DETAILED ACTION
Status of the Application
The following is a Final Office Action. In response to Examiner's communication of August 27, 2025, Applicant, on January 29, 2026, amended claims 1-4, 6-9, 11, 12, 14, & 15. Claims 1-15 are now pending in this application and have been rejected below.
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . 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 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.
Response to Amendment
Applicant's amendments are sufficient to overcome the 35 USC 101 rejection of claim 14 set forth in the previous action for being directed to software per se. Therefore, this rejection of claim 14 for being directed to software per se is withdrawn.
Applicant's amendments are not sufficient to overcome the 35 USC 101 rejections set forth in the previous action for being directed to an abstract idea. Therefore, these rejections are updated and maintained below.
Applicant's amendments are not sufficient to overcome the prior art rejections set forth in the previous action. Therefore, these rejections are updated and maintained below.
Response to Arguments - 35 USC § 101
Applicant’s arguments and amendments with respect to the 35 USC 101 rejection of claim 14 for being directed to software per se have been fully considered, and they are persuasive. Therefore, the 35 USC 101 rejection of claim 14 for being directed to software per se are withdrawn.
Applicant’s arguments with respect to the 35 USC 101 rejections for being directed to an abstract idea have been fully considered, but they are not persuasive.
Applicant argues that the claims integrate any abstract idea into a practical application by reciting specific technological improvements to computer-functionality by reciting "wherein utilizing the relationships among the plurality of assets in the asset hierarchy as part of the solution learning process improves accuracy of the one or more model solutions compared to component-based learning without the relationships” and generating a knowledge graph representation with nodes and edges "wherein the knowledge graph representation facilitates more efficient querying compared to a relational database representation." Examiner respectfully disagrees.
Pursuant to 2019 Revised Patent Subject Matter Eligibility Guidance, in order to determine whether a claim is directed to an abstract idea, under Step 2A, we first (1) determine whether the claims recite limitations, individually or in combination, that fall within the enumerated subject matter groupings of abstract ideas (mathematical concepts, certain methods of organizing human activity, or mental processes), and (2) determine whether any additional elements beyond the recited abstract idea, individually and as an ordered combination, integrate the judicial exception into a practical application. 84 Fed. Reg. 52, 54-55. Next, if a claim (1) recites an abstract idea and (2) does not integrate that exception into a practical application, in order to determine whether the claim recites an “inventive concept,” under Step 2B, we then determine whether any of the additional elements beyond the recited abstract idea, individually and in combination, are significantly more than the abstract idea itself. 84 Fed. Reg. 56.
Under prong 1 of Step 2A, claim 1, and similarly claims 2-15, recites “generating an asset hierarchy from a plurality of assets, the asset hierarchy indicative of relationships among the plurality of assets from a lowest level to a highest level, wherein the relationships comprise one or more of physical relationships or logical relationships among the plurality of assets; executing a solution learning process to learn one or more model solutions for each of the plurality of assets based on the relationships among the plurality of assets in the asset hierarchy, wherein outputs of the one or more model solutions in lower levels of the asset hierarchy are utilized as inputs for the solution learning process to learn the one or more model solutions for each of the plurality of assets in higher levels of the asset hierarchy, wherein utilizing the relationships among the plurality of assets in the asset hierarchy as part of the solution learning process improves accuracy of the one or more model solutions compared to component-based learning without the relationships; generating a knowledge graph representation of the asset hierarchy, the knowledge graph representation comprising a plurality of nodes and a plurality of edges, each of the plurality of nodes representative of an asset from the plurality of assets and associated with the one or more model solutions for the asset, each of the plurality of edges representative of the relationships among the plurality of assets, wherein the knowledge graph representation facilitates more efficient querying …; and storing … the knowledge graph representation, the one or more model solutions for the each one of the plurality of assets, and knowledge for a solution explanation for the one or more model solutions.” Claims 1-15, in view of the claim limitations, recite the abstract idea of generating an asset hierarchy of assets, executing a solution learning process to learn solutions for each of the assets based on the relationships of assets in the asset hierarchy with solutions in lower levels are inputs to learn solutions for assets in higher levels, generating a knowledge graph representation of the asset hierarchy representing assets associated with solutions and relationships between assets allowing for querying the assets, and storing the knowledge graph, the solutions, and knowledge for a solution explanation the solutions.
A claim recites mental processes when the claim recites concepts performed in the human mind (including an observation, evaluation, judgment, opinion), wherein if the claim, under its broadest reasonable interpretation, covers the claim being practically performed in the mind but for the recitation of generic computer components, then the claim is in the mental process category. 84 Fed. Reg. 52 n.14. Here, as a whole, in view of the claim limitations, but for the computer components and systems performing the claimed functions, the broadest reasonable interpretation of the recited generating an asset hierarchy of assets, executing a solution learning process to learn solutions for each of the assets based on the relationships of assets in the asset hierarchy with solutions in lower levels are inputs to learn solutions for assets in higher levels, generating a knowledge graph representation of the asset hierarchy representing assets associated with solutions and relationships between assets allowing for querying the assets, and storing the knowledge graph, the solutions, and knowledge for a solution explanation the solutions could all be reasonably interpreted as a human making observations regarding assets and using judgement to decide an asset hierarchy, a human performing an evaluation and using judgment based on the observed information to learn solutions based for assets in higher levels of the hierarchy based on related assets in lower levels in the hierarchy, a human using their mind and/or pen and paper to generate a knowledge graph of the assets and query the assets in the graph, and a human using their mind to memorize or pen and paper to output and record the resulting knowledge graph, solutions, and solution explanation; therefore, the claims, including the claim elements referred to by Applicant, recite mental processes. Accordingly, since the claims, including the claim elements referred to by Applicant, recite mental processes, the claims recite an abstract idea under the first prong of Step 2A.
Specifically with respect to Applicant’s assertions that the above limitations recite an improvement in computer technology, the recitations of “wherein utilizing the relationships among the plurality of assets in the asset hierarchy as part of the solution learning process improves accuracy of the one or more model solutions compared to component-based learning without the relationships” and “wherein the knowledge graph representation facilitates more efficient querying,” these limitations are part of and directed to the abstract idea because a human can mentally evaluate information in the asset hierarchy regarding the assets and their relationships to use relationships among the assets as part of the solution learning process to accuracy of the one or more model solutions and human can perform an evaluation and use their mind and/or pen and paper to generate a knowledge graph of the assets and query the assets in the knowledge graph, and thus, these features recite a mental process.
Simply storing the knowledge graph in storage, querying the knowledge graph, and comparing the querying to querying a relational database is nothing more than applying the recited abstract idea with generic computer components and generally linking the abstract idea to a technical environment, namely the generic environment of a generic computer performing a query.
Querying a knowledge graph is not an improvement in a technology, but rather querying a knowledge graph is a technique that is well-known in the art as evinced by, for example, Jiao, et al. (US 20220067030 A1), hereinafter Jiao, at [0006], discussing evaluating the efficiency and accuracy of querying “well-known public knowledge graphs.”
The MPEP makes clear that "[m]ere automation of manual processes” is not an improvement in computer technology. See MPEP 2106.05(a). As in the claims at issue in Electric Power Group, the present claims are not focused on a specific improvement in computers or any other technology, but instead on certain independently abstract ideas that simply invokes computers as tools to implement the abstract idea. Electric Power Group, LLC v. Alstom S.A., et al., No. 2015-1778, slip op. at 8 (Fed. Cir. Aug. 1, 2016); MPEP 2106.05(a).
This judicial exception is not integrated into a practical application under the second prong of Step 2A. In particular, the claims recite the additional elements beyond the recited abstract idea of “[a] method, comprising,” “compared to a relational database representation,” and “in storage” in claim 1, “deep neural network” in claim 7, 11, & 12, “machine learning model algorithms” in claims 8 & 9, “[a] non-transitory computer-readable storage medium storing instructions for executing a process, the instructions comprising,” “compared to a relational database representation,” and “in storage” in claim 14, and “[a]n apparatus, comprising: a processor, configured to,” “compared to a relational database representation,” and “in storage” in claim 15; however, individually and when viewed as an ordered combination, and pursuant to the broadest reasonable interpretation, each of the additional elements are computing elements recited at high level of generality implementing the abstract idea on a computer (i.e. apply it), and thus, are no more than applying the abstract idea with generic computer components. Moreover, aside from the aforementioned additional elements, the remaining elements of dependent claims 2-13 do not integrate the abstract idea into a practical application because these claims merely recite further limitations that provide no more than simply narrowing the recited abstract idea.
Applicant argues that even if the claims were found directed to an abstract idea not integrated into a practical application, the knowledge graph representation with its specific structure is not well-understood, routine, or conventional, the specification identifies that convention relational databases are "ineffective" for this purpose, and the improvements in querying efficiency and model accuracy represent concrete technological benefits beyond the abstract idea. Examiner respectfully disagrees.
As noted above, a human can perform an evaluation and use their mind and/or pen and paper to generate a knowledge graph of the assets and query the assets in the knowledge graph, and thus, these features recite a mental process.
Simply storing the knowledge graph in storage, querying the knowledge graph, and comparing the querying to querying a relational database is nothing more than applying the recited abstract idea with generic computer components and generally linking the abstract idea to a technical environment, namely the generic environment of a generic computer performing a query.
Furthermore, querying a knowledge graph is not an improvement in a technology, but rather querying a knowledge graph is a technique that is well-known in the art as evinced by, for example, Jiao, et al. (US 20220067030 A1), hereinafter Jiao, at [0006], discussing evaluating the efficiency and accuracy of querying “well-known public knowledge graphs.”
The MPEP makes clear that "[m]ere automation of manual processes” is not an improvement in computer technology. See MPEP 2106.05(a). As in the claims at issue in Electric Power Group, the present claims are not focused on a specific improvement in computers or any other technology, but instead on certain independently abstract ideas that simply invokes computers as tools to implement the abstract idea. Electric Power Group, LLC v. Alstom S.A., et al., No. 2015-1778, slip op. at 8 (Fed. Cir. Aug. 1, 2016); MPEP 2106.05(a).
Under Step 2B, the claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception because, as noted above, the aforementioned additional elements beyond the recited abstract idea of “[a] method, comprising,” “compared to a relational database representation,” and “in storage” in claim 1, and similarly claims 14 and 15, as an order combination, are no more than mere instructions to implement the idea using generic computer components (i.e. apply it), and further, generally link the abstract idea to a field of use, which is not sufficient to amount to significantly more than an abstract idea; therefore, the additional elements are not sufficient to amount to significantly more than an abstract idea. Additionally, these recitations as an ordered combination, simply append the abstract idea to recitations of generic computer structure performing generic computer functions that are well-understood, routine, and conventional in the field as evinced by Applicant’s Specification at [0138] (describing the operations can be performed by software with the methods may be executed by a processor, such as a general-purpose computer, based on instructions stored on a computer-readable medium). Furthermore, as an ordered combination, these elements amount to generic computer components performing repetitive calculations, receiving or transmitting data over a network, which, as held by the courts, are well-understood, routine, and conventional. See MPEP 2106.05(d); July 2015 Update, p. 7.
Response to Arguments - Prior Art
Applicant’s arguments with respect to the prior art rejections have been fully considered, but they are not persuasive.
Applicant argues that Mathur fails to disclose “executing a solution learning process to learn one or more model solutions for each of the plurality of assets based on the relationships among the plurality of assets in the asset hierarchy, wherein outputs of the one or more model solutions in lower levels of the hierarchy are utilized as inputs for the solution learning process to learn the one or more model solutions for each of the plurality of assets in higher levels of the asset” as recited by the amended claims because Mathur does not disclose a solution learning process where model solutions are learned for assets at lower levels and the outputs of those model solutions are then used as inputs to learn model solutions for assets at higher levels in an iterative, hierarchical manner. Examiner respectfully disagrees.
Mathur, et al. (US 20200387804 A1), hereinafter Mathur, discloses in paragraph [0077] a set of nodes representing each representing devices and device states and a set of edges connecting two of the set of nodes represents one of a plurality types of relations, including indicating that one of the two devices as a child class that belongs to the other of the device classes of a parent class (i.e., assets in lower levels and assets in higher levels).
Mathur discloses “executing a solution learning process to learn one or more model solutions for each of the plurality of assets based on the relationships among the plurality of assets in the asset hierarchy, wherein outputs of the one or more model solutions in lower levels of the hierarchy are utilized as inputs for the solution learning process to learn the one or more model solutions for each of the plurality of assets in higher levels of the asset” using these nodes and parent and child relationships in paragraphs [0080]-[0082], [0085], wherein the server, in step 508, obtains rule data related to a plurality of issue resolution rules mapping at least one device state corresponding to a known issue to a resolution of the known issue, in step 510, receives a support bundle for a computer system that has encountered an unknown issue, and in step 512, identifies a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, including for a first device state of the one or more device states extracted from the support bundle represented by a first of the set of nodes, identifies an issue resolution rule that does not refer to the first device state but refers to a second device class represented by a second of the set of nodes such that the first node and the second node are connected through one or more edges that represents the third type of relation (i.e. outputs of model solutions for assets in lower levels are utilized as inputs to learn model solutions for assets in higher levels), and creates a new issue resolution rule that refers to the second device class (i.e., output from solutions from the lower level asset represented by the first node are used as input to learn one or more model solutions for the higher level asset represented by the second node).
Here, Mathur executes a learning process that learns a resolution, i.e., a solution, for an unknown issue of a second device represented by a second node from known resolution to a known issue of a first device represented by a first node connected to the first node by an edge representing a relationship between the first and second nodes by creating a new issue resolution rule that refers to the second device using the relationship between the nodes, and Mathur explicitly discloses the devices represented by the nodes can be connected through edges representing the nodes are related as parents, children, and siblings. Therefore, by creating a new issue resolution rule that refers to the second device class from the issue resolution of the first device based on edge representing their parent/child relationship, Mathur discloses the output from solutions from the lower-level asset represented by the first node are used as input to learn one or more model solutions for the higher-level asset represented by the second node.
In a particular example, in paragraphs [0065], [0070], when the server 102 receives receive a support bundle for a target computer system that has encountered an unknown issue, e.g., an issue resolution rule identified as being potentially applicable may refer to a first device state of a first device class represented by a first node connected through an edge representing “is coupled with” a second node representing a second device state of a second device class, the second node also connected via an edge representing “is child of” to a third node of a third device class, the third node also connected class through an edge representing the “is parent of” relation to a fourth node of a fourth device class, and the second node and the fourth node are then implicitly in a “is sibling of” relation, upon determining that the issue resolution rule otherwise matches the support bundle, the server 102 derives a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time. Here, by using the resolution from second node representing the second device and the “parent of” and “child of” between the nodes to learn the resolution for the fourth node - the outputs of model solutions for assets in lower levels, e.g., the first, second, and third nodes, are utilized as inputs to learn model solutions for assets in higher levels, e.g., the fourth node.
Accordingly, contrary to Applicant’s assertion, Mathur does indeed disclose “executing a solution learning process to learn one or more model solutions for each of the plurality of assets based on the relationships among the plurality of assets in the asset hierarchy, wherein outputs of the one or more model solutions in lower levels of the hierarchy are utilized as inputs for the solution learning process to learn the one or more model solutions for each of the plurality of assets in higher levels of the asset.”
Applicant asserts Mathur does not disclose "wherein utilizing the relationships among the plurality of assets in the asset hierarchy as part of the solution learning process improves accuracy of the one or more model solutions compared to component-based learning without the relationships" as recited by amended claim1 because Mathur does not compare its approach to component-based learning or disclose any improvement in model accuracy through a hierarchical relationship-based learning. Examiner respectfully disagrees.
Mathur explicitly states in paragraph [0028] that the server enables an increase in the efficiency and accuracy of determining the applicability of the issue resolution rules to the support bundle. As disclosed in paragraphs [0082], [0085], the server 102 identifies a specific issue resolution rule of the plurality of issue resolution rules applicable to the support bundle by identifying an issue resolution rule that does not refer to the first device state but refers to a second device class represented by a second of the set of nodes such that the first node and the second node are connected through one or more edges (i.e., utilizing the relationships among the plurality of assets), including one that represents the third type of relation, and the server 102 creates a new issue resolution rule that refers to the second device class (i.e., as part of the solution learning) Therefore, Mathur explicitly discloses that techniques employed by the server utilizing the relationships among the plurality of assets in the asset hierarchy (i.e. parent or child relationships of the first and second device and first and second node) as part of the solution learning (i.e., determining applicability of issue resolution) improves accuracy of the of the model solution.
Further, Examiner notes “improves accuracy of the one or more model solutions compared to component-based learning without the relationships” is the intended use and result of the use of the learned model solutions, which does not alter the structure or function of the claimed invention, and thus, since Mathur discloses utilizing relationships to learn the model solutions as claimed, the comparison to component-based learning without the relationships does not patentably distinguish the claimed invention from the cited prior art.
Moreover, since Mathur discloses learning model solutions and utilizing relationships to learn the model solutions as claimed, utilizing the relationships among the assets in the asset hierarchy as part of the solution learning process disclosed in Mathur also improves accuracy of the model solutions compared to component-based learning without the relationships.
Applicant argues that Mathur does not disclose "wherein the knowledge graph representation facilitates more efficient querying compared to a relational database representation" as recited by amended claim1 because while Mathur discloses a knowledge graph, it does not compare the querying efficiency of its knowledge graph to relational databases or disclose that the knowledge graph provides more efficient querying. Examiner respectfully disagrees.
Mathur explicitly states in paragraphs [0051]-[0052], the knowledge graph (KG) can support various queries, wherein based on the edges that represent the “is parent of” or “is child of” relations, a hierarchy of device classes can be obtained from the KG, based on the edges that represent the “is variant of” relations, a collection of devices classes that are variants of one another can be obtained from the KG, and the field or attribute values of each node or each edge (e.g., name, weight) can also be obtained from the KG through simple queries. Therefore, Mathur explicitly discloses the knowledge graph can support various simple and efficient queries.
Further, Examiner notes “facilitates more efficient querying compared to a relational database representation” is the intended use and result of the knowledge graph, which does not alter the structure or function of the claimed invention, and thus, since Mathur discloses the knowledge graph and using it to query information regarding the devices, the comparison to a relational database representation does not patentably distinguish the claimed invention from the cited prior art.
Moreover, since Mathur discloses the knowledge graph and using it to query information regarding the devices, the knowledge graph disclosed by Mathur also “facilitates more efficient querying compared to a relational database representation.”
Applicant argues the combination of Mathur and Creed fails to render obvious the claimed invention because neither reference, alone or in combination, teaches or suggests the hierarchical solution learning process recited in base claim 1. Examiner respectfully disagrees.
As noted above, Mathur discloses claim 1, and thus, the argument that the combination of Creed fails to render the claimed invention in claim 1 is moot.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-15 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. Claim 1, and similarly claims 2-15, recites “generating an asset hierarchy from a plurality of assets, the asset hierarchy indicative of relationships among the plurality of assets from a lowest level to a highest level, wherein the relationships comprise one or more of physical relationships or logical relationships among the plurality of assets; executing a solution learning process to learn one or more model solutions for each of the plurality of assets based on the relationships among the plurality of assets in the asset hierarchy, wherein outputs of the one or more model solutions in lower levels of the asset hierarchy are utilized as inputs for the solution learning process to learn the one or more model solutions for each of the plurality of assets in higher levels of the asset hierarchy, wherein utilizing the relationships among the plurality of assets in the asset hierarchy as part of the solution learning process improves accuracy of the one or more model solutions compared to component-based learning without the relationships; generating a knowledge graph representation of the asset hierarchy, the knowledge graph representation comprising a plurality of nodes and a plurality of edges, each of the plurality of nodes representative of an asset from the plurality of assets and associated with the one or more model solutions for the asset, each of the plurality of edges representative of the relationships among the plurality of assets, wherein the knowledge graph representation facilitates more efficient querying …; and storing … the knowledge graph representation, the one or more model solutions for the each one of the plurality of assets, and knowledge for a solution explanation for the one or more model solutions.” Claims 1-15, in view of the claim limitations, recite the abstract idea of generating an asset hierarchy of assets, executing a solution learning process to learn solutions for each of the assets based on the relationships of assets in the asset hierarchy with solutions in lower levels are inputs to learn solutions for assets in higher levels, generating a knowledge graph representation of the asset hierarchy representing assets associated with solutions and relationships between assets allowing for querying the assets, and storing the knowledge graph, the solutions, and knowledge for a solution explanation the solutions.
As a whole, in view of the claim limitations, but for the computer components and systems performing the claimed functions, the broadest reasonable interpretation of the recited generating an asset hierarchy of assets, executing a solution learning process to learn solutions for each of the assets based on the relationships of assets in the asset hierarchy with solutions in lower levels are inputs to learn solutions for assets in higher levels, generating a knowledge graph representation of the asset hierarchy representing assets associated with solutions and relationships between assets allowing for querying the assets, and storing the knowledge graph, the solutions, and knowledge for a solution explanation the solutions could all be reasonably interpreted as a human making observations regarding assets and using judgement to decide an asset hierarchy, a human performing an evaluation and using judgment based on the observed information to learn solutions based for assets in higher levels of the hierarchy based on related assets in lower levels in the hierarchy, a human using their mind and/or pen and paper to generate a knowledge graph of the assets and query the assets in the graph, and a human using their mind to memorize or pen and paper to output and record the resulting knowledge graph, solutions, and solution explanation; therefore, the claims recite mental processes. Further, with respect to the dependent claims, aside from the additional elements beyond the recited abstract idea addressed below under the second prong of Step 2A and 2B, the limitations of dependent claims 2-13 recite similar further abstract limitations to those discussed above that narrow the abstract idea recited in the independent claims because, aside from the computer components and systems performing the claimed functions the limitations of claims recite mental processes that can be practically performed mentally by observing, evaluating, and judging information mentally and/or with a pen and paper. Accordingly, since the claims recite mental processes, the claims recite an abstract idea under the first prong of Step 2A.
This judicial exception is not integrated into a practical application under the second prong of Step 2A. In particular, the claims recite the additional elements beyond the recited abstract idea of “[a] method, comprising,” “compared to a relational database representation,” and “in storage” in claim 1, “deep neural network” in claim 7, 11, & 12, “machine learning model algorithms” in claims 8 & 9, “[a] non-transitory computer-readable storage medium storing instructions for executing a process, the instructions comprising,” “compared to a relational database representation,” and “in storage” in claim 14, and “[a]n apparatus, comprising: a processor, configured to,” “compared to a relational database representation,” and “in storage” in claim 15; however, individually and when viewed as an ordered combination, and pursuant to the broadest reasonable interpretation, each of the additional elements are computing elements recited at high level of generality implementing the abstract idea on a computer (i.e. apply it), and thus, are no more than applying the abstract idea with generic computer components. Moreover, aside from the aforementioned additional elements, the remaining elements of dependent claims 2-13 do not integrate the abstract idea into a practical application because these claims merely recite further limitations that provide no more than simply narrowing the recited abstract idea.
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception under Step 2B. As noted above, the aforementioned additional elements beyond the recited abstract idea, as an order combination, are no more than mere instructions to implement the idea using generic computer components (i.e. apply it), and further, generally link the abstract idea to a field of use, which is not sufficient to amount to significantly more than an abstract idea; therefore, the additional elements are not sufficient to amount to significantly more than an abstract idea. Additionally, these recitations as an ordered combination, simply append the abstract idea to recitations of generic computer structure performing generic computer functions that are well-understood, routine, and conventional in the field as evinced by Applicant’s Specification at [0138] (describing the operations can be performed by software with the methods may be executed by a processor, such as a general-purpose computer, based on instructions stored on a computer-readable medium). Furthermore, as an ordered combination, these elements amount to generic computer components performing repetitive calculations, receiving or transmitting data over a network, which, as held by the courts, are well-understood, routine, and conventional. See MPEP 2106.05(d); July 2015 Update, p. 7. Moreover, aside from the aforementioned additional elements, the remaining elements of dependent claims 2-13 do not transform the recited abstract idea into a patent eligible invention because these claims merely recite further limitations that provide no more than simply narrowing the recited abstract idea.
Looking at these limitations as an ordered combination adds nothing additional that is sufficient to amount to significantly more than the recited abstract idea because they simply provide instructions to use a generic arrangement of generic computer components and recitations of generic computer structure that perform well-understood, routine, and conventional computer functions that are used to “apply” the recited abstract idea. Thus, the elements of the claims, considered both individually and as an ordered combination, are not sufficient to ensure that the claims as a whole amount to significantly more than the abstract idea itself. Since there are no limitations in these claims that transform the exception into a patent eligible application such that these claims amount to significantly more than the exception itself, claims 1-15 are rejected under 35 U.S.C. 101 as being directed to non-statutory subject matter.
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.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 1-5, 8, & 13-15 are rejected under 35 U.S.C. 102(a)(1), (a),(2) as being anticipated by Mathur, et al. (US 20200387804 A1), hereinafter Mathur.
Regarding claim 1, Mathur discloses a method (Abstract, [0024]), comprising:
generating an asset hierarchy from a plurality of assets, the asset hierarchy indicative of relationships among the plurality of assets from a lowest level to a highest level ([0075], in step 502, a server 102 is programmed or configured to receive an initial hierarchy for a plurality of device classes), wherein the relationships comprise one or more of physical relationships or logical relationships among the plurality of assets ([0049], hierarchical structure is imposed upon nodes through edges, e.g., a node representing a product type of a switch, and a first node that represents a device class at a higher level of the hierarchy can be connected to a second node via an edge that represents the “is parent of” relation and an edge that represents the “is child of” relation in the reverse direction, [0048], each directed edge connects two nodes representing two device classes represents a certain relation between the two device classes, e.g., a relation include “is coupled with” indicates a coupling via a physical component or a communication network within an IT environment, “is variant of” indicates a proximity in function except for a few features, “is substitute for” for two device classes having similar features);
executing a solution learning process to learn one or more model solutions for each of the plurality of assets based on the relationships among the plurality of assets in the asset hierarchy, wherein outputs of the one or more model solutions in lower levels of the asset hierarchy are utilized as inputs for the solution learning process to learn the one or more model solutions for each of the plurality of assets in higher levels of the asset hierarchy ([0077], in step 506, the server 102 creates a knowledge graph having a set of nodes and a set of edges based on the device data, wherein each of the set of edges connecting two of the set of nodes represents one of a plurality types of relations, including indicating that one of the two device classes as a child class belongs to the other of the device classes as a parent class (i.e., outputs of model solutions for assets in lower levels are utilized as inputs to learn model solutions for assets in higher levels), [0080]-[0082], [0085], in step 508, the server 102 obtains rule data related to a plurality of issue resolution rules mapping at least one device state corresponding to a known issue to a resolution of the known issue, in step 510, the server 102 receives a support bundle for a computer system that has encountered an unknown issue, and in step 512, identifies a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, including for a first device state of the one or more device states extracted from the support bundle, the first device state being of a first device class represented by a first of the set of nodes, identify an issue resolution rule that does not refer to the first device state but refers to a second device class represented by a second of the set of nodes such that the first node and the second node are connected through one or more edges (i.e. outputs of model solutions for assets in lower levels are utilized as inputs to learn model solutions for assets in higher levels), including one that represents the third type of relation, and the server 102 creates a new issue resolution rule that refers to the second device class (i.e., to learn one or more model solutions), [0042], the issue resolution rule application instructions 204 can enable creating additional issue resolution rules (i.e., to learn one or more model solutions) enhancing the KG based on information regarding the specific IT infrastructure where the unknown issue has arisen, [0065], [0070], when the server 102 receives receive a support bundle for a target computer system that has encountered an unknown issue, e.g., an issue resolution rule identified as being potentially applicable may refer to a first device state of a first device class represented by a first node connected through an edge representing “is coupled with” a second node representing a second device state of a second device class, the second node also connected via an edge representing “is child of” to a third node of a third device class, the third node also connected class through an edge representing the “is parent of” relation to a fourth node of a fourth device class, and the second node and the fourth node are then implicitly in a “is sibling of” relation, upon determining that the issue resolution rule otherwise matches the support bundle, the server 102 derives a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time (i.e., outputs of model solutions for assets in lower levels, e.g., the first, second, and third nodes, are utilized as inputs to learn model solutions for assets in higher levels, e.g., the fourth node)), wherein utilizing the relationships among the plurality of assets in the asset hierarchy as part of the solution learning process improves accuracy of the one or more model solutions compared to component-based learning without the relationships ([0028], the server enables an increase in the efficiency and accuracy of determining the applicability of the issue resolution rules to the support bundle, [0082], [0085], in step 512 the server 102 identifies a specific issue resolution rule of the plurality of issue resolution rules applicable to the support bundle by identifying an issue resolution rule that does not refer to the first device state but refers to a second device class represented by a second of the set of nodes such that the first node and the second node are connected through one or more edges (i.e. utilizing the relationships among the plurality of assets), including one that represents the third type of relation, and the server 102 creates a new issue resolution rule that refers to the second device class (i.e., as part of the solution learning); Examiner notes “improves accuracy of the one or more model solutions compared to component-based learning without the relationships” is the intended use and result of the use of the learned model solutions, which does not alter the structure or function of the claimed invention, and Mathur discloses learning model solutions as claimed and utilizes relationships to learn the model solutions, the comparison to component-based learning without the relationships does not patentably distinguish the claimed invention from the cited prior art); and
generating a knowledge graph representation of the asset hierarchy, the knowledge graph representation comprising a plurality of nodes and a plurality of edges, each of the of nodes representative of an asset from the plurality of asserts and associated with the one or more model solutions for the asset each of the plurality of edges representative of the relationships among the plurality of assets ([0077], in step 506, the server 102 creates a knowledge graph having a set of nodes and a set of edges based on the device data, wherein each of the set of edges connecting two of the set of nodes represents one of a plurality types of relations, including a third type indicating that one of the two device classes as a child class belongs to the other of the device classes as a parent class), wherein the knowledge graph representation facilitates more efficient querying compared to a relational database representation ([0051]-[0052], the KG can support various queries, wherein based on the edges that represent the “is parent of” or “is child of” relations, a hierarchy of device classes can be obtained from the KG, based on the edges that represent the “is variant of” relations, a collection of devices classes that are variants of one another can be obtained from the KG, and the field or attribute values of each node or each edge (e.g., name, weight) can also be obtained from the KG through simple queries; Examiner notes “facilitates more efficient querying compared to a relational database representation” is the intended use and result of the knowledge graph, which does not alter the structure or function of the claimed invention, and Mathur discloses the knowledge graph and using it to query information regarding the devices, the comparison to a relational database representation does not patentably distinguish the claimed invention from the cited prior art);
storing, in storage, the knowledge graph representation, the one or more model solutions for the each one of the plurality of assets, and knowledge for a solution explanation for the one or more model solutions ([0085], the server 102 creates a new issue resolution rule that refers to the second device class, [0037], the server 102 maintains a database of issue resolution rules, each mapping device states of the device classes that correspond to a known issue to a known resolution of the known issue, and [0044], the server database 220 manages storage of and access to relevant data, such as definitions of nodes and edges in the knowledge graph (KG), device data regarding one or more device classes, digital models for building the KG, the KG, issue resolution rules).
Regarding claim 2, Mathur discloses the method of claim 1 (as above), further comprising generating the solution explanation for each of the outputs of the one or more model solutions for the each of the plurality of assets from the highest level to the lowest level ([0077], in step 506, the server 102 creates a knowledge graph having a set of nodes and a set of edges based on the device data, wherein each of the set of edges connecting two of the set of nodes represents one of a plurality types of relations, including a third type indicating that one of the two device classes as a child class belongs to the other of the device classes as a parent class, [0081]-[0082], [0085], in step 510, the server 102 receives a support bundle for a computer system that has encountered an unknown issue, and in step 512, identifies a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, including for a first device state of the one or more device states extracted from the support bundle, the first device state being of a first device class represented by a first of the set of nodes, identify an issue resolution rule that does not refer to the first device state but refers to a second device class represented by a second of the set of nodes such that the first node and the second node are connected through one or more edges, including one that represents the third type of relation, and the server 102 creates a new issue resolution rule that refers to the second device class, [0065], [0070], when the server 102 receives receive a support bundle for a target computer system that has encountered an unknown issue, the server 102 derives a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time, and [0071], the server 102 can create a new issue resolution rule that refers to the device states of the vendor-specific device classes instances of which are part of the target computer system, and over time, create a new issue resolution rule that refers to the device states of higher-level device class that include the various vendor-specific device classes).
Regarding claim 3, Mathur discloses the method of claim 2 (as above), wherein the storing, in storage, the knowledge graph representation, the one or more model solutions for the each one of the plurality of assets, and the knowledge for the solution explanation for the one or more model solutions comprises:
generating a first knowledge graph, the first knowledge graph comprising a plurality of first nodes and a plurality of first edges, each of the plurality of first nodes representative of an asset from the plurality of assets and associated with the one or more model solutions for the asset from the plurality of assets, each of the plurality of first edges representative of the relationships among the plurality of assets ([0077], in step 506, the server 102 creates a knowledge graph having a set of nodes and a set of edges based on the device data, wherein each of the set of edges connecting two of the set of nodes represents one of a plurality types of relations, including a third type indicating that one of the two device classes as a child class belongs to the other of the device classes as a parent class, [0081]-[0082], [0085], in step 510, the server 102 receives a support bundle for a computer system that has encountered an unknown issue, and in step 512, identifies a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, including for a first device state of the one or more device states extracted from the support bundle, the first device state being of a first device class represented by a first of the set of nodes, identify an issue resolution rule that does not refer to the first device state but refers to a second device class represented by a second of the set of nodes such that the first node and the second node are connected through one or more edges, including one that represents the third type of relation, and the server 102 creates a new issue resolution rule that refers to the second device class, [0065], [0070], when the server 102 receives receive a support bundle for a target computer system that has encountered an unknown issue, the server 102 derives a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time, and [0071], the server 102 can create a new issue resolution rule that refers to the device states of the vendor-specific device classes instances of which are part of the target computer system, and over time, create a new issue resolution rule that refers to the device states of higher-level device class that include the various vendor-specific device classes);
generating a second knowledge graph ([0052], a new version of a switch may have fewer ports and is no longer compatible with an output device compared to the original version, and therefore, the change from the original version may be obtained by computing a graph difference between two portions of the KG corresponding to the two versions of the switch that includes a change in the value of the “number of ports” attribute of a node and an elimination of the edge representing the “is coupled with” relation with the node representing the output device, Claim 5, new device data related to a first device class of the hierarchy of device classes, the new device data indicating that the first device class is no longer compatible with a second device class of the hierarchy of device classes related to the first device class; creating a new version of the knowledge graph based on the new device data), the second knowledge graph comprising a plurality of second nodes and a plurality of second edges, each of the plurality of second nodes comprising the knowledge for the solution explanation of the one or more model solutions for the each one of the plurality of assets, and each of the plurality of second edges representing a relationship between the knowledge for the solution explanation for the one or more model solutions ([0077], in step 506, the server 102 creates a knowledge graph having a set of nodes and a set of edges based on the device data, wherein each of the set of edges connecting two of the set of nodes represents one of a plurality types of relations, including a third type indicating that one of the two device classes as a child class belongs to the other of the device classes as a parent class, [0081]-[0082], [0085], in step 510, the server 102 receives a support bundle for a computer system that has encountered an unknown issue, and in step 512, identifies a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, including for a first device state of the one or more device states extracted from the support bundle, the first device state being of a first device class represented by a first of the set of nodes, identify an issue resolution rule that does not refer to the first device state but refers to a second device class represented by a second of the set of nodes such that the first node and the second node are connected through one or more edges, including one that represents the third type of relation, and the server 102 creates a new issue resolution rule that refers to the second device class, [0065], [0070], when the server 102 receives receive a support bundle for a target computer system that has encountered an unknown issue, the server 102 derives a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time, and [0071], the server 102 can create a new issue resolution rule that refers to the device states of the vendor-specific device classes instances of which are part of the target computer system, and over time, create a new issue resolution rule that refers to the device states of higher-level device class that include the various vendor-specific device classes); and
storing the first knowledge graph and the second knowledge graph as the solution representation ([0044], the server database 220 is programmed or configured to manage storage of and access to relevant data, such as definitions of nodes and edges in the KG, device data regarding one or more device classes, digital models for building the KG, the KG, issue resolution rules).
Regarding claim 4, Mathur discloses the method of claim 2 (as above), wherein generating the solution explanation for the outputs of the one or more model solutions for the each of the plurality of assets from the highest level to the lowest level, comprises: determining a root cause for each output of the one or more model solutions based on one or more of execution of a trace-down process from a higher level to a lower level of the each of the plurality of assets, execution of an explaining scheme according to cross-level ones of the relationships among the plurality of assets, and execution of a learning scheme by using as the target the root cause derived from outputs of the one or more model solutions of the each of the plurality of assets; and incorporating the root cause as the knowledge for the solution explanation for the one or more model solutions ([0077], in step 506, the server 102 creates a knowledge graph having a set of nodes and a set of edges based on the device data, wherein each of the set of edges connecting two of the set of nodes represents one of a plurality types of relations, including a third type indicating that one of the two device classes as a child class belongs to the other of the device classes as a parent class, [0081]-[0082], [0085], in step 510, the server 102 receives a support bundle for a computer system that has encountered an unknown issue, and in step 512, identifies a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, including for a first device state of the one or more device states extracted from the support bundle, the first device state being of a first device class represented by a first of the set of nodes, identify an issue resolution rule that does not refer to the first device state but refers to a second device class represented by a second of the set of nodes such that the first node and the second node are connected through one or more edges, including one that represents the third type of relation, and the server 102 creates a new issue resolution rule that refers to the second device class, [0065], [0070], when the server 102 receives receive a support bundle for a target computer system that has encountered an unknown issue, the server 102 derives a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time, and [0071], the server 102 can create a new issue resolution rule that refers to the device states of the vendor-specific device classes instances of which are part of the target computer system, and over time, create a new issue resolution rule that refers to the device states of higher-level device class that include the various vendor-specific device classes);
Regarding claim 5, Mathur discloses the method of claim 1 (as above), wherein the solution learning process comprises: learning the one or more model solutions for the each of the plurality of assets from the lowest level first; wherein the outputs of the one or more model solutions of lower levels are utilized as inputs to learn the one or more model solutions in higher levels of the asset hierarchy in an iterative manner from the lowest level to the highest level ([0077], in step 506, the server 102 creates a knowledge graph having a set of nodes and a set of edges based on the device data, wherein each of the set of edges connecting two of the set of nodes represents one of a plurality types of relations, including a third type indicating that one of the two device classes as a child class belongs to the other of the device classes as a parent class, [0081]-[0082], [0085], in step 510, the server 102 receives a support bundle for a computer system that has encountered an unknown issue, and in step 512, identifies a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, including for a first device state of the one or more device states extracted from the support bundle, the first device state being of a first device class represented by a first of the set of nodes, identify an issue resolution rule that does not refer to the first device state but refers to a second device class represented by a second of the set of nodes such that the first node and the second node are connected through one or more edges, including one that represents the third type of relation, and the server 102 creates a new issue resolution rule that refers to the second device class, [0065], [0070], when the server 102 receives receive a support bundle for a target computer system that has encountered an unknown issue, the server 102 derives a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time, and [0071], the server 102 can create a new issue resolution rule that refers to the device states of the vendor-specific device classes instances of which are part of the target computer system, and over time, create a new issue resolution rule that refers to the device states of higher-level device class that include the various vendor-specific device classes).
Regarding claim 8, Mathur discloses the method of claim 5 (as above), wherein the solution learning process further comprises: using model performance metrics from the one or more model solutions at the lower levels as the input to learn the one or more model solutions at the higher levels ([0048], each directed edge in the KG can have a weight that represents the strength of or the confidence score for the represented relation (i.e. model performance), [0072], the server 102 is programmed to prepare a recommendation for resolving the unknown issue encountered by the target computer system based on the applicable issue resolution rules by ranking these issue resolution rules can based on the extent of applicability to the support bundle (i.e. weight));
wherein each of the plurality of assets is associated with the one or more model solutions for one or more tasks ([0061], [0072], the server 102 the server 102 is programmed to builds or receives a set of issue resolution rules, wherein an issue resolution rule generally maps a list of device states of a computer system that correspond to a known issue (an abnormality or a failure) encountered by the computer system to a resolution of the known issue, which then generally involves altering one of more of the device states, and where the computer devices in the computer system or the corresponding device classes are represented as nodes and the device states correspond to the attributes of the nodes or the edges in the KG taking on specific values, the issue resolution rule can be tied to the KG, and the server 102 recommendations for resolving the unknown issue encountered by the target computer system based on the applicable issue resolution rules by e.g., including only the known resolutions from a number of highest-ranking issue resolution rules in the recommendation);
wherein each of the plurality of assets is associated with one or more versions of the one or more model solutions for each of the one or more tasks ([0072], different issue resolution rules may conflict, e.g., one known resolution might require changing the maximum number of open sessions of an output device from ten to twenty, while another might require changing the maximum number to five or uncoupling the output device);
wherein the one or more model solutions ([0069]-[0070], the server 102 can be programmed to derive a new rule that refers to a third device state of the third device class instead of the second device state as an alternative to traversing the edge representing the “is variant of ” relation of the KG next time, and the server 102 can be programmed to derive a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time) are based on one or more of machine learning model algorithms or physics-based models ([0049], hierarchical structure can be imposed upon the nodes in the KG through the edges, e.g., there can be a node representing a product type of a switch, and a first node that represents a device class at a higher level of the hierarchy can be connected to a second node via an edge that represents the “is parent of” relation and an edge that represents the “is child of” relation in the reverse direction, [0069], node 306 that represents the load balancer version 0.22 is also connected to the node 302 through the edge 312 that represents a “is variant of” relation with a strength of 0.9 indicating how functionally or physically related the two variants are, [0048], each directed edge in the KG that connects two nodes representing two device classes represents a certain relation between the two device classes, e.g., “is coupled with” corresponds to a coupling via a physical component or a communication network within an IT environment, “is variant of” indicates a proximity in function except for a few features, and “is substitute for” is for two device classes having similar features but made by different vendors);
wherein the one or more model solutions are configured to identify and utilize fault tolerance relationships among one or more of sensors or assets in the asset hierarchy ([0055], the server 102 is programmed to further identify items from the device data that may correspond to the nodes or the attributes thereof in the KG based on named-entity recognition (NER) , the server can define domain-specific entities, such as product names, versions, technology-related terms (e.g., IP, network, packet), persons, locations, and organizations, e.g., the name of a device class consisting of four tokens “Cisco 2900 series routers” can be detected and classified based on the presence of the keyword “Cisco” or “router,” and the output of the NER technique may include a confidence score for a determined entity, which may be encoded in the KG as an attribute of a node or an edge and used to rank the nodes or edges), where some of the one or more of sensors or assets are configured to have a similar function or role in a system ([0079], the server 102 can be configured to compute a weight for each of the set of edges, wherein the weight can represent a frequency of coupling for an edge that represents the first type of relation, a functional similarity between the variants for an edge that represents the second type of relation, and a degree of distribution of the parent class for an edge that represents the third type of relation);
wherein the one or more model solutions are configured to capture and utilize cross-level relationships among assets, where the input for the one or more model solutions for an asset is from one or more of the assets or the sensors at differing lower levels ([0077], in step 506, the server 102 creates a knowledge graph having a set of nodes and a set of edges based on the device data, wherein each of the set of edges connecting two of the set of nodes represents one of a plurality types of relations, including a third type indicating that one of the two device classes as a child class belongs to the other of the device classes as a parent class, [0081]-[0082], [0085], in step 510, the server 102 receives a support bundle for a computer system that has encountered an unknown issue, and in step 512, identifies a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, including for a first device state of the one or more device states extracted from the support bundle, the first device state being of a first device class represented by a first of the set of nodes, identify an issue resolution rule that does not refer to the first device state but refers to a second device class represented by a second of the set of nodes such that the first node and the second node are connected through one or more edges, including one that represents the third type of relation, and the server 102 creates a new issue resolution rule that refers to the second device class, [0065], [0070], when the server 102 receives receive a support bundle for a target computer system that has encountered an unknown issue, the server 102 derives a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time, and [0071], the server 102 can create a new issue resolution rule that refers to the device states of the vendor-specific device classes instances of which are part of the target computer system, and over time, create a new issue resolution rule that refers to the device states of higher-level device class that include the various vendor-specific device classes); and
wherein the asset hierarchy is refined by removing connections based on the feature importance in the one or more model solutions ([0066], the server 102 is programmed to determine which issue resolution rules to apply, wherein to minimize the number of issue resolution rules for consideration, the server 102 includes only those edges that represent one or more specific types of relations with weights above a certain threshold and consider only those issue resolution rules that are associated with the nodes in the resulting subset of the KG in terms of applicability to the support bundle).
Regarding claim 13, Mathur discloses the method of claim 1 (as above), wherein the asset hierarchy is representative of one or more of a physical hierarchy or a logical hierarchy of the plurality of assets ([0049], hierarchical structure can be imposed upon the nodes in the KG through the edges, e.g., there can be a node representing a product type of a switch, and a first node that represents a device class at a higher level of the hierarchy can be connected to a second node via an edge that represents the “is parent of” relation and an edge that represents the “is child of” relation in the reverse direction, [0069], node 306 that represents the load balancer version 0.22 is also connected to the node 302 through the edge 312 that represents a “is variant of” relation with a strength of 0.9 indicating how functionally or physically related the two variants are, [0048], each directed edge in the KG that connects two nodes representing two device classes represents a certain relation between the two device classes, e.g., “is coupled with” corresponds to a coupling via a physical component or a communication network within an IT environment, “is variant of” indicates a proximity in function except for a few features, and “is substitute for” is for two device classes having similar features but made by different vendors).
Regarding claim 14, this claim is substantially similar to claim 1, and is, therefore, rejected on the same basis as claim 1. While claim 14 is directed to a computer program storing instructions for executing a process, Mathur discloses a computer program, as claimed. [0089]-[0093].
Regarding claim 15, this claim is substantially similar to claim 1, and is, therefore, rejected on the same basis as claim 1. While claim 15 is directed to an apparatus comprising a processor, Mathur discloses an apparatus, as claimed. [0089]-[0093].
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 6, 7, & 9-12 are rejected under 35 U.S.C. 103 as being unpatentable over Mathur, et al. (US 20200387804 A1), hereinafter Mathur, in view of Creed, et al. (US 20210081717 A1), hereinafter Creed.
Regarding claim 6, Mathur discloses the method of claim 1 (as above). Further, while Mathur discloses all of the above and wherein the solution learning process comprises: calculating model performance metrics for the one or more model solutions and a weight for each of the inputs to the each of the plurality of assets ([0048], each directed edge in the KG can have a weight that represents the strength of or the confidence score for the represented relation (i.e. model performance), [0072], the server 102 is programmed to prepare a recommendation for resolving the unknown issue encountered by the target computer system based on the applicable issue resolution rules by ranking these issue resolution rules can based on the extent of applicability to the support bundle (i.e. weight));
for the model performance metrics meeting a success criteria, proceeding with the solution learning process to a next one of the each of the plurality of assets ([0066], the server 102 is programmed to determine which issue resolution rules to apply, wherein to minimize the number of issue resolution rules for consideration, the server 102 includes only those edges that represent one or more specific types of relations with weights above a certain threshold and consider only those issue resolution rules that are associated with the nodes in the resulting subset of the KG in terms of applicability to the support bundle);
for the model performance metrics not meeting the success criteria: traversing, from the each of the plurality of assets, ones of the plurality of assets that are at the lower levels ([0070], the server 102 can be programmed to derive a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time) in the asset hierarchy in a descending order of weights at each level ([0072], the server 102 is programmed to prepare a recommendation for resolving the unknown issue encountered by the target computer system based on the applicable issue resolution rules, and these issue resolution rules can be ranked based on the extent of applicability to the support bundle, e.g., an exact match of the device states referred to in an issue resolution rule may ranked higher than a match that requires expanding the scope of the issue resolution rule via various attributes of or edges connecting the nodes in the KG, and, the server 102 can be configured to include only the known resolutions from a number of highest-ranking issue resolution rules in the recommendation, Claims 1 & 13, a computer-implemented method of resolving infrastructure issues receives a support bundle for a computer system that has encountered an unknown issue and identifying a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, wherein the identifying comprises determining multiple issue resolution rules of the plurality of issue resolution rules that are applicable to the support bundle based on the knowledge graph; assigning ranks to the multiple issue resolution rules based on an extent of applicability to the support bundle, the specific issue resolution rule having one of highest ranks among the multiple issue resolution rules);
for each of the ones of the traversed plurality of assets, executing a broader set of model algorithms and parameter sets to generate a plurality of a model solution ([0072], the resolutions from different issue resolution rules are in conflict, and the server 102 can be configured to present a list of choices for user selection or similarly rank the resolutions based on how well the corresponding issue resolution rules match the support bundle, Claim 13, identifying comprises assigning ranks to the multiple issue resolution rules based on an extent of applicability to the support bundle, the specific issue resolution rule having one of highest ranks among the multiple issue resolution rules, [0071], the server 102 can create a new issue resolution rule that refers to the device states of the vendor-specific device classes instances of which are part of the target computer system, and over time, create a new issue resolution rule that refers to the device states of higher-level device class that include the various vendor-specific device classes), Mathur does not expressly disclose the remaining elements of the following limitations, including applying hyperparameter optimization to the plurality of model solutions to select the model solution, which however, are taught by further teachings in Creed.
Creed teaches calculating model performance metrics for the one or more model solutions and a weight for each of the inputs to the each of the plurality of assets (Abstract, a method for generating a graph neural network (GNN) model based on an entity-entity graph comprises generating an embedding based on data representative of the entity-entity graph for the GNN model, wherein the embedding comprises an attention weight assigned to each relationship edge of the entity-entity graph; and updating weights of the GNN model including the attention weights by minimising a loss function associated with at least the embedding);
for the model performance metrics meeting a success criteria, proceeding with the solution learning process to a next one of the each of the plurality of assets;
for the model performance metrics not meeting the success criteria: traversing ([0011], the computer-implemented method further comprising repeating the steps of generating the embedding and/or scoring, and updating the weights until the GNN model is determined to be trained), from the each of the plurality of assets, ones of the plurality of assets that are at the lower levels in the asset [based on] order of weights at each level ([0114], in step 134, generating a filtered entity-entity relationship graph by weighting each edge with the corresponding trained attention weight, wherein filtering the entity-entity graph based on the trained attention weights may include removing those relationship edge(s) from the entity-entity graph or portion thereof having a corresponding trained attention weight that is below or equal or above or equal to an attention relevancy threshold, and additionally or alternatively, step 134 may further include identifying uncertain relationship edges in the entity-entity graph or portion thereof based on a set of attention weights retrieved from the trained GNN model associated with the entity-entity graph and an attention relevancy threshold, and the identified uncertain relationship edges may be culled and rules may be derived defining actions for excluding the relationship edges);
for each of the ones of the traversed plurality of assets, executing a broader set of model algorithms and parameter sets to generate a plurality of model solutions ([0008], [0114], the entity-entity graph may be filtered based on the attention weights of a trained GNN model, the filtered entity-entity graph may be used to update the GNN model or train another GNN model, and the trained GNN model may be used to predict link relationship between a first entity and a second entity associated with the entity-entity graph), and applying hyperparameter optimization to the plurality of model solutions to select the model solution ([0112], [0134], the loss function includes a hyperparameter, and [0165], hyperparameters are optimized, [0106]-[0107], the process of generating the GNN model includes 118, updating the GNN model including the attention weights based on the computed loss by minimising the GNN loss function, step 120, it is determined whether further training of the graph model is necessary, and if so (e.g. Y), then repeating step 112 to118, including computing the loss and updating weights based on the loos function).
Mathur and Creed are analogous fields of invention because both address the problem of generating a graphical structure with nodes and edges linking nodes to each other and weights indicting the relationship between nodes. At the time the invention was effectively filed, it would have been obvious to one of ordinary skill in the art to include in the system of Mathur the ability to apply hyperparameter optimization as taught by Creed since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art would have recognized that the combination would produce the predictable results of applying hyperparameter optimization, as claimed. Further, it would have been obvious to one of ordinary skill in the art to have modified Mathur with the aforementioned teachings of Creed in order to produce the added benefit of improving knowledge graph quality. [0170].
Regarding claim 7, Mathur discloses the method of claim 1 (as above). Further, while Mathur discloses all of the above and wherein the solution learning process comprises: generating a … network to represent the asset hierarchy ([0049], a hierarchical structure can be imposed upon the nodes in the KG through the edges), the … network comprising: … representative of sensors associated with the plurality of assets ([0055], the server 102 identifies items from the device data that may correspond to the nodes or the attributes thereof in the KG, the server 102 can define domain-specific entities, such as product names, versions, technology-related terms (e.g., IP, network, packet), persons, locations, and organizations, e.g., the name of a device class consisting of four tokens “Cisco 2900 series routers” can be detected and classified based on the presence of the keyword “Cisco” or “router”); … representative of ones of the plurality of assets at the highest level in the asset hierarchy; and … representative of assets at other levels in the asset hierarchy; wherein connections between the … layers represent one or more of a physical or a logical relationship in the asset hierarchy ([0049], a first node that represents a device class at a higher level of the hierarchy can be connected to a second node via an edge that represents the “is parent of” relation and an edge that represents the “is child of” relation in the reverse direction), Mathur does not expressly disclose the remaining elements of the following limitations, including a deep neural network, which however, are taught by further teachings in Creed.
Creed teaches wherein the solution learning process comprises: generating a deep neural network to represent the asset hierarchy ([0084], NNs or NN structures that may be used by the invention as described herein may include at least one or more neural network structures from the group of: artificial NNs (ANNs); deep NNs; deep learning; deep learning ANNs; deep belief networks; deep Boltzmann machines), the deep neural network comprising: an input layer … associated with the plurality of assets; an output layer …; and one or more hidden layers … ([0096], neural networks are used to generate graph models by using a loss function to update weights of hidden layers of the GNN model); wherein connections between the input layer, the output layer, and the one or more hidden layers of the deep neural network layers represent … (Abstract, the method of generating a graph neural network (GNN) model based on an entity-entity graph comprises generating an embedding based on data representative of the entity-entity graph for the GNN model comprising an attention weight assigned to each relationship edge of the entity-entity graph indicating the relevancy of each relationship edge between entity nodes of the entity-entity graph, [0096], neural networks are used to generate graph models by using a loss function to update weights of hidden layers of the GNN model, [0108], data representative of a first entity, a second entity and a relationship may be input to the GNN, and the GNN may output a relationship score indicating whether the first and second entity have the relationship).
Mathur and Creed are analogous fields of invention because both address the problem of generating a graphical structure with nodes and edges linking nodes to each other and weights indicting the relationship between nodes. At the time the invention was effectively filed, it would have been obvious to one of ordinary skill in the art to include in the system of Mathur the ability to generate a deep learning network as taught by Creed since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art would have recognized that the combination would produce the predictable results of generating a deep learning network, as claimed. Further, it would have been obvious to one of ordinary skill in the art to have modified Mathur with the aforementioned teachings of Creed in order to produce the added benefit of improving knowledge graph quality. [0170].
Regarding claim 9, the combined teaches of Mathur and Creed teaches the method of claim 6 (as above). Further, Mathur discloses wherein the solution learning process further comprises: using the model performance metrics from the one or more model solutions at the lower levels as the input to learn the one or more model solutions at the higher levels ([0048], each directed edge in the KG can have a weight that represents the strength of or the confidence score for the represented relation (i.e. model performance), [0072], the server 102 is programmed to prepare a recommendation for resolving the unknown issue encountered by the target computer system based on the applicable issue resolution rules by ranking these issue resolution rules can based on the extent of applicability to the support bundle (i.e. weight));
wherein each of the plurality of assets is associated with the one or more model solutions for one or more tasks ([0061], [0072], the server 102 the server 102 is programmed to builds or receives a set of issue resolution rules, wherein an issue resolution rule generally maps a list of device states of a computer system that correspond to a known issue (an abnormality or a failure) encountered by the computer system to a resolution of the known issue, which then generally involves altering one of more of the device states, and where the computer devices in the computer system or the corresponding device classes are represented as nodes and the device states correspond to the attributes of the nodes or the edges in the KG taking on specific values, the issue resolution rule can be tied to the KG, and the server 102 recommendations for resolving the unknown issue encountered by the target computer system based on the applicable issue resolution rules by e.g., including only the known resolutions from a number of highest-ranking issue resolution rules in the recommendation);
wherein each of the plurality of assets is associated with one or more versions of the one or more model solutions for each of the one or more tasks ([0072], different issue resolution rules may conflict, e.g., one known resolution might require changing the maximum number of open sessions of an output device from ten to twenty, while another might require changing the maximum number to five or uncoupling the output device);
wherein the one or more model solutions ([0069]-[0070], the server 102 can be programmed to derive a new rule that refers to a third device state of the third device class instead of the second device state as an alternative to traversing the edge representing the “is variant of ” relation of the KG next time, and the server 102 can be programmed to derive a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time) are based on one or more of machine learning model algorithms or physics-based models ([0049], hierarchical structure can be imposed upon the nodes in the KG through the edges, e.g., there can be a node representing a product type of a switch, and a first node that represents a device class at a higher level of the hierarchy can be connected to a second node via an edge that represents the “is parent of” relation and an edge that represents the “is child of” relation in the reverse direction, [0069], node 306 that represents the load balancer version 0.22 is also connected to the node 302 through the edge 312 that represents a “is variant of” relation with a strength of 0.9 indicating how functionally or physically related the two variants are, [0048], each directed edge in the KG that connects two nodes representing two device classes represents a certain relation between the two device classes, e.g., “is coupled with” corresponds to a coupling via a physical component or a communication network within an IT environment, “is variant of” indicates a proximity in function except for a few features, and “is substitute for” is for two device classes having similar features but made by different vendors);
wherein the one or more model solutions are configured to identify and utilize fault tolerance relationships among one or more of sensors or assets in the asset hierarchy ([0055], the server 102 is programmed to further identify items from the device data that may correspond to the nodes or the attributes thereof in the KG based on named-entity recognition (NER) , the server can define domain-specific entities, such as product names, versions, technology-related terms (e.g., IP, network, packet), persons, locations, and organizations, e.g., the name of a device class consisting of four tokens “Cisco 2900 series routers” can be detected and classified based on the presence of the keyword “Cisco” or “router,” and the output of the NER technique may include a confidence score for a determined entity, which may be encoded in the KG as an attribute of a node or an edge and used to rank the nodes or edges), wherein some of the one or more of sensors or assets are configured to have a similar function or role in a system ([0079], the server 102 can be configured to compute a weight for each of the set of edges, wherein the weight can represent a frequency of coupling for an edge that represents the first type of relation, a functional similarity between the variants for an edge that represents the second type of relation, and a degree of distribution of the parent class for an edge that represents the third type of relation);
wherein the one or more model solutions are configured to capture and utilize cross-level relationships among assets, where the input for the one or more model solutions for an asset is from one or more of the assets or the sensors at differing lower levels ([0077], in step 506, the server 102 creates a knowledge graph having a set of nodes and a set of edges based on the device data, wherein each of the set of edges connecting two of the set of nodes represents one of a plurality types of relations, including a third type indicating that one of the two device classes as a child class belongs to the other of the device classes as a parent class, [0081]-[0082], [0085], in step 510, the server 102 receives a support bundle for a computer system that has encountered an unknown issue, and in step 512, identifies a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, including for a first device state of the one or more device states extracted from the support bundle, the first device state being of a first device class represented by a first of the set of nodes, identify an issue resolution rule that does not refer to the first device state but refers to a second device class represented by a second of the set of nodes such that the first node and the second node are connected through one or more edges, including one that represents the third type of relation, and the server 102 creates a new issue resolution rule that refers to the second device class, [0065], [0070], when the server 102 receives receive a support bundle for a target computer system that has encountered an unknown issue, the server 102 derives a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time, and [0071], the server 102 can create a new issue resolution rule that refers to the device states of the vendor-specific device classes instances of which are part of the target computer system, and over time, create a new issue resolution rule that refers to the device states of higher-level device class that include the various vendor-specific device classes); and
wherein the asset hierarchy is refined by removing connections based on the feature importance in the one or more model solutions ([0066], the server 102 is programmed to determine which issue resolution rules to apply, wherein to minimize the number of issue resolution rules for consideration, the server 102 includes only those edges that represent one or more specific types of relations with weights above a certain threshold and consider only those issue resolution rules that are associated with the nodes in the resulting subset of the KG in terms of applicability to the support bundle).
Regarding claim 10, the combined teaches of Mathur and Creed teaches the method of claim 6 (as above). Further, Mathur discloses wherein the solution learning process further comprises: using a traversal algorithm to traverse ones of the plurality of assets below a current asset by following the descending order of the weights; and using the weights on the connections to limit ones of the plurality of assets to be traversed ([0072], the server 102 is programmed to prepare a recommendation for resolving the unknown issue encountered by the target computer system based on the applicable issue resolution rules, and these issue resolution rules can be ranked based on the extent of applicability to the support bundle, e.g., an exact match of the device states referred to in an issue resolution rule may ranked higher than a match that requires expanding the scope of the issue resolution rule via various attributes of or edges connecting the nodes in the KG, and, the server 102 can be configured to include only the known resolutions from a number of highest-ranking issue resolution rules in the recommendation, Claims 1 & 13, a computer-implemented method of resolving infrastructure issues receives a support bundle for a computer system that has encountered an unknown issue and identifying a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, wherein the identifying comprises determining multiple issue resolution rules of the plurality of issue resolution rules that are applicable to the support bundle based on the knowledge graph; assigning ranks to the multiple issue resolution rules based on an extent of applicability to the support bundle, the specific issue resolution rule having one of highest ranks among the multiple issue resolution rules).
Regarding claim 11, the combined teaches of Mathur and Creed teaches the method of claim 7 (as above). Further, while Mathur discloses all of the above and wherein the solution learning process further comprises: building the … network to generate multiple outputs with each output for an asset from the plurality of assets in the asset hierarchy ([0072], to prepare a recommendation for resolving the unknown issue encountered by the target computer system based on the applicable issue resolution rules, the resolutions from different issue resolution rules are in conflict, and the server 102 can be configured to present a list of choices for user selection or similarly rank the resolutions based on how well the corresponding issue resolution rules match the support bundle, Claim 13, identifying comprises assigning ranks to the multiple issue resolution rules based on an extent of applicability to the support bundle, the specific issue resolution rule having one of highest ranks among the multiple issue resolution rules, [0071], the server 102 can create a new issue resolution rule that refers to the device states of the vendor-specific device classes instances of which are part of the target computer system, and over time, create a new issue resolution rule that refers to the device states of higher-level device class that include the various vendor-specific device classes); wherein each of the plurality of assets is associated with the one or more model solutions for one or more tasks; and the one or more model solutions are configured to capture and utilize cross-level relationships among the plurality of assets, wherein the connections between … comprise links that connect pairs of assets in non-adjacent layers in the … network ([0077], in step 506, the server 102 creates a knowledge graph having a set of nodes and a set of edges based on the device data, wherein each of the set of edges connecting two of the set of nodes represents one of a plurality types of relations, including a third type indicating that one of the two device classes as a child class belongs to the other of the device classes as a parent class (i.e., outputs of model solutions for assets in lower levels are utilized as inputs to learn model solutions for assets in higher levels), [0081]-[0082], [0085], in step 510, the server 102 receives a support bundle for a computer system that has encountered an unknown issue, and in step 512, identifies a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, including for a first device state of the one or more device states extracted from the support bundle, the first device state being of a first device class represented by a first of the set of nodes, identify an issue resolution rule that does not refer to the first device state but refers to a second device class represented by a second of the set of nodes such that the first node and the second node are connected through one or more edges, including one that represents the third type of relation, and the server 102 creates a new issue resolution rule that refers to the second device class (i.e., outputs of model solutions for assets in lower levels are utilized as inputs to learn model solutions for assets in higher levels), [0042], the issue resolution rule application instructions 204 can enable creating additional issue resolution rules enhancing the KG based on information regarding the specific IT infrastructure where the unknown issue has arisen, [0065], [0070], when the server 102 receives receive a support bundle for a target computer system that has encountered an unknown issue, the server 102 derives a new rule that refers to the fourth device state instead of the second device state as an alternative to traversing the “is child of” and “is parent of” edges of the KG next time), Mathur does not expressly disclose the remaining elements of the following limitations, including a deep neural network, which however, are taught by further teachings in Creed.
Creed teaches building the deep neural network to generate multiple outputs …([0084], NNs or NN structures that may be used by the invention as described herein may include at least one or more neural network structures from the group of: artificial NNs (ANNs); deep NNs; deep learning; deep learning ANNs; deep belief networks; deep Boltzmann machines);
wherein the connections between the input layer, the output layer, and the one or more hidden layers of the deep neural network comprise links that connect pairs of … in non-adjacent layers in the deep neural network ([0096], neural networks are used to generate graph models by using a loss function to update weights of hidden layers of the GNN model, Abstract, the method of generating a graph neural network (GNN) model based on an entity-entity graph comprises generating an embedding based on data representative of the entity-entity graph for the GNN model comprising an attention weight assigned to each relationship edge of the entity-entity graph indicating the relevancy of each relationship edge between entity nodes of the entity-entity graph, [0082], relationship connections between entity-entity nodes and accuracy of current relationship connections of the entity-entity knowledge graph may be determined by training the GCNN on the entity-entity knowledge graph to generate an embedding/representation of the entity-entity graph in the form of a graph model, [0108], data representative of a first entity, a second entity and a relationship may be input to the GNN, and the GNN may output a relationship score indicating whether the first and second entity have the relationship).
Mathur and Creed are analogous fields of invention because both address the problem of generating a graphical structure with nodes and edges linking nodes to each other and weights indicting the relationship between nodes. At the time the invention was effectively filed, it would have been obvious to one of ordinary skill in the art to include in the system of Mathur the ability to generate a deep learning network as taught by Creed since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art would have recognized that the combination would produce the predictable results of generating a deep learning network, as claimed. Further, it would have been obvious to one of ordinary skill in the art to have modified Mathur with the aforementioned teachings of Creed in order to produce the added benefit of improving knowledge graph quality. [0170].
Regarding claim 12, Mathur discloses the method of claim 1 (as above). Further, while Mathur discloses all of the above and further comprising generating the asset hierarchy through a … scheme, the generating the asset hierarchy comprising: identifying ones of the plurality of assets at each level ([0049], hierarchical structure can be imposed upon the nodes in the KG through the edges, e.g., there can be a node representing a product type of a switch, and a first node that represents a device class at a higher level of the hierarchy can be connected to a second node via an edge that represents the “is parent of” relation and an edge that represents the “is child of” relation in the reverse direction);
generating a … network comprising a plurality of nodes, wherein ones of the plurality of nodes at the each level are connected to other ones of the plurality of nodes at a higher level over a plurality of connections ([0025], where each node represents a device class and each edge connecting two nodes represents a relation between the two device classes represented by the two nodes, e.g., an edge connecting the first and second nodes can represent a “is coupled with” relation to indicate a potential coupling of a router and a load balancer in a computer system, and each edge can also have one or more attributes, including a weight, [0049], a first node that represents a device class at a higher level of the hierarchy can be connected to a second node via an edge that represents the “is parent of” relation and an edge that represents the “is child of” relation in the reverse direction);
… obtaining weights for each of the plurality of connections ([0048], each directed edge in the KG can have a weight that represents the strength of or the confidence score for the represented relation, [0072], the server 102 is programmed to prepare a recommendation for resolving the unknown issue encountered by the target computer system based on the applicable issue resolution rules by ranking these issue resolution rules can based on the extent of applicability to the support bundle); and
pruning the plurality of connections in the fully connected neural network by removing ones of the plurality of connections having weights that are lower than a predefined threshold ([0066], the server 102 is programmed to determine which issue resolution rules to apply, wherein to minimize the number of issue resolution rules for consideration, the server 102 includes only those edges that represent one or more specific types of relations with weights above a certain threshold and consider only those issue resolution rules that are associated with the nodes in the resulting subset of the KG in terms of applicability to the support bundle, [0072], the server 102 is programmed to prepare a recommendation for resolving the unknown issue encountered by the target computer system based on the applicable issue resolution rules, and these issue resolution rules can be ranked based on the extent of applicability to the support bundle, e.g., an exact match of the device states referred to in an issue resolution rule may ranked higher than a match that requires expanding the scope of the issue resolution rule via various attributes of or edges connecting the nodes in the KG, and, the server 102 can be configured to include only the known resolutions from a number of highest-ranking issue resolution rules in the recommendation, Claims 1 & 13, a computer-implemented method of resolving infrastructure issues receives a support bundle for a computer system that has encountered an unknown issue and identifying a specific issue resolution rule of the plurality of issue resolution rules that is applicable to the support bundle based on the knowledge graph, wherein the identifying comprises determining multiple issue resolution rules of the plurality of issue resolution rules that are applicable to the support bundle based on the knowledge graph; assigning ranks to the multiple issue resolution rules based on an extent of applicability to the support bundle, the specific issue resolution rule having one of highest ranks among the multiple issue resolution rules), Mathur does not expressly disclose the remaining elements of the following limitations, which however, are taught by further teachings in Creed.
Creed teaches further comprising generating the asset hierarchy through a deep learning scheme, the generating the asset hierarchy comprising ([0084], NNs or NN structures that may be used by the invention as described herein may include at least one or more neural network structures from the group of: artificial NNs (ANNs); deep NNs; deep learning; deep learning ANNs; deep belief networks; deep Boltzmann machines, [0096], neural networks are used to generate graph models by using a loss function to update weights of hidden layers of the GNN model, Abstract, the method of generating a graph neural network (GNN) model based on an entity-entity graph comprises generating an embedding based on data representative of the entity-entity graph for the GNN model comprising an attention weight assigned to each relationship edge of the entity-entity graph indicating the relevancy of each relationship edge between entity nodes of the entity-entity graph): …
generating a fully connected neural network comprising a plurality of nodes, wherein ones of the plurality of nodes at the each level are connected to other ones of the plurality of nodes at a higher level over a plurality of connections ([0096], neural networks are used to generate graph models for predicting/classifying problems associated with training datasets by determining embeddings or a latent space associated with the training dataset and using a loss function to update weights of hidden layers of the GNN model (e.g. the encoding network 104 and scoring network 106);
training the fully connected neural network and obtaining weights for each of the plurality of connections ([0031], method comprises repeating the steps of generating the embedding and/or scoring, and updating the weights until the GNN model is determined to be trained or updated [0109], the trained attention weights of the GNN model may be output for use as the application demands, and the trained attention weights that are output may be used to filter the entity-entity relationship graph by weighting each edge with the corresponding trained attention weight; and
pruning the plurality of connections in the fully connected neural network by removing ones of the plurality of connections having weights that are lower than a predefined threshold ([0025], generating the filtered entity-entity graph based on the trained attention weights further comprises removing any relationship edges from the entity-entity graph having a corresponding trained attention weight that is below or equal to an attention relevancy threshold, [0114], in step 134, filtering the entity-entity graph based on the trained attention weights may include removing those relationship edge(s) from the entity-entity graph or portion thereof having a corresponding trained attention weight that is below or equal or above or equal to an attention relevancy threshold, and additionally or alternatively, may further include identifying uncertain relationship edges in the entity-entity graph or portion thereof based on a set of attention weights retrieved from the trained GNN model associated with the entity-entity graph and an attention relevancy threshold, and the identified uncertain relationship edges may be culled and rules may be derived defining actions for excluding the relationship edges).
Mathur and Creed are analogous fields of invention because both address the problem of generating a graphical structure with nodes and edges linking nodes to each other and weights indicting the relationship between nodes. At the time the invention was effectively filed, it would have been obvious to one of ordinary skill in the art to include in the system of Mathur the ability to generate a deep learning network as taught by Creed since the claimed invention is merely a combination of old elements, and in the combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art would have recognized that the combination would produce the predictable results of generating a deep learning network, as claimed. Further, it would have been obvious to one of ordinary skill in the art to have modified Mathur with the aforementioned teachings of Creed in order to produce the added benefit of improving knowledge graph quality. [0170].
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 CHARLES A GUILIANO whose telephone number is (571)272-9859. The examiner can normally be reached Mon-Fri 10:00 am - 6:00 pm.
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, Rutao Wu can be reached at 571-272-6045. 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.
CHARLES GUILIANO
Primary Examiner
Art Unit 3623
/CHARLES GUILIANO/Primary Examiner, Art Unit 3623
Prosecution Insights