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 .
This action is responsive to the claims dated 12/29/2023.
Claims 1-8 are presented for examination.
Specification
The use of the terms RASPBERRY PI 4B, BROADCOM BCM2711, RASPBIAN, JETSON TX2, NVIDIA PASCAL, ARM CORTEX-A57, UBUNTU, WIFI, and PYTORCH (paragraph [0128], as also shown in FIG. 3), which are trade names or marks used in commerce, has been noted in this application. Each term should be accompanied by the generic terminology; furthermore each term should be capitalized wherever it appears or, where appropriate, include a proper symbol indicating use in commerce such as TM, SM, or (R) following the term. Although the use of trade names and marks used in commerce (i.e., trademarks, service marks, certification marks, and collective marks) is permissible in patent applications, the proprietary nature of the marks should be respected and every effort made to prevent their use in any manner which might adversely affect their validity as commercial marks. Appropriate correction is required.
The disclosure is objected to because of the following informalities: (a) at paragraph [0052], the phrase "Proposed in the present invention is a Edge-Client Collaborative Federated Graph Learning with Adaptive Neighbor Generation" recites the article "a" where the grammatically correct article "an" is required before the vowel-initial term "Edge-Client"; and (b) at paragraph [0044], the brief description of FIG. 3 recites "FedGL and SpreadFG," whereas the proposed distributed framework is named "SpreadFGL" throughout the remainder of the disclosure, so that "SpreadFG" should read "SpreadFGL" for consistent terminology. Appropriate correction is required.
Claim Rejections - 35 U.S.C. 112(a)
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 1-8 are rejected under 35 U.S.C. 112(a) 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, at the time the application was filed, had possession of the claimed invention.
Claim 1 recites, in its entirety, that the invention comprises: consider a typical FGL scenario with distributed graph datasets; based on this setting, first propose an improved centralized FGL framework, named FedGL; next, extend the FedGL to a scenario of multi-edge collaboration and propose a novel distributed FGL framework, named SpreadFGL. This language does not describe the structure, components, or operative steps of the claimed invention. Rather, it recites only the inventor's research objective, namely to propose two frameworks identified solely by the names FedGL and SpreadFGL and by the desired results those frameworks are said to achieve (an improved centralized framework and a novel distributed framework for federated graph learning with adaptive neighbor generation). A claim limitation defined by a name and a desired result, rather than by the features that achieve that result, does not convey possession of the invention. Possession is shown by what is described, not by what is named or aspired to (MPEP 2163; Ariad Pharmaceuticals, Inc. v. Eli Lilly & Co., 598 F.3d 1336, 1351 (Fed. Cir. 2010) (en banc)).
So construed under its broadest reasonable interpretation, claim 1 reads on any centralized and any distributed federated graph learning framework that achieves the recited adaptive neighbor generation result and that may be called FedGL and SpreadFGL. This is a genus defined by function and result. The specification, however, describes only a single embodiment of each framework, namely the particular FedGL and SpreadFGL architectures set out at specification paragraphs [0054] through [0124] (a graph imputation generator, an autoencoder, a versatile assessor, a negative sampling mechanism, graphic patchers, and a trace-normalization weight regularizer with a specified multi-edge parameter-averaging rule). The specification does not describe a representative number of additional species, nor does it identify structural features common to the claimed genus that would let a person of ordinary skill in the art recognize the full range of frameworks embraced by claim 1. A single disclosed species with a particular architecture does not support a genus defined solely by the function or result it is said to achieve (Abbvie Deutschland GmbH & Co. v. Janssen Biotech, Inc., 759 F.3d 1285, 1300-01 (Fed. Cir. 2014); see also MPEP 2163). Although original claim 1 is part of the original disclosure, the presumption that an original claim provides its own written description is rebutted here because the broadest reasonable scope of claim 1 exceeds what the specification as a whole conveys the inventor possessed (In re Koller, 613 F.2d 819, 823 (CCPA 1980)).
Dependent claims 2 through 8 do not cure the deficiency of claim 1. Each incorporates the aspirational framework recitation of claim 1 by reference. The additional limitations of the dependent claims (for example, the L-layer GNN node classifier and parameter-update rule of claim 2, the embedding fusion of claim 3, the autoencoder of claim 4, the versatile assessor of claim 5, the negative sampling of claim 6, the graphic patcher of claim 7, and the multi-edge SpreadFGL aggregation of claim 8) describe particular components of the single disclosed embodiment and therefore do not broaden the disclosure to the full genus claimed in claim 1. Claims 2-8 are accordingly rejected for the same lack of written description as claim 1.
Claim 1 is rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the enablement requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to enable one skilled in the art to which it pertains, or with which it is most nearly connected, to make and/or use the invention.
Claim 1 is further rejected under 35 U.S.C. 112(a) as failing to comply with the enablement requirement. The claim contains subject matter which was not described in the specification in such a way as to enable one skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the invention commensurate with the full scope of the claim. The determination of whether undue experimentation would be required is made under the factors set out in In re Wands, 858 F.2d 731, 737 (Fed. Cir. 1988) see MPEP 2164.01(a):
(A) The breadth of the claims. Claim 1 is broad. As construed above, it reaches any improved centralized and any novel distributed federated graph learning framework named FedGL and SpreadFGL that achieves adaptive neighbor generation, without limitation to the disclosed architecture. This breadth weighs against enablement.
(B) The nature of the invention. The invention is a federated graph learning system implemented in software on edge servers and clients, an incremental improvement on known federated graph learning frameworks.
(C) The state of the prior art. The art of federated graph learning and graph neural networks is well developed, as the specification itself recounts at paragraphs [0003] through [0005] and [0061] through [0075] (GCN, GAT, GraphSAGE, FedAvg, FedSage+). A person of ordinary skill can draw on substantial existing knowledge.
(D) The level of one of ordinary skill. The level of skill is high, typically a researcher or engineer with graduate-level training in machine learning and distributed systems and routine familiarity with graph neural network toolkits.
(E) The level of predictability in the art. Software and machine learning is a predictable art relative to the chemical and biological arts, so the effect of routine modifications to a disclosed architecture is generally foreseeable to a person of ordinary skill.
(F) The amount of direction provided by the inventor. For the single disclosed embodiment the direction is substantial: the specification supplies the architecture, the training procedure, and concrete parameter settings at paragraph [0141] (GraphSAGE with two layers, an autoencoder of neuron counts {c,16,d} and {d,16,c}, an assessor of {c,128,16,1}, the Adam optimizer, learning rates, the threshold theta = 1/c, and the value of k). For the full claimed genus, however, the specification provides no direction beyond that one embodiment.
(G) The existence of working examples. The specification provides working examples, including a real-world hardware testbed (Raspberry Pi 4B clients and Jetson TX2 edge servers, paragraph [0128]) and experiments on the Cora, Citeseer, WikiCS, and CoauthorCS benchmark datasets (paragraphs [0129] through [0149]). These examples are limited to the single disclosed FedGL and SpreadFGL architecture.
(H) The quantity of experimentation needed. Practicing the single disclosed embodiment would require only routine experimentation. Practicing the full aspirational scope of claim 1, namely any other improved centralized or novel distributed framework embraced by the named-and-result recitation, would require a person of ordinary skill to design and validate frameworks that the specification neither identifies nor teaches, amounting to undue experimentation.
On balance, the Wands factors show that while the single embodiment disclosed at specification paragraphs [0054] through [0149] is enabled, the full scope of claim 1 is not. The specification enables only the particular FedGL and SpreadFGL architecture it describes, whereas claim 1, by reciting two frameworks defined by name and by desired result, embraces substantially more. A specification that enables only a narrow embodiment does not enable a claim of materially greater breadth (Liebel-Flarsheim Co. v. Medrad, Inc., 481 F.3d 1371, 1379-80 (Fed. Cir. 2007); Amgen Inc. v. Sanofi, 598 U.S. 594, 610-14 (2023)). Dependent claims 2 through 8, which narrow claim 1 to particular components of the disclosed embodiment, are not separately rejected under the enablement requirement.
To overcome the written description and enablement rejections, applicant may amend claim 1 to recite the structural and operative limitations of the disclosed embodiment (for example, the graph imputation generator, autoencoder, versatile assessor, negative sampling mechanism, graphic patchers, and multi-edge aggregation actually described in the specification) in place of the present recitation that two named frameworks are proposed. No new matter should be introduced (35 U.S.C. 132).
Claim Rejections - 35 U.S.C. 112(b)
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 1-8 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. The specific bases for the rejection are set forth below.
Claim 1 recites, in its entirety, that the invention comprises: consider a typical FGL scenario with distributed graph datasets; based on this setting, first propose an improved centralized FGL framework, named FedGL; next, extend the FedGL to a scenario of multi-edge collaboration and propose a novel distributed FGL framework, named SpreadFGL. This language is cast as a research proposal rather than as a claim. The body of claim 1 recites no structure and no method steps; it recites only the act of proposing two applicant-coined frameworks ("FedGL" and "SpreadFGL") and an aspirational result (adaptive neighbor generation). A person of ordinary skill in the art cannot determine the metes and bounds of the claim, because the scope is defined solely by the coined framework names and a desired outcome, with no recited components, configuration, or operations by which the result is achieved. The claim therefore fails to particularly point out and distinctly claim the subject matter regarded as the invention. See MPEP 2173.05 and 2172; In re Packard, 751 F.3d 1307 (Fed. Cir. 2014); Nautilus, Inc. v. Biosig Instruments, Inc., 572 U.S. 898, 901, 910 (2014). Claim 1 is further rejected under 35 U.S.C. 112(b) as failing to set forth the subject matter which the inventor or a joint inventor regards as the invention: the specification states that the very "purpose of the present invention" is, in the same proposal narrative, to "first propose an improved centralized FGL framework, named FedGL" and to "propose a novel distributed FGL framework, named SpreadFGL" (specification paragraph [0006]), confirming that what is claimed is the act of proposing named frameworks rather than a particular apparatus or method. For the purpose of examination on the merits, and under the broadest reasonable interpretation in light of the specification (MPEP 2111), the limitation is interpreted to be directed to a federated graph learning system, or method, comprising a centralized framework that uses an edge server as an intermediary among clients to generate latent cross-subgraph links (the disclosed FedGL, specification paragraphs [0006]-[0007], [0054]-[0059]) and a distributed multi-edge extension thereof (the disclosed SpreadFGL, specification paragraphs [0007], [0034], and [0117]-[0118]) that achieves adaptive neighbor generation; the coined names FedGL and SpreadFGL are construed as those disclosed embodiments.
Claims 2-8 are each rejected under 35 U.S.C. 112(b) for lack of antecedent basis. Each of claims 2-8 depends directly from claim 1, and claim 1 does not introduce any of the elements to which the dependent claims later refer using the definite article. Accordingly, there is insufficient antecedent basis for the following limitations: the edge server (claims 2, 3, 4, 7, and 8); the clients (claims 3 and 8); the graph imputation generator (claims 3 and 4, the element being first introduced only in claim 2, which is not an antecedent of claims 3 and 4); the autoencoder (claims 5 and 6, the element being first introduced only in claim 4, which is not an antecedent of claims 5 and 6); the assessor (claims 5 and 6, no antecedent introduction of an assessor appearing in any claim from which claims 5 and 6 depend); the local node classifier (claim 2, recited with the definite article without a prior introduction); the global topology graph, the random noisy vector, and the encoder (claim 4); and the local graphic patcher and the subgraph (claim 7). There is insufficient antecedent basis for these limitations in the claims. See MPEP 2173.05(e) and 2173.05(f). For examination on the merits, under the broadest reasonable interpretation each such limitation is read on the correspondingly named component of the single disclosed embodiment, namely the edge server (specification paragraph [0007]), the clients (paragraph [0007]), the graph imputation generator (paragraphs [0028], [0056], and [0147]), the L-layer GNN local node classifier (paragraph [0012]), the autoencoder (paragraph [0057]), the versatile assessor (paragraphs [0027] and [0150]), the global topology graph (paragraph [0018]-[0019]), the random noisy vector and encoder (paragraph [0135]), the local graphic patcher (paragraph [0660]), and the assigned subgraphs (paragraphs [0032] and [0135]).
Claims 1-8 recite a nominal Edge-Client Collaborative Federated Graph Learning framework while also reciting acts such as "consider," "propose," "extend," "incorporate," "employ," "upload," "utilizes," "adopts," "set," "select," "divides," "design," "aggregates," and "broadcasts." This appears to recite hybrid product/process or apparatus/method claim under MPEP 2173.05(p). Claim 1 and its dependent claims are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for improperly mixing method steps and product or framework elements. It is unclear whether infringement of the claim occurs when the framework exists, when the recited software components are configured, or when the recited acts are performed. See IPXL Holdings v. Amazon.com, 430 F.3d 1377, 1384 (Fed. Cir. 2005); MPEP 2173.05(p). For examination the said claims are interpreted as to cover the disclosed FedGL and SpreadFGL software framework and its operation, as described at spec paragraphs [0077], [0087], [0097], [0099], [0102], [0109], [0117]-[0124], and [0150]. Under that BRI, the claims do not identify whether the protected subject matter is the framework as a product or the performance of the listed operations.
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 1-8 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The analysis of the claims will follow the 2019 Revised Patent Subject Matter Eligibility Guidance, 84 Fed. Reg. 50 (“2019 PEG”).
Claim 1
Step 1: Claims 1-8 recites “An Edge-Client Collaborative Federated Graph Learning with Adaptive Neighbor Generation, comprising:”; while per the 35 U.S.C. 112(b) rejection above, it is not clear whether the claims are directed to an apparatus, .e.g. Edge-Client, or A method of, is directed to the statutory category of a process.
Step 2A Prong 1:
Independent Claim
Claim 1 recites:
consider a typical FGL scenario with distributed graph datasets: These limitations recite a mentally performable process of evaluation a typical FGL scenario with distributed graph datasets with the aid of pen and paper.
based on this setting, first propose an improvement centralized FGL framework, named FedGL; next, extend the FedGL to a scenario of multi-edge collaboration and propose a novel distributed FGL framework, named SpreadFGL: These limitations recite a mentally performable process with the aid of pen and paper of observing the setting when considering the typical FGL scenario setting to first propose an improvement centralized FGL framework labeled FedGL, and then consider, e.g. mental know the FedGL to a scenario of multi-edge collaboration to propose a novel distributed FGL framework labeled SpreadFGL.
Dependent Claims
Claims 2-8 further recites additional abstract ideas of the judicial exception with mathematical formulas or equations to further narrow the FGL framework.
Step 2A Prong 2: This judicial exception is not integrated into a practical application. The additional elements of this claim are as follows:
Independent Claim
Claim 1:
An Edge-Client Collaborative Federated Graph Learning with Adaptive Neighbor Generation: These additional elements are recited at a high level of generality an merely represent generic computer components, e.g. Edge-Client, and/or a field of use/technological environment generically linking the underlying judicial exception. See MPEP 2106.05(f)(h).
Dependent Claims
Claims 2-8: Due to great deal of confusion and uncertainty as to the proper interpretation of the limitations of a claim as described in the 35 U.S.C. 112 rejections above, any additional elements recited cannot be shown to integrate the judicial exception into a practical application as there is no clear reflection of how they reflect the specific technological improvement. Thus, the way in which the additional elements use or interact with the judicial exception when analyzed with this claim as a whole do not integrate the judicial exception into a practical application.
Step 2B: The additional elements from Step 2A Prong 2 viewed individually or in combination, do not provide an inventive concept or otherwise amount to significantly more than the abstract idea itself. See MPEP 2106.05.
Conclusion
When there is a great deal of confusion and uncertainty as to the proper interpretation of the limitations of a claim, it is not proper to reject such a claim on the basis of prior art. See e.g., MPEP 2173.06. Because the Examiner was unable to ascertain the metes and bounds (i.e., the scope) of the invention as claimed, the Examiner was unable to conduct a specific prior art search on the subject matter of these claims. However, the Examiner did perform a general art search and, to the extent the Examiner understands the invention as claimed, the Examiner has compared the claims to the prior art, citing several references pertinent:
Chen “FedGL: Federated Graph Learning Framework with Global Self-Supervision” (2021) (Abstract: Graph data are ubiquitous in the real world. Graph learning (GL) tries to mine and analyze graph data so that valuable information can be discovered. Existing GL methods are designed for centralized scenarios. However, in practical scenarios, graph data are usually distributed in different organizations, i.e., the curse of isolated data islands. To address this problem, we incorporate federated learning into GL and propose a general Federated Graph Learning framework FedGL, which is capable of obtaining a high-quality global graph model while protecting data privacy by discovering the global self-supervision information during the federated training. Concretely, we propose to upload the prediction results and node embeddings to the server for discovering the global pseudo label and global pseudo graph, which are distributed to each client to enrich the training labels and complement the graph structure respectively, thereby improving the quality of each local model. Moreover, the global self-supervision enables the information of each client to flow and share in a privacy-preserving manner, thus alleviating the heterogeneity and utilizing the complementarity of graph data among different clients. Finally, experimental results show that FedGL significantly outperforms baselines on four widely used graph datasets).
Liu “Client-Edge-Cloud Hierarchical Federated Learning” (2019) (Abstract: Federated Learning is a collaborative machine learning framework to train a deep learning model without accessing clients’ private data. Previous works assume one central parameter server either at the cloud or at the edge. The cloud server can access more data but with excessive communication overhead and long latency, while the edge server enjoys more efficient communications with the clients. To combine their advantages, we propose a client-edge-cloud hierarchical Federated Learning system, supported with a HierFAVG algorithm that allows multiple edge servers to perform partial model aggregation. In this way, the model can be trained faster and better communication-computation trade-offs can be achieved. Convergence analysis is provided for HierFAVG and the effects of key parameters are also investigated, which lead to qualitative design guidelines. Empirical experiments verify the analysis and demonstrate the benefits of this hierarchical architecture in different data distribution scenarios. Particularly, it is shown that by introducing the intermediate edge servers, the model training time and the energy consumption of the end devices can be simultaneously reduced compared to cloud-based Federated Learning).
Peng “FedNI: Federated Graph Learning with Network Inpainting for Population-Based Disease Prediction” (2021) (Abstract: Graph Convolutional Neural Networks (GCNs) are widely used for graph analysis. Specifically, in medical applications, GCNs can be used for disease prediction on a population graph, where graph nodes represent individuals and edges represent individual similarities. However, GCNs rely on a vast amount of data, which is challenging to collect for a single medical institution. In addition, a critical challenge that most medical institutions continue to face is addressing disease prediction in isolation with incomplete data information. To address these issues, Federated Learning (FL) allows isolated local institutions to collaboratively train a global model without data sharing. In this work, we propose a framework, FedNI, to leverage network inpainting and inter-institutional data via FL. Specifically, we first federatively train missing node and edge predictor using a graph generative adversarial network (GAN) to complete the missing information of local networks. Then we train a global GCN node classifier across institutions using a federated graph learning platform. The novel design enables us to build more accurate machine learning models by leveraging federated learning and also graph learning approaches. We demonstrate that our federated model outperforms local and baseline FL methods with significant margins on two public neuroimaging datasets).
Zhang “Subgraph Federated Learning with Missing Neighbor Generation” (2021) (Abstract: Graphs have been widely used in data mining and machine learning due to their unique representation of real-world objects and their interactions. As graphs are getting bigger and bigger nowadays, it is common to see their subgraphs separately collected and stored in multiple local systems. Therefore, it is natural to consider the subgraph federated learning setting, where each local system holds a small subgraph that may be biased from the distribution of the whole graph. Hence, the subgraph federated learning aims to collaboratively train a powerful and generalizable graph mining model without directly sharing their graph data. In this work, towards the novel yet realistic setting of subgraph federated learning, we propose two major techniques: (1) FedSage, which trains a GraphSage model based on FedAvg to integrate node features, link structures, and task labels on multiple local subgraphs; (2) FedSage+, which trains a missing neighbor generator along FedSage to deal with missing links across local subgraphs. Empirical results on four real-world graph datasets with synthesized subgraph federated learning settings demonstrate the effectiveness and efficiency of our proposed techniques. At the same time, consistent theoretical implications are made towards their generalization ability on the global graphs).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KUANG FU CHEN whose telephone number is (571)272-1393. The examiner can normally be reached M-F 9:00-5:30pm 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, Jennifer Welch can be reached on (571) 272-7212. 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.
/KC CHEN/Primary Patent Examiner, Art Unit 2143
Prosecution Insights