DETAILED ACTION
Notice of 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 .
Priority
Regarding Chinese Patent App. No. 202210841172.6 (filed 7/18/2022), receipt is acknowledged of certified copies of papers required by 37 CFR 1.55.
Specification / Abstract / Drawings
The disclosure is objected to because of the following informalities:
The examiner suggests updating all paragraphs in the specification, Abstract, and in the drawings (see S102 and S203) to correct the misspelling of “Monte Carlo”.
Appropriate correction is required.
Claim Objections
Claim 1-2, 4, 6-11, 13, 15-16, 18, and 20 are objected to because of the following informalities:
In claims 1-2, 4, 6-11, 13, 15-16, 18, and 20, the limitation “Monto-Carlo Tree-Search algorithm” should be corrected to fix the misspelling so that it recites “Monte-Carlo Tree-Search algorithm”
In claim 4, line 7, “searching, by the optimized modularized parametric model, a program expression” should read “searching for, by the optimized modularized parametric model, a program expression”
In claim 8, line 5, “searching, by the hierarchical Monto-Carlo Tree-Search algorithm, appropriate” should read “searching for, by the hierarchical Monte-Carlo Tree-Search algorithm, appropriate”
In claim 13, line 7, “searching, by the optimized modularized parametric model, a program expression” should read “searching for, by the optimized modularized parametric model, a program expression”
In claim 18, line 7, “searching, by the optimized modularized parametric model, a program expression” should read “searching for, by the optimized modularized parametric model, a program expression”
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.
Claims 4, 13, and 18 are rejected under 35 U.S.C. 112(b) as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor regards as the invention.
Claim 4, last 2 lines, recite: “integrating, by the optimized modularized parametric model, the target program expression in the result program expression and outputting.” However, it is unclear what is actually “output”. The output could be the “result program expression”, the “optimized modularized parametric model”, or other output data that corresponds to the “inputting data to be processed” as recited in claim 2. MPEP 2173.02 I explains that “if the language of a claim, given its broadest reasonable interpretation, is such that a person of ordinary skill in the relevant art would read it with more than one reasonable interpretation, then a rejection under 35 U.S.C. 112(b) or pre-AIA 35 U.S.C. 112, second paragraph is appropriate.” Because there is more than one reasonable interpretation of what would be “output” the claim is indefinite.
Claims 13 and 18 are dependent claims that are substantially similar to claim 4, and are therefore rejected for the same reasons.
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 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding Step 1 of the Alice/Mayo framework, Claims 1-8 are directed to a method (a process), Claims 9 and 11-15 are directed to an electronic device (a machine), which each fall within one of the four statutory categories of inventions.
However, Claims 10 and 16-20 are directed to a computer-readable storage medium (an article of manufacture). Because the broadest reasonable interpretation of “computer-readable storage medium” does not exclude transitory forms of signal transmission (often referred to as “signals per se”), claims 10 and 16-20 are not directed to an eligible statutory category and are therefore rejected under 35 U.S.C. 101. The examiner recommends amending such claims to recite “a non-transitory computer-readable storage medium” to overcome the rejections.
Regarding Claim 1
Step 2A, prong 1 (Is the claim directed to a law of nature, a natural phenomenon or an abstract idea).
Claim 1 recites the following mental processes, that in each case under the broadest reasonable interpretation, covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper.
A modularized parametric visual program induction method, comprising: (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) a set of rules that each have input/output parameters, for deriving a set of rules to draw a visual representation, such as rules for drawing circles, lines, triangles, and rectangles having input parameters for lengths, angles, and line thicknesses, where the rules can be implemented in a sequence to draw a picture)
constructing a modularized parametric model, wherein the modularized parametric model comprises a plurality of parametric submodels, different parametric submodels have different types of parameters, and the different types of parameters are configured for describing a plurality of attributes of data; (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) a set of rules that each have input/output parameters (each a different “parametric submodel”), for deriving a set of rules (the set of rules being the ”modularized parametric model, where each rule is its own module) to draw a visual representation, such as rules for drawing circles, lines, triangles, and rectangles having input parameters for lengths, angles, and line thicknesses, where the rules can be implemented in a sequence to draw a picture, and where the parameters for a circle (e.g., radius) are different than parameters for a triangle (e.g., angles and side lengths), where the different parameters describe different attributes of data (radius vs. angles/side lengths))
generating an augmented training data set based on a hierarchical Monto-Carlo Tree- Search algorithm and a basic training data set; and (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) a basic training data set including 2 or more drawings and the set of rules/parameters to achieve said drawings, and then perform a hierarchical Monto-Carlo Tree-Search algorithm, which is a mathematical concept, to make augmentations to the set of 2 or more drawings and then explore paths amongst such augmentations using the MCTS algorithm)
training and optimizing the modularized parametric model based on the augmented training data set to obtain an optimized modularized parametric model, wherein the optimized modularized parametric model comprises optimized parametric submodels. (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) train and optimize the set of rules for drawing the geometric shapes and lines, for example, by creating more drawings, the individual rules for drawing shapes can be optimized and improved (for example, to place limits to keep the drawings within a bounding box), such that each individual rule is optimized)
Regarding Step 2A, Prong 2, the claim does not include any additional elements that integrate the judicial exception into a practical application and regarding Step 2B, there are no additional elements recited that amount to significantly more than the judicial exception.
Regarding Claim 2
Step 2A, Prong 1
inputting data to be processed into the optimized modularized parametric model; and (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) input a simple geometric drawing to be processed by the set of rules, such that the output of the model is a series of steps (and associated parameters) for drawing the input geometric drawing)
processing the data to be processed by the optimized modularized parametric model in conjunction with the hierarchical Monto-Carlo Tree-Search algorithm, and (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) input a simple geometric drawing to be processed by the set of rules, such that the output of the model is a series of steps (and associated parameters) derived using the mathematical concept of the Monto-Carlo Tree-search algorithm, for drawing the input geometric drawing)
outputting a result program expression. (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) output the series of steps (and associated parameters) derived using the mathematical concept of the Monto-Carlo Tree-search algorithm, for drawing the input geometric drawing, where such rules are written in terms of a program expression, e.g., (circle(1, 2) for a circle of radius 1 and line thickness 2))
Regarding Step 2A, Prong 2, the claim does not include any additional elements that integrate the judicial exception into a practical application and regarding Step 2B, there are no additional elements recited that amount to significantly more than the judicial exception.
Regarding Claim 3
Step 2A, Prong 1
constructing the plurality of parametric submodels based on different types of meta- functions of a target domain, ...; and (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) a set of rules-based on meta-functions, for example, the rules for drawing shapes can be based on the basic geometric shapes of circles, lines, triangles, rectangles (each a meta-function) in the target domain of 2-dimensional x-y coordinate systems)
combining the modularized parametric model with the plurality of parametric submodels corresponding to each of the different types of parameters, wherein the modularized parametric model is a set of the plurality of parametric submodels for the target domain. (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) can combine each of the rules for geometric shapes into a combined model, where each rule is its own model and has its own different type of parameters (e.g., circle has a radius parameter, triangle has parameters for angles and side lengths))
Step 2A, Prong 2
Regarding the “wherein the plurality of parametric submodels are defined as sub-neural networks corresponding to the different types of parameters” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception. In particular, the claim only recites the additional element of a sub-neural network. This additional element is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (a sub-neural network). Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Step 2B
Regarding the “wherein the plurality of parametric submodels are defined as sub-neural networks corresponding to the different types of parameters” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, because the limitation merely provides instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not add significantly more than the judicial exception. (See MPEP 2106.05(f)).
Regarding Claim 4
Step 2A, Prong 1
performing a program expression on the data to be processed based on the different types of optimized parametric submodels in the optimized modularized parametric model; (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) have each rule be represented by a pseudocode program expression (e.g., circle(1, 2) for a circle of radius 1 with line thickness 2)), and then implement such pseudocode program expression by drawing the shape on a piece of paper)
searching, by the optimized modularized parametric model, a program expression in conformity with the data to be processed in the program expression based on the hierarchical Monto-Carlo Tree-Search algorithm as a target program expression; and (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) use the Monto-Carlo Tree-Search algorithm, which is a mental concept, for a particular program expression (e.g., circle(1,2)) that matches input data to be processed (e.g., a drawing having a circle) to explore different permutations for the circle in order to match the input drawings (e.g., a tree having parameters for radius and line thicknesses, and searching the tree using Monto-Carlo search for the rule that matches the desired circle in the input drawing)
integrating, by the optimized modularized parametric model, the target program expression into the result program expression and outputting. (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) the series of rules (and associated parameters) that would result in drawing the equivalent of the input drawing, and then outputting such rules by writing them on paper)
Regarding Step 2A, Prong 2, the claim does not include any additional elements that integrate the judicial exception into a practical application and regarding Step 2B, there are no additional elements recited that amount to significantly more than the judicial exception.
Regarding Claim 5
Step 2A, Prong 1
PNG
media_image1.png
312
592
media_image1.png
Greyscale
(under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) define the input and output parameters for a neural network based on a set of mathematical concepts and relationships as set forth in this limitation)
Step 2A, Prong 2
Regarding the “wherein the sub-neural network is” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception. In particular, the claim only recites the additional element of a sub-neural network. This additional element is recited at a high-level of generality and amounts to no more than mere instructions to apply the exception using a generic computer component (a sub-neural network). Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Step 2B
Regarding the “wherein the sub-neural network is” limitation, such limitation is recited at a high-level of generality and amounts to no more than adding the words “apply it” (or an equivalent) with the judicial exception, because the limitation merely provides instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not add significantly more than the judicial exception. (See MPEP 2106.05(f)).
Regarding Claim 6
Step 2A, Prong 1
wherein after constructing the modularized parametric model, the operation of generating the augmented training data set based on the hierarchical Monto-Carlo Tree-Search algorithm and the basic training data set comprises: (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) contrast the modularized parametric model (set of finished rules) after generating the augmented training dataset as recited in claim 1)
providing, by the plurality of parametric submodels, a search guidance for the hierarchical Monto-Carlo Tree-Search algorithm based on the different types of parameters; (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) provide search guidance for the Monto-Carlo Tree-Search algorithm, such as in the example of searching for a matching circle rule to match an input circle, guiding the algorithm to only search for a circle radius between 0.5 and 2 and a line thickness between 1-10)
augmenting, by the hierarchical Monto-Carlo Tree-Search algorithm, the basic training data set based on the search guidance to obtain the augmented training data set; and (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) a basic training data set including 2 or more drawings and the set of rules/parameters to achieve said drawings, and then perform a hierarchical Monto-Carlo Tree-Search algorithm, which is a mathematical concept, to make augmentations to the set of 2 or more drawings (using the search guidance of only searching for a circle radius between 0.5 and 2 and a line thickness between 1-10) and then explore paths amongst such augmentations using the MCTS algorithm)
augmenting, by the hierarchical Monto-Carlo Tree-Search algorithm, a combination mode of the plurality of parametric submodels to obtain an augmented parametric submodel combination. (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) can change a combination mode (e.g., combination by addition + scaling) of the plurality of rules to obtain an augmented set of rules that creates the input drawing)
Regarding Claim 7
Step 2A, Prong 1
PNG
media_image2.png
466
608
media_image2.png
Greyscale
(under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) define the search space for a Monto-Carlo Tree-Search algorithm using the mathematical concepts and relationships recited in this claim)
Regarding Step 2A, Prong 2, the claim does not include any additional elements that integrate the judicial exception into a practical application and regarding Step 2B, there are no additional elements recited that amount to significantly more than the judicial exception.
Regarding Claim 8
Step 2A, Prong 1
wherein after generating the augmented training data set, the operation of training and optimizing the modularized parametric model based on the augmented training data set to obtain the optimized modularized parametric model comprises: (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) perform training and optimizing after generating the augmented training data set as recited in claim 1)
searching, by the hierarchical Monto-Carlo Tree-Search algorithm, appropriate augmented parametric submodel combinations in turn based on the search guidance; (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) use the search guidance (of claim 6), using the mathematical concept of Monto-Carlo Tree-Search algorithm, to derive rule combinations)
inputting the augmented training data set into the appropriate augmented parametric submodel combinations for modeling, and calculating a loss function value; (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) the resulting drawing from the combination of rules created by the model, and then compare the input drawing to the output drawing with respect to the mathematical concept of using a loss function to determine the amount of difference between the input and output drawings)
optimizing the different types of parameters corresponding to the plurality of parametric submodels in the appropriate augmented parametric submodel combinations based on the loss function value to obtain the optimized parametric submodels; and (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) train and optimize the set of rules for drawing the geometric shapes and lines, for example, by creating more drawings, the individual rules for drawing shapes can be optimized and improved (for example, to place limits to keep the drawings within a bounding box), such that each individual rule is optimized by minimizing the amount of error from the loss function value)
combining the optimized parametric submodels into the optimized modularized parametric model. (under the broadest reasonable interpretation, a human can mentally (or using pencil and paper) can perform this limitation, for example, a human can mentally (or write down on paper) can combine each of the rules for geometric shapes into a combined model, where each rule is its own model and has its own different type of parameters (e.g., circle has a radius parameter, triangle has parameters for angles and side lengths))
Regarding Step 2A, Prong 2, the claim does not include any additional elements that integrate the judicial exception into a practical application and regarding Step 2B, there are no additional elements recited that amount to significantly more than the judicial exception.
Regarding Claim 9
Step 2A, Prong 1
Claim 9 recites an electronic device that corresponds to the method of claim 1, and therefore the analysis under Step 2A, Prong 1 with respect to claim 1 also applies to this claim 9. While claim 9 recites additional generic computing components (“memory”, “processor”, “computer program”), such additional generic computing components do not change the analysis under Step 2A, Prong 1.
Step 2A, Prong 2
Claim 9 recites an electronic device that corresponds to the method of claim 1, and therefore the analysis under Step 2A, Prong 2 with respect to claim 1 also applies to this claim 9. While claim 9 recites additional generic computing components (“memory”, “processor”, “computer program”), such additional generic computing components do not change the analysis under Step 2A, Prong 2 because such limitations are recited at a high-level of generality and amount to no more than adding the words “apply it” (or an equivalent) with the judicial exception. In particular, the claim only recites the additional elements (“memory”, “processor”, “computer program”). These additional elements are recited at a high-level of generality and amount to no more than mere instructions to apply the exception using generic computer components. Accordingly, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Step 2B
Claim 9 recites an electronic device that corresponds to the method of claim 1, and therefore the analysis under Step 2B with respect to claim 1 also applies to this claim 9. While claim 9 recites additional generic computing components (“memory”, “processor”, “computer program”), such additional generic computing components do not change the analysis under Step 2B because such limitations are recited at a high-level of generality and amount to no more than adding the words “apply it” (or an equivalent) with the judicial exception, because the limitations merely provides instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not add significantly more than the judicial exception. (See MPEP 2106.05(f)).
Regarding Claim 10
Step 2A, Prong 1
Claim 10 recites a computer-readable storage medium that corresponds to the method of claim 1, and therefore the analysis under Step 2A, Prong 1 with respect to claim 1 also applies to this claim 10. While claim 10 recites additional generic computing components (“computer-readable storage medium”, “processor”, “computer program/instruction”), such additional generic computing components do not change the analysis under Step 2A, Prong 1.
Step 2A, Prong 2
Claim 10 recites a computer-readable storage medium that corresponds to the method of claim 1, and therefore the analysis under Step 2A, Prong 2 with respect to claim 1 also applies to this claim 10. While claim 10 recites additional generic computing components (“computer-readable storage medium”, “processor”, “computer program/instruction”), such additional generic computing components do not change the analysis under Step 2A, Prong 2 because such limitations are recited at a high-level of generality and amount to no more than adding the words “apply it” (or an equivalent) with the judicial exception. In particular, the claim only recites the additional elements (“computer-readable storage medium”, “processor”, “computer program/instruction”). These additional elements are recited at a high-level of generality and amount to no more than mere instructions to apply the exception using generic computer components. Accordingly, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea (See MPEP 2106.05(f)).
Step 2B
Claim 10 recites a computer-readable storage medium that corresponds to the method of claim 1, and therefore the analysis under Step 2B with respect to claim 1 also applies to this claim 10. While claim 10 recites additional generic computing components (“computer-readable storage medium”, “processor”, “computer program/instruction”), such additional generic computing components do not change the analysis under Step 2B because such limitations are recited at a high-level of generality and amount to no more than adding the words “apply it” (or an equivalent) with the judicial exception, because the limitations merely provides instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. Accordingly, this additional element does not add significantly more than the judicial exception. (See MPEP 2106.05(f)).
Regarding Claims 11-15
Claims 11-15 depend from claim 9 and claim an electronic device that corresponds to the methods of claims 2-6, respectively, and are therefore rejected for the same reasons explained above with respect to claims 9 and 2-6, respectively.
Regarding Claims 16-20
Claims 16-20 depend from claim 10 and claim a computer-readable storage medium that corresponds to the methods of claims 2-6, respectively, and are therefore rejected for the same reasons explained above with respect to claims 10 and 2-6, respectively.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-20 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Duan, Xuguang, et al. "Parametric visual program induction with function modularization." International Conference on Machine Learning. (published July 12, 2022), hereinafter referenced as DUAN.
First, the examiner notes that the DUAN reference was available at least as early as July 12, 2022 at https://proceedings.mlr.press/v162/ as shown by the provided archive.org printout. Therefore, the DUAN reference was publicly available at least as early as July 12, 2022, which is prior to the 7/18/2022 effective filing date of the present application.
Second, the examiner notes that DUAN has 4 co-authors each affiliated with Tsinghua University, Beijing, China, and that 3 of such co-authors appear to be named inventors of the present application. However, the 4th co-author, Ziwei Zhang, has not been identified as a named inventor.
MPEP 2153.01(a) recites: “If, however, the application names fewer joint inventors than a publication (e.g., the application names as joint inventors A and B, and the publication names as authors A, B and C), it would not be readily apparent from the publication that it is an inventor-originated disclosure and the publication would be treated as prior art under AIA 35 U.S.C. 102(a)(1) unless there is evidence of record that an exception under AIA 35 U.S.C. 102(b)(1) applies.”
In the present case, the inclusion of Ziwei Zhang as co-authors of the DUAN reference means that it is not “readily apparent from the publication that it is an inventor-originated disclosure.”
The examiner suggests that Applicant review MPEP 2155 regarding the use of affidavits or declarations under 37 CFR 1.130 to overcome prior art rejections.
Regarding Claim 1
DUAN teaches:
A modularized parametric visual program induction method, comprising: (DUAN, p. 1, section 1: “In this paper, we are the first to propose the concept of Parametric Visual Program Induction, i.e., generating programs with parametric primitive functions for complex visual observations, to the best of our knowledge.”;
DUAN, p. 2, section 1: “To the best of our knowledge, we are the first to investigate the problem of parametric visual program induction by proposing the concept and method of Function Modularization, which decouples the learning of function parameters and function transitions, resulting in accurate and efficient learning of the parametric programs”)
constructing a modularized parametric model, wherein the modularized parametric model comprises a plurality of parametric submodels, different parametric submodels have different types of parameters, and the different types of parameters are configured for describing a plurality of attributes of data; (DUAN, p. 3, section 4: “4. Function Modularization. In this section, we tackle the problem of learning parametric functions by function modularization. Specifically, we transform the parametric program induction problem as learning inter-function transition and intra-function parameter prediction. The former focuses on selecting which function should be used, and the latter models each function along with its parameters as a self-contained module to obtain the most suitable parameters for that function.”;
Examiner’s Note: each “function” corresponds to a recited “submodel” for the overall program, where each “function” is modularized to have suitable parameters for such function)
generating an augmented training data set based on a hierarchical Monto-Carlo Tree- Search algorithm and a basic training data set; and (DUAN, p. 3, section 3.2: “To train P in Eq. (3), synthetic data is generated and utilized (Shin et al., 2019). Specifically, a random program P is first generated.”;
DUAN, p. 4, section 5: “In this section, we propose the Hierarchical-Heterogeneous Monto-Carlo tree search (H2MCTS) algorithm in conjunction with the modularized function. Compared to using the naive synthetic data introduced in Section 3.2, our proposed H2MCTS can provide high-quality uncorrelated training data during training, and serves as an efficient search technique during inference.”;
DUAN, p. 5, section 5.2: “In this subsection, we introduce how to transform the synthetic data as introduced in Section 3.2 using H2MCTS and adopt the transformed data to more effectively train the model.”
Examiner’s Note: the synthetic data introduced in section 3.2 corresponds to the recited “basic training data set” and the H2MCTS algorithm is used to provide augmented training data corresponding to the recited “augmented training data”)
training and optimizing the modularized parametric model based on the augmented training data set to obtain an optimized modularized parametric model, wherein the optimized modularized parametric model comprises optimized parametric submodels. (DUAN, p. 5, section 5.2:
PNG
media_image3.png
412
388
media_image3.png
Greyscale
Examiner’s Note: DUAN teaches that during training using the H2MCTS-derived training data, equation (12) is optimized for each selected function f (which corresponds to the recited “optimized parametric submodels”) and the combination of optimized functions corresponds to the recited “modularized parametric model”)
Regarding Claim 2
DUAN teaches the method of claim 1 as explained above. DUAN further teaches:
inputting data to be processed into the optimized modularized parametric model; and (DUAN, p. 3.1:
PNG
media_image4.png
96
388
media_image4.png
Greyscale
DUAN, p. 8, section 6.2, Figure 7:
PNG
media_image5.png
374
770
media_image5.png
Greyscale
Examiner’s Note: as shown in Fig. 7, the input to the H2MCTS function modularization is the drawing in the top left corner)
processing the data to be processed by the optimized modularized parametric model in conjunction with the hierarchical Monto-Carlo Tree-Search algorithm, and outputting a result program expression. (DUAN, p. 5, section 5.3:
PNG
media_image6.png
168
384
media_image6.png
Greyscale
DUAN, p. 8, section 6.2, Figure 7:
PNG
media_image5.png
374
770
media_image5.png
Greyscale
Examiner’s Note: as shown in Fig. 7, the input to the H2MCTS function modularization is the drawing in the top left corner, and the output is the returned program of function calls with parameters (corresponding to recited “outputting a result program expression”)
Regarding Claim 3
DUAN teaches the method of claim 1 as explained above. DUAN further teaches:
constructing the plurality of parametric submodels based on different types of meta- functions of a target domain, (DUAN, p. 7, section 6.2: “This dataset contains 3 primitive functions: Circle, Line, and Rectangle, each of which draws on a discrete 16 × 16 grid coordinates. The synthesized data contains randomly generated programs, while the real hand-drawn images aim to show certain structures. Figure 5 shows some examples of the dataset.”
DUAN, p. 8, section 6.2, Figure 7:
PNG
media_image5.png
374
770
media_image5.png
Greyscale
Examiner’s Note: as shown in Fig. 7, each of the modules (in green) for a function corresponds to the recited “plurality of parameter submodels” and the primitive Circle, Line, and Rectangle primitives correspond to the recited “meta-functions” of the target domain of 2D Latex Drawing)
wherein the plurality of parametric submodels are defined as sub-neural networks corresponding to the different types of parameters; and (DUAN, p. 4, section 4.2:
PNG
media_image7.png
280
378
media_image7.png
Greyscale
DUAN, p. 8, section 6.2: “First, our proposed multi-head self-contained neural module is more flexible to handle different parameters adaptively.”;
Examiner’s Note: DUAN explains that a separate neural network module is utilized to predict a function, where there are individual heads (corresponding to recited “sub-neural networks”) for each parameter)
combining the modularized parametric model with the plurality of parametric submodels corresponding to each of the different types of parameters, wherein the modularized parametric model is a set of the plurality of parametric submodels for the target domain. (DUAN, p. 4, section 4.2:
PNG
media_image7.png
280
378
media_image7.png
Greyscale
Examiner’s Note: DUAN, equation 10, shows the combination of sub-functions Qf,j to form Qf)
Regarding Claim 4
DUAN teaches the method of claim 2 as explained above, including the “wherein the operation of processing the data to be processed by the optimized modularized parametric model in conjunction with the hierarchical Monto-Carlo Tree-Search algorithm, and outputting the result program expression comprises:” limitation. DUAN further teaches:
performing a program expression on the data to be processed based on the different types of optimized parametric submodels in the optimized modularized parametric model;
(DUAN, p. 8, section 6.2, Figure 7:
PNG
media_image5.png
374
770
media_image5.png
Greyscale
Examiner’s Note: as shown in Fig. 7, each of the modules (in green) executes a function with respect to the input data using different submodels (sub-functions for rectangle, line, circle))
searching, by the optimized modularized parametric model, a program expression in conformity with the data to be processed in the program expression based on the hierarchical Monto-Carlo Tree-Search algorithm as a target program expression; and
(DUAN, p. 8, section 6.2, Figure 7:
PNG
media_image5.png
374
770
media_image5.png
Greyscale
Examiner’s Note: as shown in Fig. 7, the H2MCTS algorithm searches for different program expressions that can be used in the resulting and returned program expression)
integrating, by the optimized modularized parametric model, the target program expression into the result program expression and outputting.
(DUAN, p. 8, section 6.2, Figure 7:
PNG
media_image5.png
374
770
media_image5.png
Greyscale
Examiner’s Note: as shown in Fig. 7, the returned program of 14 primitive functions corresponds to the recited “target program expression”)
Regarding Claim 5
DUAN teaches the method of claim 3 as explained above. DUAN further teaches:
PNG
media_image1.png
312
592
media_image1.png
Greyscale
(DUAN, p. 3, section 4.1:
PNG
media_image8.png
270
380
media_image8.png
Greyscale
DUAN, p. 4, section 4.2:
PNG
media_image7.png
280
378
media_image7.png
Greyscale
DUAN, p. 5, sections 5.2-5.3:
PNG
media_image9.png
284
386
media_image9.png
Greyscale
Examiner’s Note: DUAN shows the same equations and variables using slightly different notation. Qf for a particular parameter (θ) corresponds to recited “
PNG
media_image10.png
30
138
media_image10.png
Greyscale
” as shown by Eq. (5) of DUAN. Section 4.2 explains that each Qf pertains to a sub neural network and the recited “set of the plurality of parametric submodels” is disclosed by the equation in the first line of section 5.3)
Regarding Claim 6
DUAN teaches the method of claim 3 as explained above, including the “wherein after constructing the modularized parametric model, the operation of generating the augmented training data set based on the hierarchical Monto-Carlo Tree-Search algorithm and the basic training data set comprises” limitation. DUAN further teaches:
providing, by the plurality of parametric submodels, a search guidance for the hierarchical Monto-Carlo Tree-Search algorithm based on the different types of parameters; (DUAN, p. 5, section 5.2:
PNG
media_image11.png
282
392
media_image11.png
Greyscale
Examiner’s Note: Section 5.2 explains how P is used to guide H2MCTS based on different parameters predicted by Qf)
augmenting, by the hierarchical Monto-Carlo Tree-Search algorithm, the basic training data set based on the search guidance to obtain the augmented training data set; and (DUAN, p. 3, section 3.2: “To train P in Eq. (3), synthetic data is generated and utilized (Shin et al., 2019). Specifically, a random program P is first generated.”;
DUAN, p. 4, section 5: “In this section, we propose the Hierarchical-Heterogeneous Monto-Carlo tree search (H2MCTS) algorithm in conjunction with the modularized function. Compared to using the naive synthetic data introduced in Section 3.2, our proposed H2MCTS can provide high-quality uncorrelated training data during training, and serves as an efficient search technique during inference.”;
DUAN, p. 5, section 5.2: “In this subsection, we introduce how to transform the synthetic data as introduced in Section 3.2 using H2MCTS and adopt the transformed data to more effectively train the model.”
Examiner’s Note: the synthetic data introduced in section 3.2 corresponds to the recited “basic training data set” and the H2MCTS algorithm is used to provide augmented training data corresponding to the recited “augmented training data”)
augmenting, by the hierarchical Monto-Carlo Tree-Search algorithm, a combination mode of the plurality of parametric submodels to obtain an augmented parametric submodel combination. (DUAN, p. 4, section 4.2:
PNG
media_image7.png
280
378
media_image7.png
Greyscale
Examiner’s Note: DUAN, equation 10, shows the combination of sub-functions Qf,j to form Qf)
Regarding Claim 7
DUAN teaches the method of claim 6 as explained above. DUAN further teaches:
PNG
media_image2.png
466
608
media_image2.png
Greyscale
(DUAN, p. 3, section 4.1:
PNG
media_image12.png
376
386
media_image12.png
Greyscale
DUAN, p. 4, section 4.1:
PNG
media_image13.png
180
380
media_image13.png
Greyscale
DUAN, p. 4, section 4.1:
PNG
media_image14.png
154
366
media_image14.png
Greyscale
Examiner’s Note: the 3 specific equations recited in claim 7 correspond to equations (4), (6), and (7) of DUAN, respectively)
Regarding Claim 8
DUAN teaches the method of claim 6 as explained above, including the “operation of training and optimizing the modularized parametric model based on the augmented training data set to obtain the optimized modularized parametric model comprises:” limitation. DUAN further teaches:
searching, by the hierarchical Monto-Carlo Tree-Search algorithm, appropriate augmented parametric submodel combinations in turn based on the search guidance; (DUAN, p. 5, section 5.2:
PNG
media_image11.png
282
392
media_image11.png
Greyscale
Examiner’s Note: Section 5.2 explains how P is used to guide H2MCTS based on different parameters predicted by Qf)
inputting the augmented training data set into the appropriate augmented parametric submodel combinations for modeling, (DUAN, p. 3.1:
PNG
media_image4.png
96
388
media_image4.png
Greyscale
DUAN, p. 8, section 6.2, Figure 7:
PNG
media_image5.png
374
770
media_image5.png
Greyscale
Examiner’s Note: as shown in Fig. 7, the input to the H2MCTS function modularization is the drawing in the top left corner)
and calculating a loss function value; (DUAN, p. 7, section 6.2: “We compare three versions where the first two use Cross-Entropy (CE) loss, Reinforcement Learning (RL) loss during training respectively”)
optimizing the different types of parameters corresponding to the plurality of parametric submodels in the appropriate augmented parametric submodel combinations based on the loss function value to obtain the optimized parametric submodels; and (DUAN, p. 5, section 5.2:
PNG
media_image3.png
412
388
media_image3.png
Greyscale
Examiner’s Note: DUAN teaches that during training using the H2MCTS-derived training data, equation (12) is optimized for each selected function f (which corresponds to the recited “optimized parametric submodels”) and the combination of optimized functions corresponds to the recited “modularized parametric model”)
combining the optimized parametric submodels into the optimized modularized parametric model. (DUAN, p. 4, section 4.2:
PNG
media_image7.png
280
378
media_image7.png
Greyscale
Examiner’s Note: DUAN, equation 10, shows the combination of sub-functions Qf,j to form Qf)
Regarding Claim 9
DUAN teaches:
An electronic device, comprising a memory, a processor, and a computer program stored in the memory, wherein the processor, when executing the computer program, implements operations ..., wherein the operations comprise: (DUAN, p. 12, Appendix A.2: “The whole framework is launched on a GPU server with two Intel(R) Xeon(R) Gold 6240 CPU @ 2.60GHz CPU processors and two Nvidia GeForce RTX 3090 GPU processors.”
Examiner’s Note: the CPU and GPU processors necessarily have on-board cache for storing instructions that are processed by the processor, where such cache corresponds to the recited “memory” that the instructions are stored on)
The remaining limitations correspond to the method of claim 1 and are rejected for the same reasons explained above with respect to claim 1.
Regarding Claim 10
DUAN teaches:
A computer-readable storage medium storing a computer program/instruction, wherein the computer program/instruction, when executed by a processor, implements operations ..., wherein the operations comprise: (DUAN, p. 12, Appendix A.2: “The whole framework is launched on a GPU server with two Intel(R) Xeon(R) Gold 6240 CPU @ 2.60GHz CPU processors and two Nvidia GeForce RTX 3090 GPU processors.”
Examiner’s Note: the CPU and GPU processors necessarily have on-board cache for storing instructions that are processed by the processor, where such cache corresponds to the recited “computer-readable storage medium” that the instructions are stored on)
The remaining limitations correspond to the method of claim 1 and are rejected for the same reasons explained above with respect to claim 1.
Regarding Claims 11-15
Claims 11-15 depend from claim 9 and claim an electronic device that corresponds to the methods of claims 2-6, respectively, and are therefore rejected for the same reasons explained above with respect to claims 9 and 2-6, respectively.
Regarding Claims 16-20
Claims 16-20 depend from claim 10 and claim a computer-readable storage medium that corresponds to the methods of claims 2-6, respectively, and are therefore rejected for the same reasons explained above with respect to claims 10 and 2-6, respectively.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Lim, Jinsuk, et al. "Field report: Applying monte carlo tree search for program synthesis." International symposium on search based software engineering. Cham: Springer International Publishing, 2016. Applies the Monte Carlo Tree Search algorithm for general purpose program synthesis, and particularly to synthesize Java Bytecode instructions.” (section 1).
Liu, Larkin, et al. "An extensible and modular design and implementation of monte carlo tree search for the JVM." arXiv preprint arXiv:2108.10061 (7/30/2021). Applies the Monte Carlo tree search during program synthesis to generate software compatible with the Java Virtual Machine. (section 2). Page 10, Fig. 3, shows an embodiment where a software program visually depicts a game.
Islam, Mohiul, et al. "Mutation operators for genetic programming using Monte Carlo tree search." Applied Soft Computing 97 (2020): 106717. Discloses using a Monte Carlo tree search during inductive program synthesis. (section 1).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MICHAEL C LEE whose telephone number is (571)272-4933. The examiner can normally be reached M-F 12:00 pm - 8:00 pm ET.
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, Omar Fernandez Rivas can be reached at 571-272-2589. 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.
/MICHAEL C. LEE/Examiner, Art Unit 2128
Prosecution Insights