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 .
Claims 1-6 have been examined and are pending. Claims 7-9 are non-elected.
Priority
Receipt is acknowledged of certified copies of papers required by 37 CFR 1.55.
Allowable Subject Matter
Claim 4 is objected to as being dependent upon a rejected base claim, but
would be allowable if rewritten in independent form including all of the limitations of the
base claim and all intervening claims.
Examiner Comments
Claim 6 is directed towards a system of training a graph neural network, comprising a plurality of distributed training terminals and a database, have been analyzed for 35 USC 101. The claims comprises a system comprising terminals and a database. Therefore, analysis of a distributed training terminals and a database; No 35 USC 101 deemed necessary since specification states: “The computer device includes, for example, a terinal device, a server, or a further processing device. The terminal device may be a user equipment (UE), a mobile device, a user terminal, a terminal, a personal digital assistant (PDA), a handheld device, a computing device, an in-vehicle device, a wearable device, or the like. In some possible implementations, the method of training a graph neural network may be implemented by a processor invoking a computer-readable instruction stored in a memory.” (para 0043).
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 10/28/2024 and 03/05/2026 was filed. The submission is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Objections
Claim 1 is objected to because of the following informalities:
Claims 1, line 5: removal of intentional use term: being. 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.
Claim 2 is 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 2 recites the limitation "…segmenting the sample…" in line 3. There is insufficient antecedent basis for this limitation in claim 2. The metes and bounds of the claim are unclear.
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 1-6 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Claim 1 recites the following limitations:
“…obtaining initial graph structure data corresponding to the terminal device, the initial graph structure data respectively obtained by the plurality of distributed training terminals being derived from the same sample graph structure data; and
performing the following graph structure data processing stage and graph neural network training stage cyclically, until a target neural network satisfying a training requirement is obtained:
determining a processing opportunity for currently performing a graph structure data processing stage based on historical execution data of historically performing a graph structure data processing stage and a graph neural network training stage...”
Each of these limitations would be practical to perform in the mind with the aid of pencil and paper, thus directed towards a mental process (see MPEP §2106.04(a)(2)(III)). Each of these limitations collectively manipulates flow data, and tracks the presence of behavior of various flow data groupings indicative of potential maliciousness. This is the type of analysis that goes into network planning and can reasonably be done in the human mind. As result, the limitations listed recite an abstract idea.
This judicial exception is not integrated into a practical application. Claim 1 further recites “performing, based on the processing opportunity, graph structure data processing on the initial graph structure data in the graph structure data processing stage, to generate target graph structure data; the graph structure data processing comprising data sampling processing and feature extraction processing.” The claimed device is a generic computer component that is being claimed as just a tool to perform the claimed mental steps. Performing an abstract idea on a computer tool does not transform the abstract idea into a practical application (see MPEP §2106.05(f)). Claim 1 also recites “training, based on the target graph structure data, the target neural network in the graph neural network training stage.” The device receiving the claimed flow data is a data gathering step which uses the data from the received flow data to perform the mental process. The claimed data gathering step is insignificant extra solution activity and does not transform the claimed abstract idea into a practical application (see MPEP §2106.05(g)). The additional elements have been considered alone, and in combination with the claimed invention as a whole, but does not integrate the abstract idea into a practical application. As result, the invention is directed towards an abstract idea.
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements of the device amounts to no more than mere instructions to apply the exception using a generic computer component. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. The reception of the claimed data packets is an insignificant extra solution activity, which is additional well understood routine, and conventional. MPEP §2106.05(g) details similar data gather steps that have been found by the courts to be well understood routine, and conventional. Additionally applicant’s specification (see ¶¶0044-0054) provides only broad disclosure of the reception of flows and makes it clear that it considers the reception of the claimed flow data as well-known aspects of the disclosure. As result, the claim is not patent eligible.
Claim 6 is directed towards a system rather than the system of claim 1, however the same rationale applies to claim 6 as provided in the rejection to claim 1. As result, claim 6 is not patent eligible.
Claim 2 recites “…wherein the initial graph structure data is allocated based on the following: segmenting the sample graph structure data based on a breadth first search algorithm, to obtain a plurality of first graph structure data; and allocating, from the plurality of first graph structure data, the initial graph structure data to the distributed training terminals based on a greedy algorithm, a number of terminals of the distributed training terminals, and a number of segments of the first graph structure data obtained by segmentation...” The claim provides additional limitations that describe mental processes. As result, when additional features of claim 2, when considered alone and in combination, are still directed to an abstract idea which contains nothing significantly more than the judicial exception itself.
Claim 3 recites “…wherein performing the graph structure data processing on the initial graph structure data comprises: generating second graph structure data for characterizing a dependency relationship between respective operators based on an execution logic of a plurality of sampling processing and a plurality of feature extraction processing; determining a topological order of the respective operators based on the second graph structure data; and performing, based on the topological order of the respective operators, a plurality of sampling processing and a plurality of feature extraction processing on the initial graph structure data...” The claim provides additional limitations that describe mental processes. As result, when additional features of claim 3, when considered alone and in combination, are still directed to an abstract idea which contains nothing significantly more than the judicial exception itself.
Claim 4 recites “…wherein determining the processing opportunity for currently performing the graph structure data processing stage based on the historical execution data of historically performing the graph structure data processing stage and the graph neural network training stage comprises: determining whether to perform the current graph structure data processing at a current time instant based on a performance parameter of the terminal device, a first historical average duration for historically performing the graph structure data processing stage, and a second historical average duration for historically performing the graph neural network training stage; and in accordance with a determination of performing the current graph structure data processing at the current time instant, determining the processing opportunity as the current time instant; otherwise, redetermining whether to perform the current graph result data processing after a predetermined time interval…” The claim provides additional limitations that describe mental processes. As result, when additional features of claim 4, when considered alone and in combination, are still directed to an abstract idea which contains nothing significantly more than the judicial exception itself.
Claim 5 recites “… wherein training the target neural network based on the target graph structure data comprises: inputting execution code corresponding to the graph neural network training stage into a target code compiler to obtain third graph structure data generated by the target code compiler after compiling the execution code; and performing, based on the third graph structure data, data processing on the target graph structure data to implement training of the target neural network...” The claim provides additional limitations that describe mental processes. As result, when additional features of claim 5, when considered alone and in combination, are still directed to an abstract idea which contains nothing significantly more than the judicial exception itself.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 1 and 5-6 are rejected under 35 U.S.C. 103 as being unpatentable over Shao et al, hereinafter (“Shao”),US PG Publication 20230419823A1, in view of, Wu et al, hereinafter (“Wu”), Chinese Application CN 111860783 A.
Regarding claims 1 and 6, Shao teaches a method of training a graph neural network, applied to any terminal device in a plurality of distributed training terminals for training a same target neural network, the method comprising; and a system of training a graph neural network, comprising a plurality of distributed training terminals and a database, wherein: [Shao¶¶0060 0066-0067 and 0069 Fig. 4 shows flowchart for determining vehicle information based on current monitoring data performed by management platform 230; where a training process is performed by management platform 230 at various layer using Convolutional Neural Networks (CNN), Recurrent Neural Network (RNN), or Neural Network (NN)
Wu abstract and p. 3, ¶4 and ¶6 purpose of invention is to employ a method of operations of a terminal device employing a graph node low-dimensional representation learning method; where the terminal device comprises a memory for storing computer program controlling the operations.]
obtaining initial graph structure data corresponding to the terminal device, the initial graph structure data respectively obtained by the plurality of distributed training terminals being derived from the same sample graph structure data [Shao¶¶0005 0060 0064 and 0069 current monitoring data in a preset area are collected, processed and stored by the one or more object platforms may be obtained through a sensor network platform managing a smart city of Internet of Things. Fig. 4 shows flowchart for determining vehicle information based on current monitoring data performed by the management platform 230. ¶¶0120-0121 determining amount of exhaust emission of each sampled road segment through processing road network graph structure data by Graph Neural Network (GNN). The road network graph structure data in the preset area may include nodes and edges between the nodes.]; and
performing the following graph structure data processing stage and graph neural network training stage cyclically, until a target neural network satisfying a training requirement is obtained [Shao¶0021 determine vehicle information, and then determine the corresponding total amount of exhaust emission, generate corresponding vehicle limit information and provide the vehicle limit information to the user terminal 130. ¶0072 plurality of labeled training samples are respectively input into corresponding initial image feature extraction layer, initial sequence feature layers, and initial data prediction layer; updated through training iterations. ¶¶0122-0124 The nodes, edges and edge attributes in the road network graph structure data may be determined according to the road network data. The input of the GNN may be the road network graph structure data of the target area, and each edge of the GNN may output the amount of exhaust emission of each sampled road segment in each preset area. The GNN and fusion layer may be trained separately.]:
determining a processing opportunity for currently performing a graph structure data processing stage based on historical execution data of historically performing a graph structure data processing stage and a graph neural network training stage [Shao¶¶0124-0125 the GNN model may be determined through training an initial GNN model based on training data. The initial GNN model may refer to a GNN model with no parameters set. The training data may include training samples and training labels, wherein the training samples may be historical road network graph structure data determined based on historical data, and the training labels may be the historical amount of exhaust emission of each sampled road segment. ];
While Shao teaches performing, based on the processing opportunity, graph structure data processing on the initial graph structure data in the graph structure data processing stage, to generate target graph structure data; the graph structure data processing comprising data sampling processing [See Shao¶¶0072-0073 and 0075 plurality of labeled training samples are respectively input into corresponding initial image feature extraction layer. ¶¶0123-0125 input of the GNN may be the road network graph structure data of the target area, and each edge of the GNN may output the amount of exhaust emission of each sampled road segment in each preset area.]; however, Shao fails to explicitly teach but Wu teaches performing, based on the processing opportunity, graph structure data processing on the initial graph structure data in the graph structure data processing stage, to generate target graph structure data; the graph structure data processing comprising data sampling processing and feature extraction processing [Wu abstract and p. 2, ¶¶1 and 3 invention of computer information processing technology field relates to graph node low-dimensional representation learning method; network (graph) structure data composed of points and sides, utilize DeepWalk,node2vec and LINE, etc. p. 3, p. 2, ¶5 realized high-dimensional non-Euclidean space data: extracting information of graph. p. 3, ¶1 and 7-8 perform random walk sampling…corresponding sampling node sequence set common sampling methods (i.e. based on similarity between nodes, etc.). Examiner interprets the graph node low-dimensional representation learning method as analogous to performing, based on the processing opportunity, graph structure data processing on the initial graph structure data in the graph structure data processing stage, to generate target graph structure data]; and
training, based on the target graph structure data, the target neural network in the graph neural network training stage [Wu ¶abstract: invention of computer information processing technology field iterative training; invention solves the problem that the sampling algorithm. p. 3, ¶¶2-3 efficiency of neural network training. Examiner interprets the iterative training as analogous to training, based on the target graph structure data, the target neural network in the graph neural network training stage].
Shao teaches all the features of claim 1 not performing, based on the processing opportunity, graph structure data processing on the initial graph structure data in the graph structure data processing stage, to generate target graph structure data; the graph structure data processing comprising data sampling processing and feature extraction processing; and training, based on the target graph structure data, the target neural network in the graph neural network training stage. Shao teaches a methods and systems for managing exhaust emission in a smart city based on industrial internet of things. Wu teaches a graph node low-dimensional representation learning method, device, terminal device and storage medium. Both Shao and Wu teach neural networks that learn how to process and generate graph structure data through sampling and feature extraction. Because both Shao and Wu teach clustering concepts, it would have been obvious to one skilled in the art before the effective filing date of the claimed invention was made to use improved and enhanced graph node low-dimensional representation learning/deep neural network low-dimensional characterization methods to by utilizing sampling algorithm/random walk sampling mode and extracting information of the graph [Wu, p. 2, ¶4].
Regarding claim 5, the combination of Shao and Wu teaches claim 1 as described above.
However, Shao fails to explicitly teach but Wu teaches wherein training the target neural network based on the target graph structure data comprises: inputting execution code corresponding to the graph neural network training stage into a target code compiler to obtain third graph structure data generated by the target code compiler after compiling the execution code; and performing, based on the third graph structure data, data processing on the target graph structure data to implement training of the target neural network. [Wu p. 5 ¶¶1-2 computer program stored in memory executed by processor, the steps of the graph node low-dimensional representation learning method. See Abstract: iterative training. Examiner interprets that the iterative training generates a plurality of graph structure data (i.e. a second graph structure data, a third graph structure data, etc.) by utilizing a computer program has already been/will be compiled and available with trained algorithms to obtain third graph structure data generated by the target code compiler after compiling the execution code; and performing, based on the third graph structure data, data processing on the target graph structure data to implement training of the target neural network.
Shao teaches all the features of claim 1 not performing, based on the processing opportunity, graph structure data processing on the initial graph structure data in the graph structure data processing stage, to generate target graph structure data; the graph structure data processing comprising data sampling processing and feature extraction processing; and training, based on the target graph structure data, the target neural network in the graph neural network training stage. Shao teaches a methods and systems for managing exhaust emission in a smart city based on industrial internet of things. Wu teaches a graph node low-dimensional representation learning method, device, terminal device and storage medium. Both Shao and Wu teach neural networks that learn how to process and generate graph structure data through sampling and feature extraction. Because both Shao and Wu teach clustering concepts, it would have been obvious to one skilled in the art before the effective filing date of the claimed invention was made to use improved and enhanced graph node low-dimensional representation learning/deep neural network low-dimensional characterization methods to by utilizing sampling algorithm/random walk sampling mode and extracting information of the graph [Wu, p. 2, ¶4].
Claim(s) 2-3 are rejected under 35 U.S.C. 103 as being unpatentable over Shao et al, hereinafter (“Shao”),US PG Publication 20230419823A1, in view of, Wu et al, hereinafter (“Wu”), Chinese Application CN 111860783 A, in view of Qi et al, hereinafter (“Qi”), Chinese Application CN-113807404-A.
Regarding claim 2, the combination of Shao and Wu teaches claim 1 as described above.
However, the combination of Shao and Wu fails to explicitly teach but Qi teaches wherein the initial graph structure data is allocated based on the following: segmenting the sample graph structure data based on a breadth first search algorithm, to obtain a plurality of first graph structure data [p. 3, ¶¶9-10 and 14 high speed all-net graph structure data represented by two-dimensional matrix; analyzes path planning algorithm. p. 4, ¶4 updating and planning source point(s) to the adjacent point(s) by the graph traversal method of breadth-first search]; and allocating, from the plurality of first graph structure data, the initial graph structure data to the distributed training terminals based on a greedy algorithm, a number of terminals of the distributed training terminals, and a number of segments of the first graph structure data obtained by segmentation [p. 3, ¶6 and 10-11 coordinate on the matrix of the high speed full-network graph structure comprising portal frame camera device state; p. 4, ¶4 Dijkstra algorithm is mainly used for planning the shortest path from a given source point to other nodes. Calculating method using greedy idea…].
The combination of Shao and Wu teach all the features of claim 1 wherein the initial graph structure data is allocated based on the following: segmenting the sample graph structure data based on a breadth first search algorithm, to obtain a plurality of first graph structure data; and allocating, from the plurality of first graph structure data, the initial graph structure data to the distributed training terminals based on a greedy algorithm, a number of terminals of the distributed training terminals, and a number of segments of the first graph structure data obtained by segmentation. Qi teaches an invention from the machine learning and statistical machine learning field. Shao teaches a methods and systems for managing exhaust emission in a smart city based on industrial internet of things. Wu teaches a graph node low-dimensional representation learning method, device, terminal device and storage medium. Because Shao, Wu, and Qi teach clustering concepts, it would have been obvious to one skilled in the art before the effective filing date of the claimed invention was made to use single-point position and high speed full-network graph structure data and Gaussian mixture model (GMM) to cluster new obtained representation characteristic [p.6, ¶9 and p. 7, ¶¶5-6].
Regarding claim 3, the combination of Shao and Wu teaches claim 1 as described above.
However, the combination of Shao and Wu fails to explicitly teach but Qi teaches wherein performing the graph structure data processing on the initial graph structure data comprises: generating second graph structure data for characterizing a dependency relationship between respective operators based on an execution logic of a plurality of sampling processing and a plurality of feature extraction processing [p. 9, ¶12 “the high speed all-net graph structure data is graph structure data represented by two-dimensional matrix, taking the portal frame collecting point as the graph node, the relation of the highway network adjacent…” p. 10, ¶1 The Flord-warding algorithm comprises: traversing all nodes from the node set V, when traversing one node, using temporary variable k to mark the node, the node is used as the intermediate node... p. 11, ¶1 …wherein, k, u, v is one node in the graph, dist [u] [v] represents the minimum transmission time delay of point u to point v p. 14, ¶2 and 4 3a) performing data cleaning…multi-source heterogeneous data is pre-processed.. sampling the image and video data of main body…3b) performing feature extraction to the unstructured data…]; determining a topological order of the respective operators based on the second graph structure data [p. 14, ¶4 3b) performing feature extraction to the unstructured data…converted into image data sequence with time by means of frame-by-frame extraction;]; and performing, based on the topological order of the respective operators, a plurality of sampling processing and a plurality of feature extraction processing on the initial graph structure data [See p. 14, ¶2 and 4 ... sampling the image and video data of main body…3b) performing feature extraction to the unstructured data…].
The combination of Shao and Wu teach all the features of claim 1 wherein the initial graph structure data is allocated based on the following: segmenting the sample graph structure data based on a breadth first search algorithm, to obtain a plurality of first graph structure data; and allocating, from the plurality of first graph structure data, the initial graph structure data to the distributed training terminals based on a greedy algorithm, a number of terminals of the distributed training terminals, and a number of segments of the first graph structure data obtained by segmentation. Qi teaches an invention from the machine learning and statistical machine learning field. Shao teaches a methods and systems for managing exhaust emission in a smart city based on industrial internet of things. Wu teaches a graph node low-dimensional representation learning method, device, terminal device and storage medium. Because Shao, Wu, and Qi teach clustering concepts, it would have been obvious to one skilled in the art before the effective filing date of the claimed invention was made to use single-point position and high speed full-network graph structure data and Gaussian mixture model (GMM) to cluster new obtained representation characteristic [p.6, ¶9 and p. 7, ¶¶5-6].
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Naser Eddin et al 20250013923 A1 teaches a method for low-latency feature extraction for training a machine-learning model.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SAKINAH WHITE-TAYLOR whose telephone number is (571)270-0682. The examiner can normally be reached Monday-Friday, 10:45a-6:45p.
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, CATHERINE THIAW can be reached at 571-270-1138. 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.
SAKINAH WHITE-TAYLOR
Primary Examiner
Art Unit 2407
/Sakinah White-Taylor/Primary Examiner, Art Unit 2407
Prosecution Insights