DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This action is responsive to the original application filed on 12/11/2023. Acknowledgment is made with respect to a claim of priority to PCT Application PCT/JP2021/022649 filed on 6/15/2021.
Claim Objections
Claims 1-15 are objected to because of the following informalities:
Independent claims 1, 10, and 13 recite the limitation “and has a configuration kernels are arranged in a kernel direction,” (emphasis added) which should read as “and has a configuration in which kernels are arranged in a kernel direction,” (emphasis added) for better grammatical clarity. Dependent claims 2-9, 11-12, and 14-15 depend on objected claims 1, 10, and 13, and are also objected to by virtue of this dependency. Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-15 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1 recites the limitations “determine a division position in a weight value group, which is a weight value group of at least one layer included in a given neural network model … add a connection layer that is a layer that connects respective output data obtained by calculating input data to the layer” (emphasis added). It is unclear whether “the layer” refers back to “at least one layer” or to the newly introduced “connection layer”. Please explain. For examination purposes, the limitations will be interpreted to mean “determine a division position in a weight value group, which is a weight value group of at least one layer included in a given neural network model … add a connection layer that is a layer that connects respective output data obtained by calculating input data to the connection layer” (emphasis added). Dependent claims 2-9 depend on indefinite claim 1, and are also rejected under 35 USC § 112(b) by virtue of this dependency. Independent claims 10 and 13 contain the same indefiniteness issues as claim 1, and are rejected under 35 USC § 112(b) for the same reasons as claim 1. Dependent claims 11-12 and 14-15 depend on indefinite claims 10 and 13, respectively, and are also rejected under 35 USC § 112(b) by virtue of this dependency Appropriate correction is required.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-15 are rejected under 35 U.S.C 101 because the claimed invention is directed to an abstract idea without significantly more. The analysis of the claims will follow the 2019 Revised Patent Subject Matter Eligibility Guidance, 84 Fed. Reg. 50 (“2019 PEG”).
When considering subject matter eligibility under 35 U.S.C. 101, it must be determined whether the claim is directed to one of the four statutory categories of invention, i.e., process, machine, manufacture, or composition of matter (Step 1). If the claim does fall within one of the statutory categories, the second step in the analysis is to determine whether the claim is directed to a judicial exception (Step 2A). The Step 2A analysis is broken into two prongs. In the first prong (Step 2A, Prong 1), it is determined whether or not the claims recite a judicial exception (e.g., mathematical concepts, mental processes, certain methods of organizing human activity). If it is determined in Step 2A, Prong 1 that the claims recite a judicial exception, the analysis proceeds to the second prong (Step 2A, Prong 2), where it is determined whether or not the claims integrate the judicial exception into a practical application. If it is determined at step 2A, Prong 2 that the claims do not integrate the judicial exception into a practical application, the analysis proceeds to determining whether the claim is a patent-eligible application of the exception (Step 2B). If an abstract idea is present in the claim, any element or combination of elements in the claim must be sufficient to ensure that the claim integrates the judicial exception into a practical application, or else amounts to significantly more than the abstract idea itself.
Claim 1
Step 1: The claim recites a device; therefore, it is directed to the statutory category of a machine.
Step 2A Prong 1: The claim recites, inter alia:
determine a division position in a weight value group, which is a weight value group of at least one layer included in a given neural network model, and has a configuration kernels are arranged in a kernel direction, each of which is obtained by arranging at least one or more weight values in a channel direction: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of determining a division position in a weight group, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper. For example, one can practically and mentally determine where to divide a tensor of weight values.
obtain multiple weight value groups by dividing the weight value group at the division position: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of dividing weight values at a position to obtain multiple weight groups, which is performed through mathematical computation.
add a connection layer that is a layer that connects respective output data obtained by calculating input data to the layer and respective weight value groups after division to make one output data: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of adding a connection layer by calculating input data and weight value groups after division, which is performed through mathematical computation as evidenced by paragraph [0153-0159] of the originally filed specification.
when regarding a ratio of the number of weight values that are 0 to the number of weight values in the weight value group as sparsity, determines the division position in the weight value group before division so that at least one weight value group after the division has sparsity higher than or equal to a predetermined value: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of determining a division position in a weight value group before performing a division operation, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper. For example, one can practically and mentally determine where to divide a tensor of weight values before dividing the values.
Step 2A Prong 2: The claim does not recite any additional limitations which integrate the abstract idea into a practical application. Specifically, the additional elements consist of “a memory configured to store instructions; a processor configured to execute the instructions to”,
The additional elements of “a memory configured to store instructions; a processor configured to execute the instructions to” amount to generic computer components used as a tool to perform an existing process. Thus, the additional elements amount to no more than a recitation of the words "apply it" (or an equivalent) or are more than mere instructions to implement an abstract idea or other exception on a computer (see MPEP § 2106.05(f)).
Thus, even when viewed individually and as an ordered combination, these additional elements do not integrate the abstract idea into a practical application and the claim is thus directed to the abstract idea.
Step 2B: Finally, the claim taken as a whole does not contain an inventive concept which provides significantly more than the abstract idea.
The additional elements of “a memory configured to store instructions; a processor configured to execute the instructions to” amount to generic computer components used as a tool to perform an existing process. Thus, the additional elements amount to no more than a recitation of the words "apply it" (or an equivalent) or are more than mere instructions to implement an abstract idea or other exception on a computer (see MPEP § 2106.05(f)).
Taken alone or in combination, the additional elements of the claim do not provide an inventive concept and thus the claim is subject-matter ineligible.
Claim 2
Step 1: A machine, as above.
Step 2A Prong 1: The claim recites, inter alia:
determines the division position so that the weight value group before the division is divided in the kernel direction: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of determining a division position, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
divides the weight value group before the division at the division position: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of dividing a weight value group, which is performed through mathematical computation.
adds the connection layer that makes the one output data by connecting the respective output data obtained by calculating the input data and the respective weight value groups after the division, in the channel direction: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of adding a connection layer by calculating input data and weight value groups after division, which is performed through mathematical computation as evidenced by paragraph [0153-0159] of the originally filed specification.
Step 2A Prong 2, Step 2B: The claim does not recite any additional elements that are sufficient to integrate the judicial exceptions into a practical application or amount to significantly more than the judicial exception. As such, the claim is ineligible.
Claim 3
Step 1: A machine, as above.
Step 2A Prong 1: The claim recites, inter alia:
sorts the kernels included in the weight value group before the division, according to a predetermined criterion: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of sorting kernels, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
determines the division position so that the weight value group after kernel sorting is divided in the kernel direction: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of determining a division position, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
divides the weight value group before the division at the division position: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of dividing a weight value group, which is performed through mathematical computation.
adds the connection layer that makes the one output data by connecting the respective output data obtained by calculating the input data and the respective weight value groups after the division, in the channel direction: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of adding a connection layer by calculating input data and weight value groups after division, which is performed through mathematical computation as evidenced by paragraph [0153-0159] of the originally filed specification.
adds an output data sorting layer that sorts channels of the one output data to correspond to order of the kernels in the weight value group before kernel sorting, based on change in order of the kernels due to sorting by the processor: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of sorting channels, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
Step 2A Prong 2, Step 2B: The claim does not recite any additional elements that are sufficient to integrate the judicial exceptions into a practical application or amount to significantly more than the judicial exception. As such, the claim is ineligible.
Claim 4
Step 1: A machine, as above.
Step 2A Prong 1: The claim recites, inter alia:
sorts the kernels included in the weight value group before the division, according to a predetermined criterion: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of sorting kernels, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
sorts channels of next layer of the layer whose weight value group is divided according to order of the kernels sorted by the processor: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of sorting channels, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
determines the division position so that the weight value group after kernel sorting is divided in the kernel direction: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of determining a division position, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
divides the weight value group before the division at the division position: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of dividing a weight value group, which is performed through mathematical computation.
adds the connection layer that makes the one output data by connecting the respective output data obtained by calculating the input data and the respective weight value groups after the division, in the channel direction: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of adding a connection layer by calculating input data and weight value groups after division, which is performed through mathematical computation as evidenced by paragraph [0153-0159] of the originally filed specification.
Step 2A Prong 2, Step 2B: The claim does not recite any additional elements that are sufficient to integrate the judicial exceptions into a practical application or amount to significantly more than the judicial exception. As such, the claim is ineligible.
Claim 5
Step 1: A machine, as above.
Step 2A Prong 1: The claim recites, inter alia:
sorts the kernels included in the weight value group before the division, in descending or ascending order of the number of weight values that are 0: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of sorting kernels, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
Step 2A Prong 2, Step 2B: The claim does not recite any additional elements that are sufficient to integrate the judicial exceptions into a practical application or amount to significantly more than the judicial exception. As such, the claim is ineligible.
Claim 6
Step 1: A machine, as above.
Step 2A Prong 1: The claim recites, inter alia:
determines the division position so that the weight value group before the division is divided in the channel direction: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of determining a division position, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
divides the weight value group before the division at the division position: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of dividing a weight value group, which is performed through mathematical computation.
adds the connection layer that derives the one output data by adding corresponding elements in the respective output data obtained by calculating the input data and the respective weight value groups after the division: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of adding a connection layer by calculating input data and weight value groups after division, which is performed through mathematical computation as evidenced by paragraph [0153-0159] of the originally filed specification.
Step 2A Prong 2, Step 2B: The claim does not recite any additional elements that are sufficient to integrate the judicial exceptions into a practical application or amount to significantly more than the judicial exception. As such, the claim is ineligible.
Claim 7
Step 1: A machine, as above.
Step 2A Prong 1: The claim recites, inter alia:
sorts the channels included in the weight value group before the division, according to a predetermined criterion: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of sorting channels, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
adds an input data sorting layer that sorts channels of the input data according to order of the channels sorted by the processor: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of sorting channels, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
determines the division position so that the weight value group after channel sorting is divided in the channel direction: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of determining a division position, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
divides the weight value group before the division at the division position: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of dividing a weight value group, which is performed through mathematical computation.
adds the connection layer that derives the one output data by adding corresponding elements in the respective output data obtained by calculating the input data and the respective weight value groups after the division: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of adding a connection layer by calculating input data and weight value groups after division, which is performed through mathematical computation as evidenced by paragraph [0153-0159] of the originally filed specification.
Step 2A Prong 2, Step 2B: The claim does not recite any additional elements that are sufficient to integrate the judicial exceptions into a practical application or amount to significantly more than the judicial exception. As such, the claim is ineligible.
Claim 8
Step 1: A machine, as above.
Step 2A Prong 1: The claim recites, inter alia:
sorts the channels included in the weight value group before the division, according to a predetermined criterion: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of sorting channels, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
sorts kernels of weight value group of previous layer of the layer whose weight value group is divided according to order of the channels sorted by the processor: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of sorting kernels, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
determines the division position so that the weight value group after channel sorting is divided in the channel direction: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of determining a division position, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
divides the weight value group before the division at the division position: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of dividing a weight value group, which is performed through mathematical computation.
adds the connection layer that derives the one output data by adding corresponding elements in the respective output data obtained by calculating the input data and the respective weight value groups after the division: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mathematical concept of adding a connection layer by calculating input data and weight value groups after division, which is performed through mathematical computation as evidenced by paragraph [0153-0159] of the originally filed specification.
Step 2A Prong 2, Step 2B: The claim does not recite any additional elements that are sufficient to integrate the judicial exceptions into a practical application or amount to significantly more than the judicial exception. As such, the claim is ineligible.
Claim 9
Step 1: A machine, as above.
Step 2A Prong 1: The claim recites, inter alia:
sorts the channels included in the weight value group before the division, in descending or ascending order of the number of weight values that are 0: Under its broadest reasonable interpretation in light of the specification, this limitation encompasses the mental process of sorting kernels, which is an evaluation or observation that is practically capable of being performed in the human mind with the assistance of pen and paper.
Step 2A Prong 2, Step 2B: The claim does not recite any additional elements that are sufficient to integrate the judicial exceptions into a practical application or amount to significantly more than the judicial exception. As such, the claim is ineligible.
Claims 10-12
Claims 10-12 recite a method (step 1: a process) to perform the steps of claims 1, 2, and 6, respectively, without any additional elements that integrate the abstract ideas into a practical application or provide significantly more than the abstract idea by itself, and are thus rejected for the same reasons set forth in the rejection of claims 1, 2, and 6, respectively.
Claims 13-15
Claims 13-15 recite a non-transitory computer readable recording medium (step 1: a manufacture) using a computer and program to perform the steps of claims 1, 2, and 6, respectively, which by MPEP 2106.05(f) (“apply it”) cannot integrate an abstract idea into a practical application or provide significantly more than the abstract idea by itself, and are thus rejected for the same reasons set forth in the rejection of claims 1, 2, and 6, respectively.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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.
Claims 1, 2, 6, and 10-15 are rejected under 35 USC § 103 as being obvious over Wang et al. (Wang et al., “Exploring Sparsity in Image Super-Resolution for Efficient Inference”, Apr. 1, 2021, arXiv:2006.09603v2, pp. 1-10, hereinafter “Wang”) in view of Aliabadi et al. (US 20180096226 A1, hereinafter “Aliabadi”) and Anwar et al. (Anwar et al., “COMPACT DEEP CONVOLUTIONAL NEURAL NETWORKS WITH COARSE PRUNING”, Oct. 30, 2016, arXiv:1610.09639v1, pp. 1-10, hereinafter “Anwar”).
Regarding claim 1, Wang discloses [a] neural network model conversion device comprising: a memory configured to store instructions; a processor configured to execute the instructions to: (Abstract; “In this paper, we explore the sparsity in image SR to improve inference efficiency of SR networks. Specifically, we develop a Sparse Mask SR (SMSR) network to learn sparse masks to prune redundant computation”, which discloses a neural network conversion or pruning device/system/method; and §5 and 5.1; the experiments section of Wang is inherently performed using a device to implement the experiments)
determine a division position in a weight value group, which is a weight value group of at least one layer included in a given neural network model, obtain multiple weight value groups by dividing the weight value group at the division position; and (§4.2.2; “During the inference phase, sparse convolution is performed based on the predicted spatial and channel masks, as shown in Fig. 5(d). Take the lth layer in the kth SMM as an example, its kernel is first splitted into four sub-kernels according to Mchk;l1 and Mch k;l to obtain four convolutions. Meanwhile, input feature F is splitted into FD and FS based on Mchk;l1. Then, FD is fed to convolutions À andÁ to produce FD2D and FD2S, while FS is fed to convolutions  and à to produce FS2D and FS2S. Note that, FD2D is produced by a vanilla “dense” convolution while FD2S, FS2D and FS2S are generated by sparse convolutions with only “important” regions (marked by Mspak ) being computed”, which discloses splitting a layer’s kernel into multiple (four) sub-kernels according to channel masks, to obtain multiple separate convolutions, thus “obtaining multiple weight value groups by dividing the weight value group”. Further, the “division position” is determined by the channel mask generated by equation 2 of §4.1, which designates which portions of the kernel or weight tensor are treated as belonging to the “dense” subgroup versus the “sparse” subgroup. The channel mask is the claimed division position because it specifies where the split occurs; and §4.1, Equation 2)
add a connection layer that is a layer that connects respective output data obtained by calculating input data to the layer and respective weight value groups after division to make one output data (§4.2.2; “Finally, features obtained from these four branches are summed and concatenated to produce the output feature Fout”, which discloses connecting/adding/concatenating respective output data from computed output branches to merge the output data into one output data, wherein the claimed “connection layer” is a concatenation operator or elementwise sum operator that merges multiple branch outputs into one tensor; and §4.2.1; “.Finally, all these features are summed up to generate the output feature Fout”).
Wang fails to explicitly disclose but Aliabadi discloses and has a configuration kernels are arranged in a kernel direction, each of which is obtained by arranging at least one or more weight values in a channel direction; ([0030]; “A kernel stack of a CNN can include M rows of kernels and N columns of kernels, with each column also referred to as a filter bank of the kernel stack. The kernels of the kernel stack can have the same width and the same height. The convolutional layer can have M input channels for receiving M input activation maps. The convolutional layer can have N output channels for producing N output activation maps. Each output activation map can be a result of a three-dimensional convolution of a filter bank of the kernel stack and the corresponding input activation maps”, which discloses a “filter bank” that corresponds to a “kernel”, and the filter banks are arranged along the column axis or the claimed “kernel direction”. Within one filter bank, the several channel weight sub blocks correspond to the claimed “weight values in a channel direction” because there is one sub block per input channel arranged along the row or channel axis; and [0057]; “The kernel stack 208 includes N columns of kernels with each column also referred to as a filter bank of the kernel stack 208. For example, the column of kernels 208 a 1, 208 b 1, . . . , and 208 m 1 forms a filter bank of the kernel stack 208”; and Figure 2).
Wang and Aliabadi are analogous art because both are concerned with convolutional neural network processing. Before the effective filing date of the claimed invention, it would have been obvious to one skilled in convolutional neural network processing to combine the kernel and channel direction configuration of Aliabadi with the device of Wang to yield to the predictable result of and has a configuration kernels are arranged in a kernel direction, each of which is obtained by arranging at least one or more weight values in a channel direction. The motivation for doing so would be to provide for the efficient implementation of a convolutional layer of a convolutional neural network (Aliabadi; Abstract).
Wang fails to explicitly disclose but Anwar discloses wherein the processor, when regarding a ratio of the number of weight values that are 0 to the number of weight values in the weight value group as sparsity, determines the division position in the weight value group before division so that at least one weight value group after the division has sparsity higher than or equal to a predetermined value (§4; “Kernel-pruning zeroes k×k kernels and is neither too fine nor too coarse”, which discloses that sparsity is defined and measured as the fraction of weight values set to zero within the convolution layer; and §3; “For a specific pruning ratio, we search for the best pruning masks which afflicts the least adversary on the pruned network. Indeed retraining can partially or fully recover the pruning losses, but the lesser the losses, the more plausible is the recovery”, which discloses selecting a candidate by achieving a specific pruning ratio, such as a target zero-weight ratio; and §5.1.1; “Training the network with batch normalization Ioffe & Szegedy (2015) enables us to directly prune a network for a target ratio, instead of taking small sized steps”; and Abstract; “Experiments with the CIFAR-10 dataset show that more than 85% sparsity can be induced in the convolution layers with less than 1% increase in the misclassification rate of the baseline network”, which discloses that at least one weight value group after the division has had sparsity equal to or higher than a predetermined value).
Wang, Aliabadi, and Anwar are analogous art because all are concerned with convolutional neural network processing. Before the effective filing date of the claimed invention, it would have been obvious to one skilled in convolutional neural network processing to combine the zero weight ratios and division position of Anwar with the device of Wang and kernel and channel directions of Alibadi to yield to the predictable result of wherein the processor, when regarding a ratio of the number of weight values that are 0 to the number of weight values in the weight value group as sparsity, determines the division position in the weight value group before division so that at least one weight value group after the division has sparsity higher than or equal to a predetermined value. The motivation for doing so would be to reduce the computational complexity of a deep convolutional neural network (Anwar; Abstract).
Regarding claim 10, it is a method claim corresponding to the steps of claim 1, and is rejected for the same reasons as claim 1.
Regarding claim 13, it is a non-transitory computer-readable medium claim corresponding to the steps of claim 1, and is rejected for the same reasons as claim 1.
Regarding claims 2, 11, and 14, the rejection of claims 1, 10, and 13 are incorporated and Wang further discloses the processor adds the connection layer that makes the one output data by connecting the respective output data obtained by calculating the input data and the respective weight value groups after the division, (§4.2.2; “Finally, features obtained from these four branches are summed and concatenated to produce the output feature Fout”, which discloses connecting/adding/concatenating respective output data from computed output branches to merge the output data into one output data, wherein the claimed “connection layer” is a concatenation operator or elementwise sum operator that merges multiple branch outputs into one tensor; and §4.2.1; “.Finally, all these features are summed up to generate the output feature Fout”).
Wang fails to explicitly disclose but Aliabadi discloses the kernel direction … in the channel direction ([0030]; and [0057]).
The motivation to combine Wang and Aliabadi is the same as discussed above with respect to claim 1.
Wang fails to explicitly disclose but Anwar discloses wherein the processor determines the division position so that the weight value group before the division is divided in the kernel (§4; and §1; and Figure 1(c))
the processor divides the weight value group before the division at the division position (§4; and §1; and Figure 1(c)).
The motivation to combine Wang, Aliabadi, and Anwar is the same as discussed above with respect to claim 1.
Regarding claims 6, 12, and 15, the rejection of claims 1, 10, and 13 are incorporated and Wang further discloses the processor adds the connection layer that derives the one output data by adding corresponding elements in the respective output data obtained by calculating the input data and the respective weight value groups after the division, (§4.2.2; “Finally, features obtained from these four branches are summed and concatenated to produce the output feature Fout”, which discloses connecting/adding/concatenating respective output data from computed output branches to merge the output data into one output data, wherein the claimed “connection layer” is a concatenation operator or elementwise sum operator that merges multiple branch outputs into one tensor; and §4.2.1; “.Finally, all these features are summed up to generate the output feature Fout”).
Wang fails to explicitly disclose but Aliabadi discloses the kernel direction … in the channel direction ([0030]; and [0057]).
The motivation to combine Wang and Aliabadi is the same as discussed above with respect to claim 1.
Wang fails to explicitly disclose but Anwar discloses wherein the processor determines the division position so that the weight value group before the division is divided (§4; and §1; and Figure 1(c))
the processor divides the weight value group before the division at the division position (§4; and §1; and Figure 1(c)).
The motivation to combine Wang, Aliabadi, and Anwar is the same as discussed above with respect to claim 1.
Claims 3-5 and 7-9 are rejected under 35 USC § 103 as being obvious over Wang in view of Aliabadi and Anwar and further in view of Li et al. (Li et al., “PRUNING FILTERS FOR EFFICIENT CONVNETS”, Mar. 10, 2017, arXiv:1608.08710v3, pp. 1-13, hereinafter “Li”).
Regarding claim 3, the rejection of claim 1 is incorporated and Wang further discloses the processor adds the connection layer that makes the one output data by connecting the respective output data obtained by calculating the input data and the respective weight value groups after the division, in the channel direction, and (§4.2.2; “Finally, features obtained from these four branches are summed and concatenated to produce the output feature Fout”, which discloses connecting/adding/concatenating respective output data from computed output branches to merge the output data into one output data, wherein the claimed “connection layer” is a concatenation operator or elementwise sum operator that merges multiple branch outputs into one tensor; and §4.2.1; “.Finally, all these features are summed up to generate the output feature Fout”).
Wang fails to explicitly disclose but Aliabadi discloses the kernel direction … in the channel direction ([0030]; and [0057]).
The motivation to combine Wang and Aliabadi is the same as discussed above with respect to claim 1.
Wang fails to explicitly disclose but Anwar discloses wherein the processor determines the division position so that the weight value group after kernel sorting is divided in the kernel (§4; and §1; and Figure 1(c))
the processor divides the weight value group before the division at the division position (§4; and §1; and Figure 1(c)).
The motivation to combine Wang, Aliabadi, and Anwar is the same as discussed above with respect to claim 1.
Wang fails to explicitly disclose but Li discloses the processor sorts the kernels included in the weight value group before the division, according to a predetermined criterion (§3.1, Steps 1-4; “1. For each filter Fi,j, calculate the sum of its absolute kernel weights sj = ni l=1 2. Sort the filters by sj. |Kl|. 3. Prune m filters with the smallest sum values and their corresponding feature maps. The kernels in the next convolutional layer corresponding to the pruned feature maps are also removed. 4. Anewkernel matrix is created for both the ith and i + 1th layers, and the remaining kernel weights are copied to the new model”)
wherein the processor adds an output data sorting layer that sorts channels of the one output data to correspond to order of the kernels in the weight value group before kernel sorting, based on change in order of the kernels due to sorting by the processor (§3.1, Steps 1-4; “1. For each filter Fi,j, calculate the sum of its absolute kernel weights sj = ni l=1 2. Sort the filters by sj. |Kl|. 3. Prune m filters with the smallest sum values and their corresponding feature maps. The kernels in the next convolutional layer corresponding to the pruned feature maps are also removed. 4. Anewkernel matrix is created for both the ith and i + 1th layers, and the remaining kernel weights are copied to the new model”).
Wang, Aliabadi, Anwar, and Li are analogous art because all are concerned with convolutional neural network processing. Before the effective filing date of the claimed invention, it would have been obvious to one skilled in convolutional neural network processing to combine the kernel sorting of Li with the device of Wang and kernel and channel directions of Alibadi and weight ratios of Anwar to yield to the predictable result of the processor sorts the kernels included in the weight value group before the division, according to a predetermined criterion … wherein the processor adds an output data sorting layer that sorts channels of the one output data to correspond to order of the kernels in the weight value group before kernel sorting, based on change in order of the kernels due to sorting by the processor. The motivation for doing so would be to reduce computation costs of a deep convolutional neural network (Li; Abstract).
Regarding claim 4, the rejection of claim 1 is incorporated and Wang further discloses the processor adds the connection layer that makes the one output data by connecting the respective output data obtained by calculating the input data and the respective weight value groups after the division, in the channel direction (§4.2.2; “Finally, features obtained from these four branches are summed and concatenated to produce the output feature Fout”, which discloses connecting/adding/concatenating respective output data from computed output branches to merge the output data into one output data, wherein the claimed “connection layer” is a concatenation operator or elementwise sum operator that merges multiple branch outputs into one tensor; and §4.2.1; “.Finally, all these features are summed up to generate the output feature Fout”).
Wang fails to explicitly disclose but Aliabadi discloses the kernel direction … in the channel direction ([0030]; and [0057]).
The motivation to combine Wang and Aliabadi is the same as discussed above with respect to claim 1.
Wang fails to explicitly disclose but Anwar discloses wherein the processor determines the division position so that the weight value group after kernel sorting is divided in the kernel direction (§4; and §1; and Figure 1(c))
the processor divides the weight value group before the division at the division position (§4; and §1; and Figure 1(c)).
The motivation to combine Wang, Aliabadi, and Anwar is the same as discussed above with respect to claim 1.
Wang fails to explicitly disclose but Li discloses the processor sorts the kernels included in the weight value group before the division, according to a predetermined criterion (§3.1, Steps 1-4; “1. For each filter Fi,j, calculate the sum of its absolute kernel weights sj = ni l=1 2. Sort the filters by sj. |Kl|. 3. Prune m filters with the smallest sum values and their corresponding feature maps. The kernels in the next convolutional layer corresponding to the pruned feature maps are also removed. 4. Anewkernel matrix is created for both the ith and i + 1th layers, and the remaining kernel weights are copied to the new model”)
the processor sorts channels of next layer of the layer whose weight value group is divided according to order of the kernels sorted by the processor (§3.1, Steps 1-4; “1. For each filter Fi,j, calculate the sum of its absolute kernel weights sj = ni l=1 2. Sort the filters by sj. |Kl|. 3. Prune m filters with the smallest sum values and their corresponding feature maps. The kernels in the next convolutional layer corresponding to the pruned feature maps are also removed. 4. Anewkernel matrix is created for both the ith and i + 1th layers, and the remaining kernel weights are copied to the new model”).
The motivation to combine Wang, Aliabadi, Anwar, and Li is the same as discussed above with respect to claim 3.
Regarding claim 5, the rejection of claims 1 and 3 are incorporated and Wang fails to explicitly disclose but Li discloses sorts the kernels included in the weight value group before the division, in descending or ascending order of the number of weight values that are 0 (§3.1, Steps 1-4; “1. For each filter Fi,j, calculate the sum of its absolute kernel weights sj = ni l=1 2. Sort the filters by sj. |Kl|. 3. Prune m filters with the smallest sum values and their corresponding feature maps. The kernels in the next convolutional layer corresponding to the pruned feature maps are also removed. 4. Anewkernel matrix is created for both the ith and i + 1th layers, and the remaining kernel weights are copied to the new model”).
The motivation to combine Wang, Aliabadi, Anwar, and Li is the same as discussed above with respect to claim 3.
Regarding claim 7, the rejection of claim 1 is incorporated and Wang further discloses the processor adds the connection layer that derives the one output data by adding corresponding elements in the respective output data obtained by calculating the input data and the respective weight value groups after the division (§4.2.2; “Finally, features obtained from these four branches are summed and concatenated to produce the output feature Fout”, which discloses connecting/adding/concatenating respective output data from computed output branches to merge the output data into one output data, wherein the claimed “connection layer” is a concatenation operator or elementwise sum operator that merges multiple branch outputs into one tensor; and §4.2.1; “.Finally, all these features are summed up to generate the output feature Fout”).
Wang fails to explicitly disclose but Aliabadi discloses the channel direction ([0030]; and [0057]).
The motivation to combine Wang and Aliabadi is the same as discussed above with respect to claim 1.
Wang fails to explicitly disclose but Anwar discloses wherein the processor determines the division position so that the weight value group after channel sorting is divided in the channel direction (§4; and §1; and Figure 1(c))
the processor divides the weight value group before the division at the division position (§4; and §1; and Figure 1(c)).
The motivation to combine Wang, Aliabadi, and Anwar is the same as discussed above with respect to claim 1.
Wang fails to explicitly disclose but Li discloses the processor sorts the channels included in the weight value group before the division, according to a predetermined criterion; and the processor adds an input data sorting layer that sorts channels of the input data according to order of the channels sorted by the processor, (§3.1, Steps 1-4; “1. For each filter Fi,j, calculate the sum of its absolute kernel weights sj = ni l=1 2. Sort the filters by sj. |Kl|. 3. Prune m filters with the smallest sum values and their corresponding feature maps. The kernels in the next convolutional layer corresponding to the pruned feature maps are also removed. 4. Anewkernel matrix is created for both the ith and i + 1th layers, and the remaining kernel weights are copied to the new model”).
The motivation to combine Wang, Aliabadi, Anwar, and Li is the same as discussed above with respect to claim 3.
Regarding claim 8, the rejection of claim 1 is incorporated and Wang further discloses the processor adds the connection layer that derives the one output data by adding corresponding elements in the respective output data obtained by calculating the input data and the respective weight value groups after the division (§4.2.2; “Finally, features obtained from these four branches are summed and concatenated to produce the output feature Fout”, which discloses connecting/adding/concatenating respective output data from computed output branches to merge the output data into one output data, wherein the claimed “connection layer” is a concatenation operator or elementwise sum operator that merges multiple branch outputs into one tensor; and §4.2.1; “.Finally, all these features are summed up to generate the output feature Fout”).
Wang fails to explicitly disclose but Aliabadi discloses the channel direction ([0030]; and [0057]).
The motivation to combine Wang and Aliabadi is the same as discussed above with respect to claim 1.
Wang fails to explicitly disclose but Anwar discloses determines the division position so that the weight value group after channel sorting is divided in the channel direction (§4; and §1; and Figure 1(c))
the processor divides the weight value group before the division at the division position (§4; and §1; and Figure 1(c)).
The motivation to combine Wang, Aliabadi, and Anwar is the same as discussed above with respect to claim 1.
Wang fails to explicitly disclose but Li discloses the processor sorts the channels included in the weight value group before the division, according to a predetermined criterion; and the processor sorts kernels of weight value group of previous layer of the layer whose weight value group is divided according to order of the channels sorted by the processor (§3.1, Steps 1-4; “1. For each filter Fi,j, calculate the sum of its absolute kernel weights sj = ni l=1 2. Sort the filters by sj. |Kl|. 3. Prune m filters with the smallest sum values and their corresponding feature maps. The kernels in the next convolutional layer corresponding to the pruned feature maps are also removed. 4. Anewkernel matrix is created for both the ith and i + 1th layers, and the remaining kernel weights are copied to the new model”).
The motivation to combine Wang, Aliabadi, Anwar, and Li is the same as discussed above with respect to claim 3.
Regarding claim 9, the rejection of claims 1 and 7 are incorporated and Wang fails to explicitly disclose but Li discloses wherein the processor sorts the channels included in the weight value group before the division, in descending or ascending order of the number of weight values that are 0 (§3.1, Steps 1-4; “1. For each filter Fi,j, calculate the sum of its absolute kernel weights sj = ni l=1 2. Sort the filters by sj. |Kl|. 3. Prune m filters with the smallest sum values and their corresponding feature maps. The kernels in the next convolutional layer corresponding to the pruned feature maps are also removed. 4. Anewkernel matrix is created for both the ith and i + 1th layers, and the remaining kernel weights are copied to the new model”).
The motivation to combine Wang, Aliabadi, Anwar, and Li is the same as discussed above with respect to claim 3.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Brent Hoover whose telephone number is (303)297-4403. The examiner can normally be reached Monday - Friday 9-5 MST.
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, Abdullah Kawsar can be reached at 571-270-3169. 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.
/BRENT JOHNSTON HOOVER/Primary Examiner, Art Unit 2127
Prosecution Insights