DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Remarks
This action is in response to the applicant’s response filed 13 January 2026, which is in response to the USPTO office action mailed 27 October 2025. Claims 1-8 and 13-19 are amended. Claims 1-20 are currently pending. Claims 11, 12 and 20 are withdrawn.
Response to Arguments
With respect to the 35 USC §102 rejection of claims 1-8 and 13-19, the applicant’s arguments are moot in view of a new grounds of rejection, as necessitated by the applicant's amendments.
However, still applicable to the current rejection, the applicant argues “Wu does not involve dividing V into different subsets, nor does it involve identifying certain nodes from the complete set of nodes V. And although Wu mentions cosine similarity, radial basis function (RBF) kernel, and attention mechanisms, Wu aims to determine whether to remove redundant edges by judging whether the similarity Sij between two nodes reaches a threshold (see formula 3 in Paragraph [0078] of Wu). Thus, Wu does not involve the removal or filtering of partial nodes.” (Remarks, pg. 11). Respectfully, this argument is not persuasive.
In response to the applicant's argument that the references fail to show certain features of the invention, it is noted that the features upon which applicant relies (i.e., dividing into different subsets and removing or filtering partial nodes) are not recited in the rejected claim(s). Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993).
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 1-10 and 13-19 are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
Claims 1 and 13 recite “wherein the first nodes are some of the second nodes” (e.g. claim 1 lines 3-4). The examiner has reviewed the specification, but was unable to find support for this limitation. In particular, the specification discloses “determining first node information based on edge data and feature information corresponding to first graph data; where the first node information corresponds to K first nodes selected from P second nodes; K is an integer greater than 1 and less than P” in [0046].
The term “some” is defined as an unspecified number or amount of, which may include all. In other words, the claims as currently drafted encompass the first nodes being all of the second nodes. However, the specification appears to disclose the first nodes are an integer number greater than 1 and less than a number of second nodes. Therefore, the claims lack written description support from the specification.
Note, the dependent claims are also rejected because they do not remedy the deficiencies inherited by their parent claims.
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 1-10 and 13-19 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.
Claims 1 and 13 recite the limitation "determining a prediction result based on the aggregated feature information of the second nodes, wherein the prediction result comprises a predicted category label of the labeled node among the second nodes " (e.g. claim 1 lines 10-12). It is unclear as to which second nodes are being referred to as at least two second nodes were introduced earlier in the claim (e.g. “second nodes” (see claim 1 line 3) as well as “each second node” (see claim 1 line 5)). Therefore, the claim is rendered indefinite because the examiner is unable to ascertain the scope of the claims.
Note, the dependent claims are also rejected because they do not remedy the deficiencies inherited by their parent claim. Appropriate action is required.
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 1-8 and 13-19 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Wu et al., US 2021/0374499 A1 (hereinafter “Wu”).
Claim 1: Wu teaches a data processing method, the method comprises:
determining node information of first nodes based on edge data and feature information of second nodes of graph data, wherein the first nodes are some of the second nodes (Wu, [0066] note Let the graph G=(V, E)ϵ 𝓖 be represented as a set of n nodes viϵV with an initial node feature matrix XϵRd×n, edges (vi, vj)ϵE (binary or weighted) formulating an initial noisy adjacency matrix A(0)ϵRn×n);
calculating feature similarity between each second node and a neighborhood node of each second node based on the node information of the first node (Wu, [0061] note a graph learning neural network uses multi-head self-attention with epsilon-neighborhood sparsification for constructing a graph, [0078] note A symmetric sparse adjacency matrix A is extracted from S by considering only the E-neighborhood for each node; specifically, those elements in S which are smaller than a non-negative threshold ε, [0074] note common options for metrics include cosine similarity, radial basis function (RBF) kernel, and attention mechanisms);
aggregating feature information of each second node based on the feature similarity and feature information of the neighborhood node, wherein the neighborhood node is a first node in a neighborhood of each second node (Wu, [0097] note at the t-th iteration, A(t) is computed based on Z(t-1) (Line 7), and Z(t) is computed based on Ã(t) (Line 9) which is computed based on A(t) (Eq. (4)). The difference between the adjacency matrices at the t-th iteration and the previous iteration is denoted by δA(t));
determining a prediction result based on the aggregated feature information of the second nodes, wherein the prediction result comprises a predicted category label of the labeled node among the second nodes (Wu, [0082] note à is the normalized adjacency matrix, σ(⋅) is a task dependent output function, and ℓ (⋅) is a task-dependent loss function. For instance, for node classification, σ(⋅) is a Softmax function for predicting a probability distribution over a set of classes, [0102] note The input features are bag of words and the task is node classification); and
iteratively updating a parameter of a classification model based on the predicted category label of the labeled node and a real category label of the labeled node (Wu, [0095] note After all iterations, the overall loss is back-propagated through all previous iterations to update the model parameters).
Claim 2: Wu teaches the method according to claim 1, wherein the determining the node information of the first node based on the edge data and the feature information of the second node of the graph data comprises:
determining a first score of each second node based on the edge data between the second node and the neighborhood node of each second node; determining feature information of third nodes based on first scores and the feature information of the second nodes; wherein the third nodes are selected from the second nodes; determining a second score of each third node based on the feature information of each third node; and determining the node information of the first node based on second scores (Wu, [0078] note A is extracted from S by considering only the E-neighborhood for each node; specifically, those elements in S which are smaller than a non-negative threshold ε… where L(0) is the normalized adjacency matrix of the initial graph, defined as, L(0)=D(0)-1/2A(0)D(0)-1/2, and D(0) is its degree matrix. A(t) and A(1) are the two adjacency matrices learned at the t-th and 1-st iterations using Eq. (2) and Eq. (3), respectively. The adjacency matrix learned by Eqs. (2) and (3) is row normalized such that each row sums to 1, so that f(A)=Aij/ΣjAij. Note that A(1) is computed from the raw node features while A(t) is computed from the intermediate node embeddings that usually reside on a low-dimensional manifold of the raw node feature space and are optimized towards the downstream prediction task).
Claim 3: Wu teaches the method according to claim 2, wherein the determining the first score of each second node based on the edge data between each second node and the neighborhood node of each second node comprises:
determining a first matrix based on an adjacency matrix of the second nodes of the graph data and a first reference matrix of the second nodes of the graph data; wherein the adjacency matrix is obtained based on the edge data, the first reference matrix of the second nodes of the graph data is a column matrix comprising a plurality of preset values; and determining the first score of each second node based on the first matrix (Wu, [0060] note The disclosed iterative method adjusts when to stop in each mini-batch, such as when the learned graph structure is similar enough to the optimized graph based on the given stopping criterion, [0124] note repeating the generating, producing, and feeding back operations continues until a corresponding updated adjacency matrix designated A(t) as is sufficiently similar to a corresponding optimized graph for use in prediction).
Claim 4: Wu teaches the method according to claim 2, wherein the determining the feature information of the third node based on the first score and the feature information of the second node comprises:
determining a fifth matrix based on an initial matrix of the second nodes of the graph data and the third node; wherein the initial matrix is obtained based on the feature information of the second node, and the fifth matrix is a matrix comprising the feature information of the third node; and determining the feature information of the third node based on the fifth matrix (Wu, [0078] note A is extracted from S by considering only the E-neighborhood for each node; specifically, those elements in S which are smaller than a non-negative threshold ε… where L(0) is the normalized adjacency matrix of the initial graph, defined as, L(0)=D(0)-1/2A(0)D(0)-1/2, and D(0) is its degree matrix. A(t) and A(1) are the two adjacency matrices learned at the t-th and 1-st iterations using Eq. (2) and Eq. (3), respectively. The adjacency matrix learned by Eqs. (2) and (3) is row normalized such that each row sums to 1, so that f(A)=Aij/ΣjAij. Note that A(1) is computed from the raw node features while A(t) is computed from the intermediate node embeddings that usually reside on a low-dimensional manifold of the raw node feature space and are optimized towards the downstream prediction task).
Claim 5: Wu teaches the method according to claim 4, wherein the determining the second score of each third node based on the feature information of each third node comprises:
determining a second matrix based on the fifth matrix, wherein the second matrix is a column matrix comprising the second scores of score of the third nodes (Wu, [0078] note A is extracted from S by considering only the E-neighborhood for each node; specifically, those elements in S which are smaller than a non-negative threshold ε… where L(0) is the normalized adjacency matrix of the initial graph, defined as, L(0)=D(0)-1/2A(0)D(0)-1/2, and D(0) is its degree matrix. A(t) and A(1) are the two adjacency matrices learned at the t-th and 1-st iterations using Eq. (2) and Eq. (3), respectively. The adjacency matrix learned by Eqs. (2) and (3) is row normalized such that each row sums to 1, so that f(A)=Aij/ΣjAij. Note that A(1) is computed from the raw node features while A(t) is computed from the intermediate node embeddings that usually reside on a low-dimensional manifold of the raw node feature space and are optimized towards the downstream prediction task).
Claim 6: Wu teaches the method according to claim 5, wherein the determining the second score of each third node based on the feature information of each third node further comprises: determining an eighth matrix based on the second matrix, wherein the eighth matrix is a column matrix comprises a difference between the second score of each third node and an average score corresponding to second scores; determining a third matrix based on the second matrix and a second reference matrix, wherein the second reference matrix is determined based on the eighth matrix; and determining the second scores of the third nodes based on the third matrix (Wu, [0078] note A is extracted from S by considering only the E-neighborhood for each node; specifically, those elements in S which are smaller than a non-negative threshold ε… where L(0) is the normalized adjacency matrix of the initial graph, defined as, L(0)=D(0)-1/2A(0)D(0)-1/2, and D(0) is its degree matrix. A(t) and A(1) are the two adjacency matrices learned at the t-th and 1-st iterations using Eq. (2) and Eq. (3), respectively. The adjacency matrix learned by Eqs. (2) and (3) is row normalized such that each row sums to 1, so that f(A)=Aij/ΣjAij. Note that A(1) is computed from the raw node features while A(t) is computed from the intermediate node embeddings that usually reside on a low-dimensional manifold of the raw node feature space and are optimized towards the downstream prediction task, [0097] note at the t-th iteration, A(t) is computed based on Z(t-1) (Line 7), and Z(t) is computed based on Ã(t) (Line 9) which is computed based on A(t) (Eq. (4)). The difference between the adjacency matrices at the t-th iteration and the previous iteration is denoted by δA(t)).
Claim 7: Wu teaches the method according to claim 2, wherein the determining the node information of the first node based on the second score comprises:
determining a sixth matrix based on the first node; wherein the sixth matrix comprises any one of a seventh matrix and a fourth matrix, the seventh matrix is a matrix comprising he feature information of the first node, and the fourth matrix is a matrix comprising a second scores of the first nodes; and determining the node information of the first node based on the sixth matrix (Wu, [0078] note A is extracted from S by considering only the E-neighborhood for each node; specifically, those elements in S which are smaller than a non-negative threshold ε… where L(0) is the normalized adjacency matrix of the initial graph, defined as, L(0)=D(0)-1/2A(0)D(0)-1/2, and D(0) is its degree matrix. A(t) and A(1) are the two adjacency matrices learned at the t-th and 1-st iterations using Eq. (2) and Eq. (3), respectively. The adjacency matrix learned by Eqs. (2) and (3) is row normalized such that each row sums to 1, so that f(A)=Aij/ΣjAij. Note that A(1) is computed from the raw node features while A(t) is computed from the intermediate node embeddings that usually reside on a low-dimensional manifold of the raw node feature space and are optimized towards the downstream prediction task, [0097] note at the t-th iteration, A(t) is computed based on Z(t-1) (Line 7), and Z(t) is computed based on Ã(t) (Line 9) which is computed based on A(t) (Eq. (4)). The difference between the adjacency matrices at the t-th iteration and the previous iteration is denoted by δA(t)).
Claim 8: Wu teaches the method according to claim 1, wherein the second nodes further comprise an unlabeled node, and the unlabeled node comprises a node to be classified, and the prediction result further comprises a predicted category label of the unlabeled node; wherein the method further comprises: determining a predicted category label of the node to be classified based on a prediction result output by a last round of training of the classification model (Wu, [0102] note The input features are bag of words and the task is node classification, [0107] note the disclosed graph learning method can greatly help the node classification task even when the graph topology is given. When the graph topology is not available, compared to GCNkNN, IDGL consistently achieves much better results on all datasets, which shows the power of jointly learning graph structures and GNN parameters).
Claim 13: Wu teaches a computer device, comprising: a processor; and a memory arranged to store computer-executable instructions, the processor, when executing the computer-executable instructions, is configured to:
determine node information of first nodes based on edge data and feature information of second nodes of graph data, wherein the first nodes are some of the second node (Wu, [0066] note Let the graph G=(V, E)ϵ 𝓖 be represented as a set of n nodes viϵV with an initial node feature matrix XϵRd×n, edges (vi, vj)ϵE (binary or weighted) formulating an initial noisy adjacency matrix A(0)ϵRn×n);
calculate feature similarity between each second node and a neighborhood node of each second node based on the node information of the first node (Wu, [0061] note a graph learning neural network uses multi-head self-attention with epsilon-neighborhood sparsification for constructing a graph, [0078] note A symmetric sparse adjacency matrix A is extracted from S by considering only the E-neighborhood for each node; specifically, those elements in S which are smaller than a non-negative threshold ε, [0074] note common options for metrics include cosine similarity, radial basis function (RBF) kernel, and attention mechanisms);
aggregate feature information of each second node based on the feature similarity and feature information of the neighborhood node, wherein the neighborhood node is a first node in a neighborhood of each second node (Wu, [0097] note at the t-th iteration, A(t) is computed based on Z(t-1) (Line 7), and Z(t) is computed based on Ã(t) (Line 9) which is computed based on A(t) (Eq. (4)). The difference between the adjacency matrices at the t-th iteration and the previous iteration is denoted by δA(t));
determine a prediction result based on the aggregated feature information of the second nodes, wherein the prediction result comprises a predicted category label of the labeled node among the second nodes (Wu, [0082] note à is the normalized adjacency matrix, σ(⋅) is a task dependent output function, and ℓ (⋅) is a task-dependent loss function. For instance, for node classification, σ(⋅) is a Softmax function for predicting a probability distribution over a set of classes, [0102] note The input features are bag of words and the task is node classification); and
iteratively update a parameter of a classification model based on the predicted category label of the labeled node and a real category label of the labeled node (Wu, [0095] note After all iterations, the overall loss is back-propagated through all previous iterations to update the model parameters).
Claim 14: Wu teaches the computer device according to claim 13, wherein the processor is configured to:
determine a first score of each second node based on the edge data between each second node and the neighborhood node of each second node; determine feature information of thirds node based on first scores and the feature information of the second nodes; wherein the third nodes are selected from the second nodes; determine a second score of each third node based on the feature information of each third node; and determine the node information of the first node based on the second score (Wu, [0078] note A is extracted from S by considering only the E-neighborhood for each node; specifically, those elements in S which are smaller than a non-negative threshold ε… where L(0) is the normalized adjacency matrix of the initial graph, defined as, L(0)=D(0)-1/2A(0)D(0)-1/2, and D(0) is its degree matrix. A(t) and A(1) are the two adjacency matrices learned at the t-th and 1-st iterations using Eq. (2) and Eq. (3), respectively. The adjacency matrix learned by Eqs. (2) and (3) is row normalized such that each row sums to 1, so that f(A)=Aij/ΣjAij. Note that A(1) is computed from the raw node features while A(t) is computed from the intermediate node embeddings that usually reside on a low-dimensional manifold of the raw node feature space and are optimized towards the downstream prediction task).
Claim 15: Wu teaches the computer device according to claim 14, wherein the processor is configured to:
determine a first matrix based on an adjacency matrix of the second nodes of the graph data and a first reference matrix of the second nodes of the graph data; wherein the adjacency matrix is obtained based on the edge data, the first reference matrix of the second nodes of the graph data is a column matrix comprising a plurality of preset values; and determine the first score of each second node based on the first matrix (Wu, [0060] note The disclosed iterative method adjusts when to stop in each mini-batch, such as when the learned graph structure is similar enough to the optimized graph based on the given stopping criterion, [0124] note repeating the generating, producing, and feeding back operations continues until a corresponding updated adjacency matrix designated A(t) as is sufficiently similar to a corresponding optimized graph for use in prediction).
Claim 16: Wu teaches the computer device according to claim 14, wherein the processor is configured to:
determine a fifth matrix based on an initial matrix of the second nodes of the graph data and the third node; wherein the initial matrix is obtained based on the feature information of the second node, and the fifth matrix is a matrix comprising the feature information of the third node; and determine the feature information of the third node based on the fifth matrix (Wu, [0078] note A is extracted from S by considering only the E-neighborhood for each node; specifically, those elements in S which are smaller than a non-negative threshold ε… where L(0) is the normalized adjacency matrix of the initial graph, defined as, L(0)=D(0)-1/2A(0)D(0)-1/2, and D(0) is its degree matrix. A(t) and A(1) are the two adjacency matrices learned at the t-th and 1-st iterations using Eq. (2) and Eq. (3), respectively. The adjacency matrix learned by Eqs. (2) and (3) is row normalized such that each row sums to 1, so that f(A)=Aij/ΣjAij. Note that A(1) is computed from the raw node features while A(t) is computed from the intermediate node embeddings that usually reside on a low-dimensional manifold of the raw node feature space and are optimized towards the downstream prediction task).
Claim 17: Wu teaches the computer device according to claim 16, wherein the processor is configured to:
determine a second matrix based on the fifth matrix, wherein the second matrix is a column matrix comprising the second scores of the third nodes (Wu, [0078] note A is extracted from S by considering only the E-neighborhood for each node; specifically, those elements in S which are smaller than a non-negative threshold ε… where L(0) is the normalized adjacency matrix of the initial graph, defined as, L(0)=D(0)-1/2A(0)D(0)-1/2, and D(0) is its degree matrix. A(t) and A(1) are the two adjacency matrices learned at the t-th and 1-st iterations using Eq. (2) and Eq. (3), respectively. The adjacency matrix learned by Eqs. (2) and (3) is row normalized such that each row sums to 1, so that f(A)=Aij/ΣjAij. Note that A(1) is computed from the raw node features while A(t) is computed from the intermediate node embeddings that usually reside on a low-dimensional manifold of the raw node feature space and are optimized towards the downstream prediction task).
Claim 18: Wu teaches the computer device according to claim 17, wherein the processor is configured to:
determine an eighth matrix based on the second matrix, wherein the eighth matrix is a column matrix comprises a difference between the second score of each third node and an average score corresponding to second scores; determine a third matrix based on the second matrix and a second reference matrix, wherein the second reference matrix is determined based on the eighth matrix; and determine the second scores of the third nodes based on the third matrix (Wu, [0078] note A is extracted from S by considering only the E-neighborhood for each node; specifically, those elements in S which are smaller than a non-negative threshold ε… where L(0) is the normalized adjacency matrix of the initial graph, defined as, L(0)=D(0)-1/2A(0)D(0)-1/2, and D(0) is its degree matrix. A(t) and A(1) are the two adjacency matrices learned at the t-th and 1-st iterations using Eq. (2) and Eq. (3), respectively. The adjacency matrix learned by Eqs. (2) and (3) is row normalized such that each row sums to 1, so that f(A)=Aij/ΣjAij. Note that A(1) is computed from the raw node features while A(t) is computed from the intermediate node embeddings that usually reside on a low-dimensional manifold of the raw node feature space and are optimized towards the downstream prediction task, [0097] note at the t-th iteration, A(t) is computed based on Z(t-1) (Line 7), and Z(t) is computed based on Ã(t) (Line 9) which is computed based on A(t) (Eq. (4)). The difference between the adjacency matrices at the t-th iteration and the previous iteration is denoted by δA(t)).
Claim 19: Wu teaches the computer device according to claim 14, wherein the processor is configured to:
determine a sixth matrix based on the first node; wherein the sixth matrix comprises any one of a seventh matrix and a fourth matrix, the seventh matrix is a matrix comprising the feature information of the first nodes, and the fourth matrix is a matrix comprising second scores of the first nodes; and determine the node information of the first node based on the sixth matrix (Wu, [0078] note A is extracted from S by considering only the E-neighborhood for each node; specifically, those elements in S which are smaller than a non-negative threshold ε… where L(0) is the normalized adjacency matrix of the initial graph, defined as, L(0)=D(0)-1/2A(0)D(0)-1/2, and D(0) is its degree matrix. A(t) and A(1) are the two adjacency matrices learned at the t-th and 1-st iterations using Eq. (2) and Eq. (3), respectively. The adjacency matrix learned by Eqs. (2) and (3) is row normalized such that each row sums to 1, so that f(A)=Aij/ΣjAij. Note that A(1) is computed from the raw node features while A(t) is computed from the intermediate node embeddings that usually reside on a low-dimensional manifold of the raw node feature space and are optimized towards the downstream prediction task, [0097] note at the t-th iteration, A(t) is computed based on Z(t-1) (Line 7), and Z(t) is computed based on Ã(t) (Line 9) which is computed based on A(t) (Eq. (4)). The difference between the adjacency matrices at the t-th iteration and the previous iteration is denoted by δA(t)).
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 9 and 10 are rejected under 35 U.S.C. 103 as being unpatentable over Wu in view of Chang et al., US 2021/0049225 A1 (hereinafter “Chang”).
Claim 9: Wu teaches the method according to claim 1, further comprising:
acquiring the graph data corresponding to a classification task; wherein, in a condition that the classification task is a task for classifying articles, graph data of the task for classifying articles is acquired, wherein the second node comprises an article node (Wu, [0060] note the disclosed iterative method aim to search for a hidden graph structure that augments the initial graph structure with the goal of optimizing the graph for supervised prediction tasks, [0102] note The task is also node classification. Finally, to demonstrate the effectiveness of the IDGL framework 400 on inductive learning problems, document classification and regression tasks were conducted on the 20Newsgroups data (20News) and the movie review data (MRD), respectively, where a document was treated as a graph containing each word as a node).
Wu does not explicitly teach in a condition that the classification task is a task for classifying risky users, graph data of the task for classifying risky users is acquired, wherein the second node comprises a user node; and in a condition that the classification task is a task for classifying pushed users, graph data of the task for classifying pushed users is acquired, wherein the second node comprises the user node.
However, Chang teaches this (Chang, [0076] note the first target sub-graph is input to a pre-trained neural network model, and the neural network model outputs the first feature vector based on the node feature of each node included in the first target sub-graph and the directional relationship of the edge between the nodes, [0084] note when the object corresponding to the first target node is a user, a user type of the user can be predicted based on the first feature vector, for example, a crowd type or a risk class type. When the object corresponding to the first target node is a product, a type of the product can be predicted based on the first feature vector, for example, a service type, a crowd type to which the product is appropriate, or a scenario type in which the product is purchased, [0100] note when the sample object is a user, the classification label can be a predetermined label of a crowd classification or a label of a user risk degree classification; when the sample object is a product, the classification label can be a label of a product classification).
It would have been obvious to one of ordinary skill in the art at the effective filing date of the application to combine the graph neural network of Wu with the graph neural network classification of Chang according it is desired to perform feature expression and modeling on participants of interactions based on the interactions (Chang, [0003]).
Claim 10: Wu and Chang teach the method according to claim 9, wherein the feature information of the second node comprises feature information of an article or a user, and the edge data between the second nodes comprises a connection edge between two second nodes, and the connection edge indicates that there is a preset association relationship between two articles or between two users (Chang, [0076] note the first target sub-graph is input to a pre-trained neural network model, and the neural network model outputs the first feature vector based on the node feature of each node included in the first target sub-graph and the directional relationship of the edge between the nodes, [0084] note when the object corresponding to the first target node is a user, a user type of the user can be predicted based on the first feature vector, for example, a crowd type or a risk class type. When the object corresponding to the first target node is a product, a type of the product can be predicted based on the first feature vector, for example, a service type, a crowd type to which the product is appropriate, or a scenario type in which the product is purchased, [0100] note when the sample object is a user, the classification label can be a predetermined label of a crowd classification or a label of a user risk degree classification; when the sample object is a product, the classification label can be a label of a product classification).
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Giuseppi Giuliani whose telephone number is (571)270-7128. The examiner can normally be reached Monday-Friday.
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, Kavita Stanley can be reached at (571)272-8352. 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.
/GIUSEPPI GIULIANI/Primary Examiner, Art Unit 2153
Prosecution Insights