 are currently pending.
Specification
Applicant's amendments made to the specification are acknowledged. Examiner’s objection to the specification are hereby withdrawn, as necessitated by Applicant’s amendments made to the specification.
Response to Arguments
Applicant’s arguments with respect to rejection of claims 1-30 under 35 U.S.C. 101 based on amendment have been considered, however, are not persuasive.
Applicant’s arguments with respect to rejection of claims 1-30 under 35 U.S.C. 101 based on amendment have been considered, however, are not persuasive.
With respect to Applicant's arguments on pp. 13-14 of the Remarks submitted 1/5/2026 that the judicial exception is integrated into a practical application, and more specifically that "generating an inference using the neural network with one or more refined parameters [...] represent a practical application of machine learning inference", Examiner respectfully disagrees. The claim is directed wholly to a judicial exception performed on a generic computer system such that the judicial exception is not seen as improving the generic computer system. Examiner notes 2106.05(a) "It is important to note, the judicial exception alone cannot provide the improvement." MPEP 2106.05(a) also recites "An important consideration in determining whether a claim improves technology is the extent to which the claim covers a particular solution to a problem or a particular way to achieve a desired outcome, as opposed to merely claiming the idea of a solution or outcome." and MPEP 2106.07(a)(II) "employing well-known computer functions to execute an abstract idea, even when limiting the use of the idea to one particular environment, does not integrate the exception into a practical application". While Examiner has generously interpreted a neural network as a generic computer component in the Non-Final Office Action mailed 10/7/2025, Examiner notes that the equation y=m*x+b fully represents a single layer perceptron without activation (a neural network), and one of ordinary skill in the art could readily apply this equation entirely in the mind with or without the assistance of tools such as pen and paper. Even if, for the sake of argument, machine learning inference using refined parameters was for some reason not itself considered an abstract mental process, Examiner notes that performing inference with neural networks with refined parameters is not only well-understood, routine, and conventional in the art, but definitional to machine learning such that Examiner, respectfully, does not agree that the claims present a technical improvement. Applicant's arguments directed towards Example 48 of the 2024 SME eligibility examples, as well as previous court examples, are seen as moot as the instant claims are of significantly different matter and scope. For at least these reasons and those further detailed below Examiner asserts that it is reasonable and appropriate to maintain the rejection under 35 U.S.C. 101.
Applicant’s arguments with respect to rejection of claims 1-30 under 35 U.S.C. 101 based on amendment have been considered, however, are not persuasive.
With respect to Applicant's arguments on p. 20 of the Remarks submitted 1/5/2026 that Hajimolahoseini does not disclose "refining one or more parameters of the first layer of the neural network", Examiner respectfully disagrees. First, Examiner respectfully disagrees with Applicant's characterization of Hajimolahoseini, specifically, Applicant argues that in Hajimolahoseini "a tensor [is] unassociated with a neural network". The entire system in Hajimolahoseini is explicitly a convolutional neural network having convolutional layers and fully connected layers, this is plainly stated ([¶0002] "The present disclosure relates to artificial neural networks, including convolutional neural networks used to perform computer vision tasks based on multi-frame video data."). Second, Examiner notes that there is nothing in the instant claims or specification that would make it unreasonable to interpret down-sampling a parameter tensor as refining one or more parameters. Examiner also notes that Hajimolahoseini also explicitly states that the parameters of the neural network are trained (refined) such that there are multiple interpretations of Hajimolahoseini that clearly read on the claim limitation. (See also FIG. 4 and FIG. 7. while Examiner has interpreted the downsampling step (shown in FIG. 4 which is explicitly based on the stored second subset of the first data) as refining one or more parameters of the first layer (1st 2D spatial convolution layer 312 which outputs 4D tensor [B*T,X/s2,Y/s2,S] comprising (based at least in part on) both the first and second subsets B and T which are explicitly stored ([¶0034] “the present disclosure describes a non-transitory processor-readable medium having stored thereon an output tensor generated according to one or more of the methods described above.”)Examiner notes that alternative interpretations of Hajimolahoseni also satisfy the claim limitation. For example, Hajimolahoseni explicitly discloses updating the CNN parameters through training ([¶0061] “After forward propagation (propagation in a direction from 124 to 138 in FIG. 2 is forward propagation) is complete a loss function similar to category (i.e. class) cross-entropy is used to compute a prediction error of the CNN 120, and back propagation (propagation in a direction from 138 to 124 in FIG. 2 is back propagation) is performed to update the parameters (e.g. weights) of the layers 128, 130, 132, 134, 136, and 128 of the CNN 120 based on the computed prediction error” [¶0047] “Each kernel of a filter uses its own distinct set of nine weights that are tuned (i.e. adjusted) during training”). For at least these reasons and those further detailed below Examiner asserts that it is reasonable and appropriate to maintain the rejections under USC 102 and 103 in view of Hajimolahoseni.
With respect to Applicant's arguments on p. 21 of the Remarks submitted 1/5/2026 that the combination of Dinh and Hajimolahoseni does not disclose "applying a multiplication operation based on a summation of an output of the one or more convolution operations and the second data input tensor", Examiner respectfully disagrees. Dinh explicitly teaches ([p. 5] "For the models presented here, both s(·) and t(·) are rectified convolutional networks") and further Dinh explicitly states that the models are convolutional networks such that all operations are interpreted as convolution operations. See Eqn. 7 and 8 both of which have element-wise multiplication to generate second output data tensor yd+1:D. xd+1:D exp s(x1:d) interpreted as an output of one or more convolution operations which is explicitly summed with t(x1:d). For at least these reasons and those further detailed below Examiner asserts that it is reasonable and appropriate to maintain the rejections under USC 102 and 103 in view of Dinh and Hajimolahoseni.
Claim Rejections - 35 USC § 101
101 Rejection
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 1-30 are rejected under 35 USC § 101 because the claimed invention is directed to non-statutory subject matter.
Regarding Claim 1: Claim 1 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 1 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: Claim 1 under its broadest reasonable interpretation is a series of mental processes. For example, but for the generic computer components language, the above limitations in the context of this claim encompass neural network processing, including the following:
generating a first subset of the first data tensor and a second subset of the first data tensor using a tensor splitting operation (observation, evaluation, and judgement),
refining one or more parameters of the first layer of the neural network based at least in part on the stored second subset of the first data tensor (observation, evaluation, and judgement)
generating an inference using the neural network with the one or more refined parameters of the first layer of the neural network (observation, evaluation, and judgement)
Therefore, claim 1 recites an abstract idea which is a judicial exception.
Step 2A Prong Two Analysis: Claim 1 recites additional elements “computer implemented method of machine learning” and “neural network”. However, these additional features are computer components recited at a high-level of generality, such that they amount to no more than mere instructions to apply the judicial exception using a generic computer component. An additional element that merely recites the words “apply it” (or an equivalent) with the judicial exception, or merely includes instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea, does not integrate the judicial exception into a practical application (See MPEP 2106.05(f)). Claim 1 also recites additional elements “generating a first data tensor as output from a first layer of a neural network”, “storing the second subset of the first data tensor”, and “providing the first subset of the first data tensor to a subsequent layer of the neural network” which amounts to gathering and outputting data which is insignificant extra-solution activity (See MPEP 2106.05(g)). Therefore, claim 1 is directed to a judicial exception.
Step 2B Analysis: Claim 1 does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the lack of integration of the abstract idea into a practical application, the additional elements recited in claim 1 amount to no more than mere instructions to apply the judicial exception using a generic computer component and insignificant extra-solution activity. The gathering and outputting of data is considered well-understood, routine, and conventional in the art (See MPEP 2106.05(d)(II)(i) and 2106.05(d)(II)(iv)).
For the reasons above, claim 1 is rejected as being directed to non-patentable subject matter under §101. This rejection applies equally to independent claims 11 and 21, which recite a system and a computer program product, respectively, as well as to dependent claims 2-10, 12-20, and 22-29.
Independent claim 11 recites additional instructions to apply the judicial exception using generic computer components “A processing system, comprising: a memory comprising computer-executable instructions; and one or more processors configured to execute the computer-executable instructions and cause the processing system to perform an operation comprising”.
Independent claim 21 recites additional instructions to apply the judicial exception using generic computer components: “A non-transitory computer-readable medium comprising computer-executable instructions that, when executed by one or more processors of a processing system, cause the processing system to perform an operation comprising:”.
The additional limitations of the dependent claims are addressed briefly below:
Dependent claims 2, 12, and 22 recite additional observation, evaluation, and judgement “wherein the tensor splitting operation comprises: a downsampling operation to generate a first intermediate data tensor having a reduced spatial dimensionality, as compared to the first data tensor, and an increased channel depth, as compared to the first data tensor; and a bijective function to delineate the first intermediate data tensor into the first and second subsets of the first data tensor.”
Dependent claims 3, 13, and 23 recite additional observation, evaluation, and judgement “the first subset of the first data tensor corresponds to a first set of channels from the first intermediate tensor, the second subset of the first data tensor corresponds to a second set of channels from the first intermediate tensor, and the first and second sets of channels are non-overlapping.
Dependent claims 4, 14, and 24 recite additional observation, evaluation, and judgement “the first data tensor has dimensionality B×C×H×W, wherein B is a batch size, C is a channel depth, and H and W are spatial dimensions, and the first intermediate data tensor has dimensionality B×4xC×H2×W2”.
Dependent claims 5, 15, and 25 recite additional observation, evaluation, and judgement “the first and second subsets of the first data tensor each have dimensionality B×2xC×H2×W2.
Dependent claims 6 and 16 recite additional observation, evaluation, and judgement “the first subset of the first data tensor has dimensionality B×1×H2×W2, and the second subset of the first data tensor has dimensionality B×(C-1)×H2×W2.”.
Dependent claims 7, 17, and 26 recite additional mathematical calculations and relationships “wherein refining the one or more parameters of the first layer of the neural network comprises: recreating the first subset of the first data tensor using backpropagation through the subsequent layer of the neural network” (Explicitly disclosed as mathematical calculations and relationships in instant specification at [¶0091] “the recreated input tensor may be generated while computing gradients and updating model parameters during backpropagation. Using the chain rule, the machine learning system can thereby iterate backwards through the model layers, refining each in turn.”). Claims 7, 17, and 26 also recite additional observation, evaluation, and judgement “generating a recreated first data tensor by combining the recreated first subset of the first data tensor and the stored second subset of the first data tensor; and generating a recreated first data tensor by combining the recreated first subset of the first data tensor and the stored second subset of the first data tensor refining the one or more parameters of the first layer using backpropagation of the recreated first data tensor”
Dependent claims 8, 18, and 27 recite additional observation, evaluation, and judgement “performs an invertible operation.” And mere instructions to apply the judicial exception using generic computer components “the first layer of the neural network”
Dependent claims 9, 19, and 28 recite additional observation, evaluation, and judgement “generates the first data tensor based on a first input data tensor and a second input data tensor; the first data tensor comprises the first input data tensor and a third data tensor, and the third data tensor data tensor comprises a non-linear combination of the first and second input data tensors” And mere instructions to apply the judicial exception using generic computer components “the first layer of the neural network”
Dependent claims 10, 20, and 29 recite additional observation, evaluation, and judgement “applying the tensor splitting operation after each layer of a plurality of layers of the neural network”
Regarding Claim 30: Claim 30 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 30 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: Claim 30 under its broadest reasonable interpretation is a series of mental processes and mathematical calculations and relationships. For example, but for the generic computer components language, the above limitations in the context of this claim encompass neural network processing, including the following:
Generating, using the layer of the neural network, a first output data tensor and a second output data tensor by processing the first input data tensor and the second input data tensor (observation, evaluation, and judgement),
the first output data tensor is equal to the first input data tensor (observation, evaluation, and judgement)
the second output data tensor is generated by: applying one or more convolution operations to the first input data tensor (mathematical calculations and relationships)
applying a multiplication operation based on a summation of an output of the one or more convolution operations and the second input data tensor to generate the second output data tensor (mathematical calculations and relationships)
the first and second input data tensors can be reconstructed based on the first and second output data tensors (observation, evaluation, and judgement)
generating an inference based on the first and second output data tensors (observation, evaluation, and judgement)
Therefore, claim 30 recites an abstract idea which is a judicial exception.
Step 2A Prong Two Analysis: Claim 30 recites additional elements “using the layer of the neural network”. However, these additional features are computer components recited at a high-level of generality, such that they amount to no more than mere instructions to apply the judicial exception using a generic computer component. An additional element that merely recites the words “apply it” (or an equivalent) with the judicial exception, or merely includes instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea, does not integrate the judicial exception into a practical application (See MPEP 2106.05(f)). Claim 30 also recites additional elements “accessing, at a layer of a neural network, a first input data tensor and a second input data tensor, the second input data tensor being different from the first input data tensor” and “outputting the first and second output data tensors from the layer of the neural network” which amounts to gathering and outputting data which is insignificant extra-solution activity (See MPEP 2106.05(g)). Therefore, claim 30 is directed to a judicial exception.
Step 2B Analysis: Claim 30 does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the lack of integration of the abstract idea into a practical application, the additional elements recited in claim 30 amount to no more than mere instructions to apply the judicial exception using a generic computer component and insignificant extra-solution activity. The gathering and outputting of data is considered well-understood, routine, and conventional in the art (See MPEP 2106.05(d)(II)(i) and 2106.05(d)(II)(iv)).
For the reasons above, claim 30 is rejected as being directed to non-patentable subject matter under §101.
Therefore, when considering the elements separately and in combination, they do not add significantly more to the inventive concept. Accordingly, claims 1-30 are rejected under 35 U.S.C. § 101.
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.
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.
Claims 1, 8, 10, 11, 18, 20, 21, 27, and 29 are rejected under U.S.C. §102(a)(1) as being anticipated by Hajimolahoseini (US20230124075A1).
PNG
media_image1.png
384
702
media_image1.png
Greyscale
FIG. 4 of US20230124075A1
PNG
media_image2.png
384
702
media_image2.png
Greyscale
Interpretation of first layer in FIG. 4 of US20230124075A1
PNG
media_image3.png
384
702
media_image3.png
Greyscale
Interpretation of first tensor in FIG. 4 of US20230124075A1
PNG
media_image4.png
384
702
media_image4.png
Greyscale
Interpretation of FIG. 4 of US20230124075A1 showing split of tensor output of 1st Conv2D
Regarding claim 1, Hajimolahoseini teaches A computer-implemented method, comprising:([¶0023] "the present disclosure describes a system for processing an input tensor to generate an output tensor. The system comprises a processor device, and a memory storing machine-executable instructions which, when executed by the processor device, cause the system to perform a number of steps. The input tensor is obtained")
generating a first data tensor as output from a first layer of a neural network;([¶0022] "the present disclosure describes a method for processing an input tensor to generate an output tensor" [¶0047] "In deep CNNs, most of the computation time is spent on convolutional layers. In a convolutional layer, filters are convolved over inputs of the layer to generate outputs. In the examples described herein, the input of each convolutional layer may be a 4D tensor of values representing features, organized into four dimensions. In 2D convolution layers, as described herein, one or more 4D filters are used to perform respective 2D convolution operations on the input tensor to generate a 4D output tensor")
generating a first subset of the first data tensor and a second subset of the first data tensor using a tensor splitting operation;([¶0020] "splitting the 3D kernels into smaller 2D kernels may significantly reduce the number of parameters of the example CNNs" [¶0071] "The reshaping operation of step 708 splits the combined batch index-temporal dimension of the 4D spatial feature tensor 402 back into two separate batch index and temporal dimensions" See also FIG. 3 and FIG. 5 which shows that the convolution block 300 of FIG. 3 is repeated sequentially as layers (blocks) which explicitly comprise sublayers. The claim is significantly broad such that the first and second subset could be interpreted as any subset of tensors 336 and 332, input batches, layer subsets, split batch from spatial processing 310, or something else altogether. In the interest of further examination B and T split from B*T of the output of convolution layer 312, X, Y, s2, and S are interpreted as subsets of the first data tensor.)
storing the second subset of the first data tensor;([¶0034] “the present disclosure describes a non-transitory processor-readable medium having stored thereon an output tensor generated according to one or more of the methods described above.” [¶0051] "A set of machine-executable instructions 110 defining a 2D convolution block 300 as part of a video-processing CNN 500 are shown stored in the memory 108" See also FIG. 3)
providing the first subset of the first data tensor to a subsequent layer of the neural network; and([¶0072] "At 710, the reshaped 4D spatial feature tensor 334 is down-sampled in the temporal dimension T using a second 2D spatial convolution layer 314." Hajimolahoseni explicitly discloses that after the split/reshape ([¶0072] “At 710, the reshaped 4D spatial feature tensor 334 is down-sampled in the temporal dimension T using a second 2D spatial convolution layer 314.” Shown in FIG. 4). This step is repeated in the process flowchart in FIG. 7)
PNG
media_image5.png
384
702
media_image5.png
Greyscale
Interpretation of FIG. 4 of US20230124075A1 showing first subset provided to subsequent layer
refining one or more parameters of the first layer of the neural network based at least in part on the stored second subset of the first data tensor.([¶0072] "At 710, the reshaped 4D spatial feature tensor 334 is down-sampled in the temporal dimension T using a second 2D spatial convolution layer 314." See also FIG. 4 and FIG. 7. while Examiner has interpreted the downsampling step (shown in FIG. 4 which is explicitly based on the stored second subset of the first data) as refining one or more parameters of the first layer (1st 2D spatial convolution layer 312 which outputs 4D tensor [B*T,X/s2,Y/s2,S] comprising (based at least in part on) both the first and second subsets B and T which are explicitly stored ([¶0034] “the present disclosure describes a non-transitory processor-readable medium having stored thereon an output tensor generated according to one or more of the methods described above.”)Examiner notes that alternative interpretations of Hajimolahoseni also satisfy the claim limitation. For example, Hajimolahoseni explicitly discloses updating the CNN parameters through training ([¶0061] “After forward propagation (propagation in a direction from 124 to 138 in FIG. 2 is forward propagation) is complete a loss function similar to category (i.e. class) cross-entropy is used to compute a prediction error of the CNN 120, and back propagation (propagation in a direction from 138 to 124 in FIG. 2 is back propagation) is performed to update the parameters (e.g. weights) of the layers 128, 130, 132, 134, 136, and 128 of the CNN 120 based on the computed prediction error” [¶0047] “Each kernel of a filter uses its own distinct set of nine weights that are tuned (i.e. adjusted) during training”).)
generating an inference using the neural network with the one or more refined parameters of the first layer of the neural network.([¶0003] "Artificial neural networks are computational structures used to create and apply models for performing inference tasks." See also FIG. 6 where the final output of the model is explicitly prediction (inference) information 139).
PNG
media_image6.png
404
704
media_image6.png
Greyscale
FIG. 3 of US20230124075A1
PNG
media_image7.png
404
704
media_image7.png
Greyscale
Markup of FIG. 3 of US20230124075A1 highlighting where FIG. 4 plugs into FIG. 3
PNG
media_image8.png
274
704
media_image8.png
Greyscale
FIG. 5 of US20230124075A1 showing each block 300 in FIG. 3 repeated
PNG
media_image9.png
422
706
media_image9.png
Greyscale
Markup of FIG. 6 of US20230124075A1 showing inference
Regarding claim 8, Hajimolahoseini teaches The computer-implemented method of claim 1, wherein the first layer of the neural network performs an invertible operation.(Hajimolahoseini [¶0070] "The 4D spatial input tensor 332, having dimensions [B*T,X,Y,C], is processed by a first 2D spatial convolution layer 312 to generate a 4D spatial feature tensor 402" 2D spatial convolutional layer interpreted as invertible operation).
Regarding claim 10, Hajimolahoseini teaches The computer-implemented method of claim 1, further comprising applying the tensor splitting operation after each layer of a plurality of layers of the neural network.(Hajimolahoseini See FIGs. 3-5).
Regarding claims 11, 18, and 20, claims 11, 18, and 20 are directed towards a system for performing the methods of claims 1, 8, and 10, respectively. Therefore, the rejections applied to claims 1, 8, and 10 also apply to claims 11, 18, and 20. Claims 11, 18, and 20 also recite additional elements A processing system, comprising: a memory comprising computer-executable instructions; and one or more processors configured to execute the computer-executable instructions and cause the processing system to perform an operation comprising: (Hajimolahoseini [¶0023] "the present disclosure describes a system for processing an input tensor to generate an output tensor. The system comprises a processor device, and a memory storing machine-executable instructions which, when executed by the processor device, cause the system to perform a number of steps. The input tensor is obtained").
Regarding claims 21, 27, and 29, claims 21, 27, and 29 are directed towards a computer program product for performing the method of claims 1, 8, and 10, respectively. Therefore, the rejections applied to claims 1, 8, and 10 also apply to claims 21, 27, and 29. Claims 21, 27, and 29 also recite additional elements A non-transitory computer-readable medium comprising computer-executable instructions that, when executed by one or more processors of a processing system, cause the processing system to perform an operation comprising: (Hajimolahoseini [¶0023] "the present disclosure describes a system for processing an input tensor to generate an output tensor. The system comprises a processor device, and a memory storing machine-executable instructions which, when executed by the processor device, cause the system to perform a number of steps. The input tensor is obtained").
Claims 2-7, 9, 12-17, 19, 22-26, and 28 are rejected under U.S.C. §103 as being unpatentable over the combination of Hajimolahoseini and Dinh (“DENSITY ESTIMATION USING REAL NVP”, 2017).
Regarding claim 2, Hajimolahoseini teaches The computer-implemented method of claim 1, wherein the tensor splitting operation comprises: a downsampling operation to generate a first intermediate data tensor having a reduced spatial dimensionality, as compared to the first data tensor, and an increased channel depth, as compared to the first data tensor; and (Hajimolahoseini [¶0072] "At 710, the reshaped 4D spatial feature tensor 334 is down-sampled in the temporal dimension T using a second 2D spatial convolution layer 314." See also FIG. 4 and FIG. 7).
However, Hajimolahoseini doesn't explicitly teach a bijective function to delineate the first intermediate data tensor into the first and second subsets of the first data tensor.
Dinh, in the same field of endeavor, teaches a bijective function to delineate the first intermediate data tensor into the first and second subsets of the first data tensor. ([p. 4 §3.2] "We will build a flexible and tractable bijective function by stacking a sequence of simple bijections. In each simple bijection, part of the input vector is updated using a function which is simple to invert, but which depends on the remainder of the input vector in a complex way. We refer to each of these simple bijections as an affine coupling layer.").
Hajimolahoseini as well as Dinh are directed towards image processing with convolutional neural networks. Therefore, Hajimolahoseini as well as Dinh are analogous art in the same field of endeavor. It would have been obvious before the effective filing date of the claimed invention to combine the teachings of Hajimolahoseini with the teachings of Dinh by applying an alternating identity shortcut to convert the convolution block 300 into a invertible bijective coupling layer. Dinh provides as additional motivation for combination ([p. 3 §3] "we define a powerful class of bijective functions which enable exact and tractable density evaluation and exact and tractable inference." [pp. 9-10] "unlike both variational autoencoders and GANs, our technique is able to learn a semantically meaningful latent space which is as high dimensional as the input space. This may make the algorithm particularly well suited to semi-supervised learning tasks").
Regarding claim 3, the combination of Hajimolahoseini, and Dinh teaches The computer-implemented method of claim 2, wherein: the first subset of the first data tensor corresponds to a first set of channels from the first intermediate tensor,(Hajimolahoseini [¶0030] "an output channel dimension S indicating a number of spatial feature maps equal to the number of filters S1")
the second subset of the first data tensor corresponds to a second set of channels from the first intermediate tensor, and(Hajimolahoseini [¶0078] "The 2D temporal convolution layer 320 applies a number S2 of 2D convolution filters to the 4D temporal input tensor 336 to generate S2 temporal feature maps (i.e. output channels).")
the first and second sets of channels are non-overlapping.(Hajimolahoseini [¶0030] "an output channel dimension S indicating a number of spatial feature maps equal to the number of filters S1" [¶0078] "The 2D temporal convolution layer 320 applies a number S2 of 2D convolution filters to the 4D temporal input tensor 336 to generate S2 temporal feature maps (i.e. output channels)." S and S2 correspond to spatial and temporal feature maps (output channels), respectively.).
Regarding claim 4, the combination of Hajimolahoseini, and Dinh teaches The computer-implemented method of claim 2, wherein: the first data tensor has dimensionality B×C×H×W, wherein B is a batch size, C is a channel depth, and H and W are spatial dimensions, and(Hajimolahoseini [¶0030] "an output channel dimension S indicating a number of spatial feature maps equal to the number of filters S1" [¶0079] "having a batch index output dimension B" See FIG. 4, B*T interpreted as B, X/s2 interpreted as H, Y/s2 interpreted as W, and S interpreted as C.)
the first intermediate data tensor has dimensionality B×4C×H2×W2.(Dinh [p. 6 §3.6] "We implement a multi-scale architecture using a squeezing operation: for each channel, it divides the image into subsquares of shape 2 × 2 × c, then reshapes them into subsquares of shape 1 × 1 × 4c. The squeezing operation transforms an s × s × c tensor into an s2 ×s2 × 4c tensor (see Figure 3), effectively trading spatial size for number of channels" See FIG. 3 where first data tensor has B=1, C=1, H=4, W=4 and intermediate data tensors have B=1, C=4, H2=2, W2=2).
Regarding claim 5, the combination of Hajimolahoseini, and Dinh teaches The computer-implemented method of claim 4, wherein the first and second subsets of the first data tensor each have dimensionality B×2C×H2×W2.(Dinh [p. 6 §3.6] "We implement a multi-scale architecture using a squeezing operation: for each channel, it divides the image into subsquares of shape 2 × 2 × c, then reshapes them into subsquares of shape 1 × 1 × 4c. The squeezing operation transforms an s × s × c tensor into an s2 ×s2 × 4c tensor (see Figure 3), effectively trading spatial size for number of channels" FIG. 3 shows two subsets (white and black) of the input tensor, the respective subsets have dimensions (B=1, C=2, H2=2, W2=2)).
Regarding claim 6, the combination of Hajimolahoseini, and Dinh teaches The method of claim 4, wherein: the first subset of the first data tensor has dimensionality B×1×H2×W2, and the second subset of the first data tensor has dimensionality B×(C-1)×H2×W2.(Dinh [p. 6 §3.6] "We implement a multi-scale architecture using a squeezing operation: for each channel, it divides the image into subsquares of shape 2 × 2 × c, then reshapes them into subsquares of shape 1 × 1 × 4c. The squeezing operation transforms an s × s × c tensor into an s2 ×s2 × 4c tensor (see Figure 3), effectively trading spatial size for number of channels" See FIG. 3. C-1 interpreted as zero such that second subset is empty set and any of the remaining subsets are interpreted as the first subset of the first data tensor.).
Regarding claim 7, Hajimolahoseini teaches The computer-implemented method of claim 1, wherein refining the one or more parameters of the first layer of the neural network comprises: recreating the first subset of the first data tensor using backpropagation through the subsequent layer of the neural network;(Hajimolahoseini [¶0061] "back propagation (propagation in a direction from 138 to 124 in FIG. 2 is back propagation) is performed to update the parameters (e.g. weights) of the layers 128, 130, 132, 134, 136, and 128 of the CNN 120 based on the computed prediction error to reduce the prediction error between an ideal result (i.e. the ground truth in the training data) and the prediction result output by the output layer 138")
generating a recreated first data tensor by combining the recreated first subset of the first data tensor and the stored second subset of the first data tensor; and(Hajimolahoseini [¶0061] "back propagation (propagation in a direction from 138 to 124 in FIG. 2 is back propagation) is performed to update the parameters (e.g. weights) of the layers 128, 130, 132, 134, 136, and 128 of the CNN 120 based on the computed prediction error to reduce the prediction error between an ideal result (i.e. the ground truth in the training data) and the prediction result output by the output layer 138" Tensors updated through backpropagation including any of the subsets of the first data tensor are interpreted as the recreated first data tensor. See also FIG. 3-5)
refining the one or more parameters of the first layer using backpropagation of the recreated first data tensor.(Hajimolahoseini [¶0061] "back propagation (propagation in a direction from 138 to 124 in FIG. 2 is back propagation) is performed to update the parameters (e.g. weights) of the layers 128, 130, 132, 134, 136, and 128 of the CNN 120 based on the computed prediction error to reduce the prediction error between an ideal result (i.e. the ground truth in the training data) and the prediction result output by the output layer 138").
However, Hajimolahoseini doesn't explicitly teach generating a recreated first data tensor by combining the recreated first subset of the first data tensor and the stored second subset of the first data tensor.
Dinh, in the same field of endeavor, teaches generating a recreated first data tensor by combining the recreated first subset of the first data tensor and the stored second subset of the first data tensor([p. 4 §3.2] "We will build a flexible and tractable bijective function by stacking a sequence of simple bijections. In each simple bijection, part of the input vector is updated using a function which is simple to invert, but which depends on the remainder of the input vector in a complex way. We refer to each of these simple bijections as an affine coupling layer." See FIG. 2 and eqn. 4 and 5).
Hajimolahoseini as well as Dinh are directed towards image processing with convolutional neural networks. Therefore, Hajimolahoseini as well as Dinh are analogous art in the same field of endeavor. It would have been obvious before the effective filing date of the claimed invention to combine the teachings of Hajimolahoseini with the teachings of Dinh by applying an alternating identity shortcut to convert the convolution block 300 into a invertible bijective coupling layer. Dinh provides as additional motivation for combination ([p. 3 §3] "we define a powerful class of bijective functions which enable exact and tractable density evaluation and exact and tractable inference." [pp. 9-10] "unlike both variational autoencoders and GANs, our technique is able to learn a semantically meaningful latent space which is as high dimensional as the input space. This may make the algorithm particularly well suited to semi-supervised learning tasks").
Regarding claim 9, Hajimolahoseini teaches The computer-implemented method of claim 8, wherein: the first layer of the neural network generates the first data tensor based on a first input data tensor and a second input data tensor;(Hajimolahoseini See FIG. 3. Tensor 123 interpreted as a first input data tensor, tensor 336 interpreted as second input data tensor,).
However, Hajimolahoseini doesn't explicitly teach the first data tensor comprises the first input data tensor and a third data tensor, and the third data tensor data tensor comprises a non-linear combination of the first and second input data tensors.
Dinh, in the same field of endeavor, teaches the first data tensor comprises the first input data tensor and a third data tensor, and the third data tensor data tensor comprises a non-linear combination of the first and second input data tensors.(See Eqn. 8: x_(d+1):D interpreted as first data tensor,).
Hajimolahoseini as well as Dinh are directed towards image processing with convolutional neural networks. Therefore, Hajimolahoseini as well as Dinh are analogous art in the same field of endeavor. It would have been obvious before the effective filing date of the claimed invention to combine the teachings of Hajimolahoseini with the teachings of Dinh by applying an alternating identity shortcut to convert the convolution block 300 into a invertible bijective coupling layer. Dinh provides as additional motivation for combination ([p. 3 §3] "we define a powerful class of bijective functions which enable exact and tractable density evaluation and exact and tractable inference." [pp. 9-10] "unlike both variational autoencoders and GANs, our technique is able to learn a semantically meaningful latent space which is as high dimensional as the input space. This may make the algorithm particularly well suited to semi-supervised learning tasks").
Regarding claims 12-17 and 19, claims 12-17 and 19 are directed towards a system for performing the method of claims 2-7 and 9, respectively. Therefore, the rejections applied to claims 2-7 and 9 also apply to claims 12-17 and 19.
Regarding claims 22-26 and 28, claims 22-26 and 28 are directed towards a system0 for performing the method of claims 2-5, 7, and 9, respectively. Therefore, the rejections applied to claims 2-5, 7, and 9 also apply to claims 22-26 and 28.
Claim 30 is rejected under U.S.C. §103 as being unpatentable over the combination of Dinh and Hajimolahoseini.
Regarding claim 30, Dinh teaches A computer-implemented method, comprising: accessing, at a layer of a neural network, a first input data tensor and a second input data tensor;([p. 4] "A coupling layer applies a simple invertible transformation consisting of scaling followed by addition of a constant offset to one part x2 of the input vector conditioned on the remaining part of the input vector x1" See FIG. 2 and 4 and Eqn. 5-8 which each take two input tensors into a coupling layer.)
the second input data tensor being different from the first input data tensor([p. 4] "A coupling layer applies a simple invertible transformation consisting of scaling followed by addition of a constant offset to one part x2 of the input vector conditioned on the remaining part of the input vector x1" First part of input vector x1 interpreted as first input tensor, second part of input vector x1 interpreted as second input data tensor. Both input data tensors are explicitly input into the coupling layer.)
generating, using the layer of the neural network, a first output data tensor and a second output data tensor by processing the first input data tensor and the second input data tensor, wherein: the first output data tensor is equal to the first input data tensor;([p. 5 §3.5] "coupling layers can be powerful, their forward transformation leaves some components unchanged. This difficulty can be overcome by composing coupling layers in an alternating pattern, such that the components that are left unchanged in one coupling layer are updated in the next (see Figure 4(a))." [p. 6] "In this alternating pattern, units which remain identical in one transformation are modified in the next." See FIG. 2 and FIG. 4a.)
the second output data tensor is generated by: applying one or more convolution operations to the first input data tensor;([p. 5] "For the models presented here, both s(·) and t(·) are rectified convolutional networks" [p. 7 §4.1] "We use the multi-scale architecture described in Section 3.6 and use deep convolutional residual networks in the coupling layers with rectifier nonlinearity and skip-connections as suggested by [46]." Dinh explicitly states that the models are convolutional networks such that all operations are interpreted as convolution operations)
applying a multiplication operation based on a summation of an output of the one or more convolution operations and the second input data tensor to generate the second output data tensor; and([p. 5] "For the models presented here, both s(·) and t(·) are rectified convolutional networks" Dinh explicitly states that the models are convolutional networks such that all operations are interpreted as convolution operations. See Eqn. 7 and 8 both of which have element-wise multiplication to generate second output data tensor yd+1:D. xd+1:D exp s(x1:d) interpreted as an output of one or more convolution operations which is explicitly summed with t(x1:d).)
the first and second input data tensors can be reconstructed based on the first and second output data tensors; and([p. 4 §3.2] "We will build a flexible and tractable bijective function by stacking a sequence of simple bijections. In each simple bijection, part of the input vector is updated using a function which is simple to invert, but which depends on the remainder of the input vector in a complex way. We refer to each of these simple bijections as an affine coupling layer." See FIG. 2 and 5 and eqn. 4-8)
outputting the first and second output data tensors from the layer of the neural network.(See FIG. 2 and 5 and eqn. 4-8)
and generating an inference based on the first and second output data tensors.([Abstract] "Unsupervised learning of probabilistic models is a central yet challenging problem in machine learning. Specifically, designing models with tractable learning, sampling, inference and evaluation is crucial in solving this task" [p. 3] "See also Figure 1. Exact and efficient inference enables the accurate and fast evaluation of the model" [p. 6] "the number of hidden layer features in s and t is doubled. All variables which have been factored out at different scales are concatenated to obtain the final transformed output").
However, Dinh does not explicitly teach that the method is A computer-implemented method.
Hajimolahoseini, in the same field of endeavor, teaches A computer-implemented method ([¶0023] "the present disclosure describes a system for processing an input tensor to generate an output tensor. The system comprises a processor device, and a memory storing machine-executable instructions which, when executed by the processor device, cause the system to perform a number of steps. The input tensor is obtained).
Dinh as well as Hajimolahoseini are directed towards image processing with convolutional neural networks. Therefore, Hajimolahoseini as well as Dinh are analogous art in the same field of endeavor. It would have been obvious before the effective filing date of the claimed invention to combine the teachings of Hajimolahoseini with the teachings of Dinh by applying the method of Dinh on a computer. Hajimolahoseini provides as additional motivation for combination ([¶0051] “The system 100 includes one or more memories 108 (collectively referred to as “memory 108”), which may include a volatile or non-volatile memory (e.g., a flash memory, a random access memory (RAM), and/or a read-only memory (ROM)). The non-transitory memory 108 may store machine-executable instructions for execution by the processor 102, such as to carry out examples described in the present disclosure”).
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any extension fee pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SIDNEY VINCENT BOSTWICK whose telephone number is (571)272-4720. The examiner can normally be reached M-F 7:30am-5:00pm EST.
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, Miranda Huang can be reached on (571)270-7092. 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.
/SIDNEY VINCENT BOSTWICK/Examiner, Art Unit 2124
/MIRANDA M HUANG/Supervisory Patent Examiner, Art Unit 2124
Prosecution Insights