CTFR 18/661,928 CTFR 94969 DETAILED ACTION The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA. This action is in response to the communication filed on 01/30/2026. Claims 11-15 and 17-31 are pending in this application. Information Disclosure Statement The information disclosure statement(s) (IDS) submitted on 02/06/2026 is/are in compliance with the provisions of 37 CFR 1.97. Accordingly, the IDS(s) is/are being considered by the examiner. Response to Arguments 07-37 AIA Applicant’s arguments filed 01/30/2026 have been fully considered but they are not persuasive. Applicant argues: a. Applicant states that “Although Hollender refers to word embeddings and vectorial representations, Hollender does not teach that its vectorial representation or embeddings are generated using a first machine learning model as recited in claim 14. Accordingly, Applicant respectfully submits that Hollender does not teach subject matter of claim 14 conceded by the Office Action to be missing from Cote (Reply, page 15).” a. Examiner respectfully disagrees. US 20240419707 A1 (hereinafter Kumar) teaches that the Natural Language Processing and Word Embeddings is closely associated with a machine learning model (Kumar, para. [0021] “… The method includes aggregating, based on the correlating, the first set of word embeddings and the second set of word embeddings into a global ML model of word embeddings …”). Therefore, Hollender continues to teach the corresponding claim limitations of claim 14 . Response to Amendment The objection to the Specification is no withdrawn in view of the amendments. The claim rejections under 35 U.S.C. 112(b) to claims 14-18 and 20 are now withdrawn in view of the claim amendments. Applicant’s arguments with respect to independent claims 11 and 19 and their corresponding dependent claims have been considered but are moot based on the new grounds of rejection necessitated by Applicant’s amendments. Specifically, the arguments present that Lu fails to provide for the amended language (“Although Lu refers to a series of models, Lu does not teach the third portion of the machine learning model of claim 11 generating an output based on two pieces of information including … the contextual representation … and the prediction (Reply, page 14)”), where the rejection below now relies on Hyde to teach this subject matter. See the rejection section for details. Claim Rejections - 35 USC § 103 07-06 AIA 15-10-15 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. 07-20-aia AIA 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. 07-23-aia AIA The factual inquiries set forth in Graham v. John Deere Co. , 383 U.S. 1, 148 USPQ 459 (1966), that are applied for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows: 1. Determining the scope and contents of the prior art. 2. Ascertaining the differences between the prior art and the claims at issue. 3. Resolving the level of ordinary skill in the pertinent art. 4. Considering objective evidence present in the application indicating obviousness or nonobviousness. 07-20-02-aia AIA This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention. 07-21-aia AIA Claim s 11-13, 19, 23 and 27-29 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 20210028973 A1 (hereinafter Cote), in view of US 20200036611 A1 (hereinafter Lu), and in further view of US 20220014422 A1 (hereinafter Hyde) . For Claim 11, Cote teaches a non-transitory machine-readable storage medium comprising instructions that upon execution cause a computer system (Cote, para. [0013] “… a non-transitory computer-readable medium that stores computer logic is described. The computer logic may have instruction that, when executed, cause a processing device to obtain information defining a topology of a multi-layer network having a plurality of Network Elements (NEs) and a plurality of links interconnecting the NEs …”) to: receive a first representation of attributes associated with a network stack connected to an underlay network that couples a first system to a computing environment, wherein the network stack comprises a plurality of layers (Cote teaches a communications network exemplifying entity 18 communicating with entity 22 in FIG. 1A, and teaches receiving PM metrics and alarms from a plurality of layers including underlay layers; FIG. 1A; para. [0054] “… Input data may be obtained in a way that is similar to known network assurance applications. Performance Monitoring (PM) metrics (e.g., signal strength, power, etc.) may be obtained from Layer 0 (e.g., physical layer, optical layer) components. PM metrices (e.g., pre-Forward Error Correction (FEC) Bit Error Rate (BER), etc.) may be obtained from Layer 1 components. PM metrics (e.g., received/transmit/discarded frames, etc.) may be obtained from Layer 2 components. Also, PM metrics (e.g., latency, jitter, etc.) may be obtained from Layer 3 components. In addition to PM metrics, the systems of the present disclosure may be configured to obtain ‘alarms’ signifying various states of conditions of the communications network. Various Layer 0-3 alarms may include ‘Loss of Signal,’ ‘Channel Degrade,’ ‘High Fiber Loss,’ ‘High Received Span Loss,’ ‘Optical Line Fail,’ ‘Shutoff Threshold Crossed,’ etc. …”; para. [0055] “… Generally speaking, a communications network is built using multiple technologies, which in many cases create underlay/overlay layers. An overlay layer depends on the services of its underlay layer. The ‘atomic’ NEs are the most granular components of the underlay layer for the problem under consideration. They typically report PM metrics and Alarms directly …”; para. [0060] “… FIG. 1A is a diagram showing the topology of a section of a communications network 10 according to one embodiment. FIG. 1B is a dependency graph of the communications network 10. The exemplary communications network 10 includes a first layer (e.g., Layer 0) having AMP 12, AMP 14, and AMP 16. A second, overlaid layer (e.g., Layer 1) includes an Optical Transport Network (OTN) transmitter 18 and an OTN receiver 20. The OTN receiver 20 may be part of a device ( e.g., switch, router, server, shelf, node, etc.) 22 having a packet-optical device 24 …”) ; receive a second representation of attributes associated with an overlay network provided over the underlay network, wherein the second representation comprises a representation of attributes associated with communication paths in the overlay network (Cote teaches receiving PM metrics and alarms from layered composite structures of an overlay network; FIG. 1A, FIG. 2A, FIG. 6A; para. [0055] “… Generally speaking, a communications network is built using multiple technologies, which in many cases create underlay/overlay layers …”; para. [0056] “… The composite structures are part of the overlay layer and are made up of several atomic elements. They typically report end-to-end PM and Alarms in the context of the overlay layer …”; para. [0067] “… The MPLS receiver 40 may be configured to receive the packets and measure end-to-end Quality of Service (QoS), latency, packet loss, jitter, and/or other metrics. The MPLS receiver 40 may be placed in the device 42 (e.g., shelf), where the packet-optical device 44 is configured to report alarms if a QoS parameter does not meet a minimum QoS threshold …”; para. [0102] “… Network topology is analyzed using graph analytics and a dependency graph is created to indicate which network elements are to be analyzed during RCI for the top-level service. A group of network elements with the same set of collected data is called a ‘network entity group’ …”; para. [0103] “… During training (i.e., FIG. 6A), data of a network entity group is overlaid and labelled with alarms from the service for the group (e.g., block 220) …”) ; generate, using a first portion of a machine learning model, a prediction of an underlay network fault associated with the underlay network based on the first representation provided as an input to the first portion of the machine learning model, the machine learning model trained to detect faults associated with communications between the first system and the computing environment (Cote teaches providing network time-series data (i.e. PM metrics and alarms) from the underlay networks to the machine learning model to identify the root cause of issues in the communications network and predict the issues; FIG. 1A, FIG. 3B; para. [0061] “… During operation of the communications network 10 , various PM metrics may be obtained from the different components over time … Also, the packet-optical device 24 may be configured to analyze conditions of the device 22 and provide alarm signals to a control (CTRL) device 26. The CTRL device 26 may be configured in a control plane of the communications network 10 for monitoring the various PM metrics and alarms from the various components and/or sensors …”; para. [0062] “… The CTRL device 26 may be configured to utilize various processes (e.g., ML processes) for performing Root Cause Identification (RCI) with network time-series object detection to determine the root cause of one or more issues in the communications network 10 … The embodiments of the present disclosure may rely on Machine Learning (ML) for performing RCI with the input time-series data obtained from the communications network 10 …”; para. [0074] “… Software applications may use Big Data and ML techniques to identify and classify network issues automatically and help with pro-active remediation process. With a ML model trained, the software applications can then utilize current data to predict when an issue is imminent. In the example of FIG. 3B, two issues (e.g., faults, failures, line breaks, etc.) 120, 122 are shown …”) ; …; generate, using … the machine learning model …., an output … representing a likelihood of a presence of a fault associated with the overlay network or the underlay network (Cote teaches utilizing a ML process to identify a problematic component in the multilayer network, FIG. 1A, FIG. 4; para. [0085] “… the root cause identifying module 144 may be configured to enable the processing device 132 to obtain information defining a topology of a multi-layer network having a plurality of Network Elements (NEs) and a plurality of links interconnecting the NEs. Then, the root cause identifying module 144 may be configured to enable the processing device 132 to receive Performance Monitoring (PM) metrics and one or more alarms from the multilayer network. Based on the information defining the topology, the PM metrics, and the one or more alarms, the processing device 132, as instructed by the root cause identifying module 144, may be configured to utilize a Machine Learning (ML) process to identify a problematic component from the plurality of NEs and links. The ML process may also identify a root cause of one or more issues in the multi-layer network, whereby the root cause may be identified as being associated with the problematic component …”; para. [0089] “… The plurality of NEs may be arranged in an underlay layer of the multi-layer network, and the multi-layer network may include a composite structure arranged in an overlay layer. The PM metrics may include optical power, pre-Forward Error Correction Bit Error Rate (pre-FEC BER), received discarded frames, transmitted discarded frames, latency, and/or jitter. The one or more alarms may include Loss of Signal (LoS), channel degrade, high fiber loss, high received span loss, optical line failure, and/or shutoff threshold crossed …”) ; and based on the output, initiate a remediation action to address the fault represented by the output . (Cote, FIG. 4; para. [0087] “… The step of utilizing the ML process may include the steps of: a) utilizing a plurality of ML models, each ML model corresponding to a respective NE, and b) hierarchically combining results from the plurality of ML models. The ML process may be configured to generate human-readable text to describe the problematic component. For example, the ML process may include an explanation generator configured to generate human-readable text describing the problematic component, whereby the human-readable text may include an explanation that provides insight for remediating the root cause or problematic component …”). Cote does not explicitly teach, but Lu teaches generate, using a second portion of the machine learning model, a contextual representation based on the second representation provided as an input to the second portion of the machine learning model (Lu teaches the traffic prediction model generating a feature vector/contextual representation of the real-time collected traffic data, Lu also teaches a traffic prediction model may include different layers/portions; FIG. 3; para. [0056] “… In this embodiment, a traffic prediction model may include a convolutional neural network, a residual network and a fully connected layer …”; para. [0057] “… Step 301, inputting the real-time collected traffic data sequence into the convolutional neural network, to obtain a feature vector of the real-time collected traffic data sequence …”; para. [0063] “… Step 302, inputting the feature vector of the real-time collected traffic data sequence into the residual network, to obtain a fused feature vector of the feature vector of the real-time collected traffic data sequence …”; as discussed above, Cote teaches receiving the traffic data such as PM metrics and alarms from layered composite structures of an overlay network and the traffic data may be input to a layer/portion of the traffic prediction model/neural network model) ; Lu and Cote are analogous art because they are both related to collecting network metrics. Before the effective filing date of the claimed invention it would have been obvious to one of ordinary skill in the art to use the multi-layered machine learning techniques of Lu with the system of Cote to timely discover the abnormality of the network traffic (Lu ¶ 0047) . Cote-Lu does not explicitly teach, but Hyde teaches generate, using a third portion of the machine learning model based on the contextual representation and the prediction, an output comprising a value representing a likelihood of a presence of a fault associated with the overlay network or the underlay network (Hyde teaches an AI model generating an anomaly impact score for an overlay network supported application by correlating anomaly predictions for each layer of a plurality of network layers, also as discussed above, the feature vector/contextual representation taught by Lu and the predicted root cause issues may be input to the cross-layer fault tolerance AI model taught by Hyde; FIG. 15; para. [0026] “… The cross-layer fault tolerance system interfaces with a distributed anomaly detection system that develops AI based models to predict anomalies in the network. Such anomalies may happen at the physical layer (e.g., anomalous signal variation or outage in a given location, etc.), at the MAC layer (e.g., prediction of handover failures, congestion prediction, etc.), application layer anomalies such as dropped calls, poor quality of experience (QoE), etc. The activations or confidence levels of the anomaly detectors are fed into an AI based correlation engine (e.g., a graph-based model, etc.) that localize a set of anomalies that are correlated as well as identify which anomalies are coupled together …”; para. [0126] “… At operation 1505, anomaly data may be obtained (e.g., via means including a logic gate of a processor, network communication circuitry, etc.) from a plurality of layers of a network … In an example the anomaly data may be obtained from monitoring of payload integrity errors (e.g., checksum errors, etc.), packet delivery delays, round trip time (RTT) measurements, etc. in an overlay network for an anomaly that may occur in an underlay network through which the overlay network traffic is tunneled …”; para. [0128] “… At operation 1515, an artificial intelligence model may be trained … using the elements of the anomaly data to generate an impact score for the application. In an example, the anomaly data may include anomaly predictions for each layer of the plurality of layers …”; para. [0129] “… At operation 1520, the impact score may be generated … for the application by evaluating current network metrics using the artificial intelligence model …”). Hyde and Cote-Lu are analogous art because they are both related to collecting network metrics. Before the effective filing date of the claimed invention it would have been obvious to one of ordinary skill in the art to use the evaluating the cross-layer network impact via an AI model techniques of Hyde with the system of Cote-Lu to facilitate the network error tracking system in order to improve the recovery mechanism of applications running on the network (Hyde ¶ 0003) . For Claim 12, Cote-Lu-Hyde teaches the non-transitory machine-readable storage medium of claim 11, wherein the machine learning model comprises a neural network comprising the first portion, the second portion, and the third portion (Lu teaches a traffic prediction model may include different layers/portions; para. [0056] “… In this embodiment, a traffic prediction model may include a convolutional neural network, a residual network and a fully connected layer …”). See motivation to combine for claim 11. For Claim 13, Cote-Lu-Hyde teaches the non-transitory machine-readable storage medium of claim 11, wherein the underlay network includes a plurality of network paths to the computing environment, the plurality of network paths comprising a first network path and a second network path, wherein the machine learning model is for the first network path and the overlay network is established over the first network path, and wherein the instructions upon execution cause the computer system to: use a different machine learning model to predict a fault associated with the second network path and another overlay network established over the second network path (Cote teaches using multiple ML models (one per NE) to identify the root cause of the network traffic faults, while different NEs being applied in different paths ; para. [0052] “… In the present disclosure, the various embodiments solve the root issue identifying problem by considering all NEs simultaneously and by processing the network topology information (e.g., using ML). Two different methods, among others, of solving this problem with ML may include 1) using a single ML model for the entire communications network (e.g., using an ML approach similar to an object recognition algorithm), and 2) using multiple ML models (one per NE) combined hierarchically (e.g., using an ML approach with a correlation of alarms) …”; para. [0054] “… Data input may also be obtained for determining the topology of the multi-layer communications network. The topology, for example, may include physical connectivity of devices, logical paths across multi-layer routes, etc. …”). For Claim 19, the claim is substantially similar to claim 11 and therefore is rejected for the same reasoning set forth above. For Claim 23, Cote-Lu-Hyde teaches the non-transitory machine-readable storage medium of claim 11, wherein the second representation further comprises a representation of attributes associated with a service that uses the overlay network for communications (Cote teaches receiving PM metrics and alarms from layered composite structures of an overlay network and application level information; FIG. 1A, FIG. 2A, FIG. 6A; para. [0055] “… Generally speaking, a communications network is built using multiple technologies, which in many cases create underlay/overlay layers …”; para. [0056] “… The composite structures are part of the overlay layer and are made up of several atomic elements. They typically report end-to-end PM and Alarms in the context of the overlay layer …”; para. [0067] “… The MPLS receiver 40 may be configured to receive the packets and measure end-to-end Quality of Service (QoS), latency, packet loss, jitter, and/or other metrics. The MPLS receiver 40 may be placed in the device 42 (e.g., shelf), where the packet-optical device 44 is configured to report alarms if a QoS parameter does not meet a minimum QoS threshold …”; para. [0102] “… Network topology is analyzed using graph analytics and a dependency graph is created to indicate which network elements are to be analyzed during RCI for the top-level service. A group of network elements with the same set of collected data is called a ‘network entity group’ …”; para. [0103] “… During training (i.e., FIG. 6A), data of a network entity group is overlaid and labelled with alarms from the service for the group (e.g., block 220) …”) , and wherein the contextual representation is based on the representation of the attributes associated with the communication paths in the overlay network and the representation of the attributes associated with the service (Lu teaches the traffic prediction model generating a feature vector/contextual representation of the real-time collected traffic data; FIG. 3; para. [0057] “… Step 301, inputting the real-time collected traffic data sequence into the convolutional neural network, to obtain a feature vector of the real-time collected traffic data sequence …”; para. [0063] “… Step 302, inputting the feature vector of the real-time collected traffic data sequence into the residual network, to obtain a fused feature vector of the feature vector of the real-time collected traffic data sequence …”; as discussed above, Cote teaches receiving the traffic data such as PM metrics and alarms from layered composite structures of an overlay network and application level information). See motivation to combine for claim 11. For Claim 27, Cote-Lu-Hyde teaches the non-transitory machine-readable storage medium of claim 11, wherein the second representation comprises a representation of attributes associated with routes in the overlay network (Cote teaches receiving PM metrics and alarms from layered composite structures of an overlay network including a router; FIG. 1A; para. [0055] “… Generally speaking, a communications network is built using multiple technologies, which in many cases create underlay/overlay layers …”; para. [0056] “… The composite structures are part of the overlay layer and are made up of several atomic elements. They typically report end-to-end PM and Alarms in the context of the overlay layer …”; para. [0065] “… The MPLS receiver 40 may be part of a device (e.g., switch, router, server, shelf, node, etc.) 42 having a packet-optical device 44 that is configured to provide alarms to a CTRL device 46 …”). For Claim 28, Cote-Lu-Hyde teaches the non-transitory machine-readable storage medium of claim 11, wherein the second representation comprises a representation of attributes associated with a tunnel according to a security protocol in the overlay network (Hyde, FIG. 15; para. [0126] “… In an example the anomaly data may be obtained from monitoring of payload integrity errors (e.g., checksum errors, etc.), packet delivery delays, round trip time (RTT) measurements, etc. in an overlay network for an anomaly that may occur in an underlay network through which the overlay network traffic is tunneled. In an example when monitoring of traffic in the overlay network is used for diagnosis of anomalies in an underlay network, security filtering may admit or exclude anomaly data providers in the overlay network so that anomaly data from only trusted overlay networks is used in identification of anomalies in the underlay network or in the diagnoses of anomalies in the underlay network …”). See motivation to combine for claim 11. For Claim 29, Cote-Lu-Hyde teaches the non-transitory machine-readable storage medium of claim 12, wherein the first portion comprises a first collection of neurons of a first hidden layer of the neural network, the second portion comprises a second collection of neurons of the first hidden layer, and the third portion comprises neurons of an output layer of the neural network (Lu teaches a traffic prediction model including multiple layers/portions of neural networks, each layer/portion having a collection of neurons; para. [0049] “… The artificial neural network is usually composed of a large number of nodes (also known as neurons) connected to each other. Each node represents a specific output function, called an activation function. The connection between every two nodes represents a weighted value, which is called a weight (also called a parameter). The output of the network varies according to the connection mode, the weighted value and the activation function of the network. The traffic prediction model often includes a plurality of layers, and each layer includes a plurality of nodes …”; para. [0056] “… a traffic prediction model may include a convolutional neural network, a residual network and a fully connected layer …”). See motivation to combine for claim 11 . Claim Rejections - 35 USC § 103 07-21-aia AIA Claim s 14-15 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 20210028973 A1 (hereinafter Cote), in view of US 20220044178 A1 (hereinafter Hollender), and in further view of US 20210083926 A1 (hereinafter Costa) . For Claim 14, Cote teaches a system comprising: a hardware processor; and a non-transitory storage medium storing machine-readable instructions executable on the hardware processor (Cote, para. [0013] “… a non-transitory computer-readable medium that stores computer logic is described. The computer logic may have instruction that, when executed, cause a processing device to obtain information defining a topology of a multi-layer network having a plurality of Network Elements (NEs) and a plurality of links interconnecting the NEs …”) to: receive a first representation of attributes associated with an underlay network stack connected to an underlay network that couples a first system to a computing environment, wherein the underlay network stack comprises a plurality of layers (Cote teaches a communications network exemplifying entity 18 communicating with entity 22 in FIG. 1A, and discloses receiving PM metrics and alarms from a plurality of layers including underlay layers; FIG. 1A; para. [0054] “… Input data may be obtained in a way that is similar to known network assurance applications. Performance Monitoring (PM) metrics (e.g., signal strength, power, etc.) may be obtained from Layer O (e.g., physical layer, optical layer) components. PM metrices (e.g., pre-Forward Error Correction (FEC) Bit Error Rate (BER), etc.) may be obtained from Layer 1 components. PM metrics (e.g., received/transmit/discarded frames, etc.) may be obtained from Layer 2 components. Also, PM metrics (e.g., latency, jitter, etc.) may be obtained from Layer 3 components. In addition to PM metrics, the systems of the present disclosure may be configured to obtain ‘alarms’ signifying various states of conditions of the communications network. Various Layer 0-3 alarms may include ‘Loss of Signal,’ ‘Channel Degrade,’ ‘High Fiber Loss,’ ‘High Received Span Loss,’ ‘Optical Line Fail,’ ‘Shutoff Threshold Crossed,’ etc. …”; para. [0055] “… Generally speaking, a communications network is built using multiple technologies, which in many cases create underlay/overlay layers. An overlay layer depends on the services of its underlay layer. The ‘atomic’ NEs are the most granular components of the underlay layer for the problem under consideration. They typically report PM metrics and Alarms directly …”; para. [0060] “… FIG. 1A is a diagram showing the topology of a section of a communications network 10 according to one embodiment. FIG. 1B is a dependency graph of the communications network 10. The exemplary communications network 10 includes a first layer (e.g., Layer 0) having AMP 12, AMP 14, and AMP 16. A second, overlaid layer (e.g., Layer 1) includes an Optical Transport Network (OTN) transmitter 18 and an OTN receiver 20. The OTN receiver 20 may be part of a device ( e.g., switch, router, server, shelf, node, etc.) 22 having a packet-optical device 24 …”) ; receive a second representation of attributes associated with an overlay network provided over the underlay network (Cote teaches receiving PM metrics and alarms from layered composite structures of an overlay network; FIG. 1A; para. [0055] “… Generally speaking, a communications network is built using multiple technologies, which in many cases create underlay/overlay layers …”; para. [0056] “… The composite structures are part of the overlay layer and are made up of several atomic elements. They typically report end-to-end PM and Alarms in the context of the overlay layer …”) ; …; provide the first … representation and the second … representation to a second machine learning model trained to detect a fault associated with communications between the first system and the computing environment (Cote teaches providing network time-series data (i.e. PM metrics and alarms) from the underlay/overlay networks to identify the root cause of issues in the communications network; FIG. 1A; para. [0061] “… During operation of the communications network 10 , various PM metrics may be obtained from the different components over time … Also, the packet-optical device 24 may be configured to analyze conditions of the device 22 and provide alarm signals to a control (CTRL) device 26. The CTRL device 26 may be configured in a control plane of the communications network 10 for monitoring the various PM metrics and alarms from the various components and/or sensors …”; para. [0062] “… The CTRL device 26 may be configured to utilize various processes (e.g., ML processes) for performing Root Cause Identification (RCI) with network time-series object detection to determine the root cause of one or more issues in the communications network 10 … The embodiments of the present disclosure may rely on Machine Learning (ML) for performing RCI with the input time-series data obtained from the communications network 10 …”) ; and generate, by the second machine learning model, an output … representing a likelihood of a presence of the fault associated with the overlay network or the underlay network (Cote teaches utilizing a ML process to identify a problematic component in the multilayer network, FIG. 1A, FIG. 4; para. [0085] “… the root cause identifying module 144 may be configured to enable the processing device 132 to obtain information defining a topology of a multi-layer network having a plurality of Network Elements (NEs) and a plurality of links interconnecting the NEs. Then, the root cause identifying module 144 may be configured to enable the processing device 132 to receive Performance Monitoring (PM) metrics and one or more alarms from the multilayer network. Based on the information defining the topology, the PM metrics, and the one or more alarms, the processing device 132, as instructed by the root cause identifying module 144, may be configured to utilize a Machine Learning (ML) process to identify a problematic component from the plurality of NEs and links. The ML process may also identify a root cause of one or more issues in the multi-layer network, whereby the root cause may be identified as being associated with the problematic component …”; para. [0089] “… The plurality of NEs may be arranged in an underlay layer of the multi-layer network, and the multi-layer network may include a composite structure arranged in an overlay layer. The PM metrics may include optical power, pre-Forward Error Correction Bit Error Rate (pre-FEC BER), received discarded frames, transmitted discarded frames, latency, and/or jitter. The one or more alarms may include Loss of Signal (LoS), channel degrade, high fiber loss, high received span loss, optical line failure, and/or shutoff threshold crossed …”). Cote does not explicitly teach, but Hollender teaches generate, using a first machine learning model, a first embedding representation of the first representation of attributes, and a second embedding representation of the second representation of attributes (Hollender teaches utilizing Natural Language Processing and Word Embeddings to generate embedding representations of alarms and events for a communicatively coupled industrial system; Examiner notes that the Natural Language Processing and Word Embeddings is closely associated with a machine learning model in US 20240419707 A1 (para. [0021]); para. [0117] “… Natural Language Processing and Word Embeddings is used to obtain vectorial representations of alarms and events in such a way that related alarms/events are mapped to close points in feature space. In this way knowhow can be transferred between identical sections/equipment (.e.g. several identical coal mills).Also the relation between closely related sensors, like redundant sensors, can be taken into account ( current approaches treat each alarm separately as soon as the name is different) …”) ; provide the first embedding representation and the second embedding representation to a second machine learning model (Hollender teaches to provide the embedding representation to a neuronal network trained to identify alarm/action relationship; para. [0118] “… A prediction model such as neuronal network is taught for alarm/action relationship based on the word embeddings …”). Hollender and Cote are analogous art because they are both related to identifying network issues. Before the effective filing date of the claimed invention it would have been obvious to one of ordinary skill in the art to use the alarms/events processing techniques of Hollender with the system of Cote to facilitate identifying the important alarms and recommending the correct actions (Hollender ¶ 0005) . Cote-Hollender does not explicitly teach, but Costa teaches generating an output comprising a value representing a likelihood of a presence of the fault associated with a network (Costa teaches using a ML model to generate probability values associated with network issues; FIGS 1-3; para. [0056] “… The issue detection model 225 can use a machine learning architecture applied to network metrics to automatically determine and classify network issues …”; para. [0065] “… the issue detection model can be trained to provide a plurality of output values, for example in a vector, array, or other type of data structure. The outputs can be probability values that are each assigned to one or more possible network issues. The greater the probability value for a given network issue, the greater the likelihood that such issue is present in the network under analysis …”). Costa and Cote-Hollender are analogous art because they are both related to identifying network issues. Before the effective filing date of the claimed invention it would have been obvious to one of ordinary skill in the art to use the classifying the probability values of the network issues techniques of Costa with the system of Cote-Hollender to “provide the ability to troubleshoot from the access layer to subscriber devices and determine where issues exist before taking more extensive troubleshooting measures” (Costa ¶ 0005) . For Claim 15, Cote-Hollender-Costa teaches the system of claim 14, wherein the output generated by the second machine learning model comprises contribution parameters that indicate faults in respective layers of the underlay network and layers of the overlay network (Cote teaches utilizing the ML process to identify a problematic component in the multilayer network, FIG. 1A, FIG. 4; para. [0085] “… the root cause identifying module 144 may be configured to enable the processing device 132 to obtain information defining a topology of a multi-layer network having a plurality of Network Elements (NEs) and a plurality of links interconnecting the NEs. Then, the root cause identifying module 144 may be configured to enable the processing device 132 to receive Performance Monitoring (PM) metrics and one or more alarms from the multilayer network. Based on the information defining the topology, the PM metrics, and the one or more alarms, the processing device 132, as instructed by the root cause identifying module 144, may be configured to utilize a Machine Learning (ML) process to identify a problematic component from the plurality of NEs and links. The ML process may also identify a root cause of one or more issues in the multi-layer network, whereby the root cause may be identified as being associated with the problematic component …”; para. [0089] “… The plurality of NEs may be arranged in an underlay layer of the multi-layer network, and the multi-layer network may include a composite structure arranged in an overlay layer. The PM metrics may include optical power, pre-Forward Error Correction Bit Error Rate (pre-FEC BER), received discarded frames, transmitted discarded frames, latency, and/or jitter. The one or more alarms may include Loss of Signal (LoS), channel degrade, high fiber loss, high received span loss, optical line failure, and/or shutoff threshold crossed …”) . Claim Rejections - 35 USC § 103 07-21-aia AIA Claim 17 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 20210028973 A1 (hereinafter Cote), in view of US 20220044178 A1 (hereinafter Hollender), in view of US 20210083926 A1 (hereinafter Costa), in view of US 20200036611 A1 (hereinafter Lu), and in further view of US 20220014422 A1 (hereinafter Hyde) . For Claim 17, Cote-Hollender-Costa teaches the system of claim 14, wherein the machine-readable instructions are executable on the hardware processor to: generate, using a first portion of the second machine learning model, a prediction of an underlay network fault associated with the underlay network based on the first embedding representation provided as an input to the first portion of the second machine learning model (Cote teaches providing network time-series data (i.e. PM metrics and alarms) from the underlay networks to the machine learning model to identify the root cause of issues in the communications network and predict the issues; FIG. 1A, FIG. 3B; para. [0061] “… During operation of the communications network 10 , various PM metrics may be obtained from the different components over time … Also, the packet-optical device 24 may be configured to analyze conditions of the device 22 and provide alarm signals to a control (CTRL) device 26. The CTRL device 26 may be configured in a control plane of the communications network 10 for monitoring the various PM metrics and alarms from the various components and/or sensors …”; para. [0062] “… The CTRL device 26 may be configured to utilize various processes (e.g., ML processes) for performing Root Cause Identification (RCI) with network time-series object detection to determine the root cause of one or more issues in the communications network 10 … The embodiments of the present disclosure may rely on Machine Learning (ML) for performing RCI with the input time-series data obtained from the communications network 10 …”; para. [0074] “… Software applications may use Big Data and ML techniques to identify and classify network issues automatically and help with pro-active remediation process. With a ML model trained, the software applications can then utilize current data to predict when an issue is imminent. In the example of FIG. 3B, two issues (e.g., faults, failures, line breaks, etc.) 120, 122 are shown …”) ; Cote-Hollender-Costa does not explicitly teach, but Lu teaches generate, using a second portion of the second machine learning model, a contextual representation based on the second embedding representation provided as an input to the second portion of the second machine learning model (Lu teaches the traffic prediction model generating a feature vector/contextual representation of the real-time collected traffic data, Lu also teaches a traffic prediction model may include different layers/portions; FIG. 3; para. [0056] “… In this embodiment, a traffic prediction model may include a convolutional neural network, a residual network and a fully connected layer …”; para. [0057] “… Step 301, inputting the real-time collected traffic data sequence into the convolutional neural network, to obtain a feature vector of the real-time collected traffic data sequence …”; para. [0063] “… Step 302, inputting the feature vector of the real-time collected traffic data sequence into the residual network, to obtain a fused feature vector of the feature vector of the real-time collected traffic data sequence …”; as discussed above, Cote teaches receiving the traffic data such as PM metrics and alarms from layered composite structures of an overlay network and the traffic data may be input to a layer/portion of the traffic prediction model/neural network model) ; and … Lu and Cote-Hollender-Costa are analogous art because they are both related to collecting network metrics. Before the effective filing date of the claimed invention it would have been obvious to one of ordinary skill in the art to use the multi-layered machine learning techniques of Lu with the system of Cote-Hollender-Costa to timely discover the abnormality of the network traffic (Lu ¶ 0047) . Cote-Hollender-Costa does not explicitly teach, but Hyde teaches generate, using a third portion of the second machine learning model based on the contextual representation and the prediction, the output comprising the value representing the likelihood of the presence of the fault associated with the overlay network or the underlay network (Hyde teaches an AI model generating an anomaly impact score for an overlay network supported application by correlating anomaly predictions for each layer of a plurality of network layers; also as discussed above, the feature vector/contextual representation taught by Lu and the predicted root cause issues may be input to the cross-layer fault tolerance AI model taught by Hyde; FIG. 15; para. [0026] “… The cross-layer fault tolerance system interfaces with a distributed anomaly detection system that develops AI based models to predict anomalies in the network. Such anomalies may happen at the physical layer (e.g., anomalous signal variation or outage in a given location, etc.), at the MAC layer (e.g., prediction of handover failures, congestion prediction, etc.), application layer anomalies such as dropped calls, poor quality of experience (QoE), etc. The activations or confidence levels of the anomaly detectors are fed into an AI based correlation engine (e.g., a graph-based model, etc.) that localize a set of anomalies that are correlated as well as identify which anomalies are coupled together …”; para. [0126] “… At operation 1505, anomaly data may be obtained (e.g., via means including a logic gate of a processor, network communication circuitry, etc.) from a plurality of layers of a network … In an example the anomaly data may be obtained from monitoring of payload integrity errors (e.g., checksum errors, etc.), packet delivery delays, round trip time (RTT) measurements, etc. in an overlay network for an anomaly that may occur in an underlay network through which the overlay network traffic is tunneled …”; para. [0128] “… At operation 1515, an artificial intelligence model may be trained … using the elements of the anomaly data to generate an impact score for the application. In an example, the anomaly data may include anomaly predictions for each layer of the plurality of layers …”; para. [0129] “… At operation 1520, the impact score may be generated … for the application by evaluating current network metrics using the artificial intelligence model …”). Hyde and Cote-Hollender-Costa-Lu are analogous art because they are both related to collecting network metrics. Before the effective filing date of the claimed invention it would have been obvious to one of ordinary skill in the art to use the evaluating the cross-layer network impact via an AI model techniques of Hyde with the system of Cote-Hollender-Costa-Lu to facilitate the network error tracking system in order to improve the recovery mechanism of applications running on the network (Hyde ¶ 0003) . Claim Rejections - 35 USC § 103 07-21-aia AIA Claim s 18 and 30 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 20210028973 A1 (hereinafter Cote), in view of US 20220044178 A1 (hereinafter Hollender), in view of US 20210083926 A1 (hereinafter Costa), in view of US 20200036611 A1 (hereinafter Lu), in view of US 20220014422 A1 (hereinafter Hyde), and in further view of “Reasoning under Uncertainty for Overlay Fault Diagnosis” (hereinafter Tang) . For Claim 18, Cote-Hollender-Costa-Lu-Hyde teaches the system of claim 17, wherein the third portion of the second machine learning model identifies the fault (Hyde teaches an AI model generating an anomaly impact score for an overlay network supported application by correlating anomaly predictions for each layer of a plurality of network layers; FIG. 15; para. [0126] - para. [0129]). Cote-Hollender-Costa-Lu-Hyde does not explicitly teach, but Tang teaches identifies the fault as being present in the underlay network based on the contextual representation indicating that no fault is present in the overlay network (Tang teaches a network fault diagnosis framework DigOver solves the overlay versus underlay fault determination by quantifying a fault likelihood for each component of overlay and underlay networks and locating faulty components including overlay nodes and underlay routers; Section III. System Overview A. Evidential Fault Diagnosis Overview page 37 “… the fault likelihoods and evaluation uncertainty of all the investigated components are fed to the plausible reasoning (PR) component. By choosing an appropriate uncertainty handling strategy specified by a risk model, a total belief metric for each investigated component (denoted as Ψ c) can be calculated. Then, a plausible fault graph (PFC) is dynamically constructed to represent the correlation between each potential faulty component c and its related evidences e i (\/e i € E c , c € e i ). Finally, based on the total belief metric values, PR locates the most likely faulty components that explain the end-user observations. This is proven to be an NP-complete problem …”). Tang and Cote-Hollender-Costa-Lu-Hyde are analogous art because they are both related to collecting network metrics. Before the effective filing date of the claimed invention it would have been obvious to one of ordinary skill in the art to use the quantifying fault likelihood of each component of overlay and underlay network techniques of Tang with the system of Cote-Hollender-Costa-Lu-Hyde to overcome the overlay network fault diagnosis challenges (Tang, Section I. Introduction) . For Claim 30, Cote-Hollender-Costa-Lu-Hyde teaches the system of claim 17, wherein the third portion of the second machine learning model identifies the fault (Hyde teaches an AI model generating an anomaly impact score for an overlay network supported application by correlating anomaly predictions for each layer of a plurality of network layers; FIG. 15; para. [0126] - para. [0129]). Cote-Hollender-Costa-Lu-Hyde does not explicitly teach, but Tang teaches identifies the fault as being present in the overlay network based on the contextual representation indicating that a fault is present in the overlay network (Tang teaches a network fault diagnosis framework DigOver solves the overlay versus underlay fault determination by quantifying a fault likelihood for each component of overlay and underlay networks and locating faulty components including overlay nodes and underlay routers; Section III. System Overview A. Evidential Fault Diagnosis Overview page 37 “… the fault likelihoods and evaluation uncertainty of all the investigated components are fed to the plausible reasoning (PR) component. By choosing an appropriate uncertainty handling strategy specified by a risk model, a total belief metric for each investigated component (denoted as Ψ c) can be calculated. Then, a plausible fault graph (PFC) is dynamically constructed to represent the correlation between each potential faulty component c and its related evidences e i (\/e i € E c , c € e i ). Finally, based on the total belief metric values, PR locates the most likely faulty components that explain the end-user observations. This is proven to be an NP-complete problem …”). See motivation to combine for claim 18 . Claim Rejections - 35 USC § 103 07-21-aia AIA Claim s 20 and 26 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 20210028973 A1 (hereinafter Cote), in view of US 20200036611 A1 (hereinafter Lu), in view of US 20220014422 A1 (hereinafter Hyde), in view of US 20220044178 A1 (hereinafter Hollender), and in further view of US 20200076841 A1 (hereinafter Hajimirsadeghi) . For Claim 20, Cote-Lu-Hyde teaches the method of claim 19 . Cote-Lu-Hyde does not explicitly teach, but Hollender teaches wherein the first representation comprises a first embedding vector of the attributes associated with the layers of the underlay network, the second representation comprises a second embedding vector of the attributes associated with the overlay network (Hollender teaches the embedding representations of alarms and events for a communicatively coupled industrial system, para. [0117] “… Natural Language Processing and Word Embeddings is used to obtain vectorial representations of alarms and events in such a way that related alarms/events are mapped to close points in feature space. In this way knowhow can be transferred between identical sections/equipment (.e.g. several identical coal mills).Also the relation between closely related sensors, like redundant sensors, can be taken into account ( current approaches treat each alarm separately as soon as the name is different) …”) , … Hollender and Cote-Lu-Hyde are analogous art because they are both related to identifying network issues. Before the effective filing date of the claimed invention it would have been obvious to one of ordinary skill in the art to use the alarms/events processing techniques of Hollender with the system of Cote-Lu-Hyde to facilitate identifying the important alarms and recommending the correct actions (Hollender ¶ 0005) . Cote-Lu-Hyde-Hollender does not explicitly teach, but Hajimirsadeghi teaches the contextual representation comprises a contextual embedding vector representing likelihoods of faults occurring in the overlay network (Hajimirsadeghi teaches contextual embedding vector being applied for network traffic anomaly detection and analysis; FIG. 11; para. [0097] “… Techniques are provided herein for contextual embedding of features of operational logs or network traffic for anomaly detection based on sequence prediction. These techniques may detect malicious activity by anomaly analysis of context-aware feature embeddings of network packet flows, log messages, and/or log traces …”; para. [0208] “… dense sequence 1140 may be applied as stimulus input into a predictive RNN (not shown) to predict a next log message in the sequence. For example, dense feature vector 1141 may be applied to a first recurrent step of the predictive RNN to generate a next predicted dense feature vector (not shown) that is expected to match/approximate next dense feature vector 1142, which may be used to detect whether or not dense feature vector 1142 is contextually anomalous …”). Hajimirsadeghi and Cote-Lu-Hyde-Hollender are analogous art because they are both related to identifying network issues. Before the effective filing date of the claimed invention it would have been obvious to one of ordinary skill in the art to use the contextual embedding vector techniques of Hajimirsadeghi with the system of Cote-Lu-Hyde-Hollender to provide a dense, reduce and optimized vector that has more semantics to facilitate processing large volume of network traffic logs (Hajimirsadeghi, para. [0020]) . Cote-Lu-Hyde further teaches and wherein the first portion of the machine learning model generates the prediction of the underlay network fault based on the first embedding vector (Cote teaches providing network time-series data (i.e. PM metrics and alarms) from the underlay networks to the machine learning model to identify the root cause of issues in the communications network and predict the issues, the embedding representations of the network metrics and alarms as taught in Hollender may be the input to the ML model; FIG. 1A, FIG. 3B; para. [0061] “… During operation of the communications network 10 , various PM metrics may be obtained from the different components over time … Also, the packet-optical device 24 may be configured to analyze conditions of the device 22 and provide alarm signals to a control (CTRL) device 26. The CTRL device 26 may be configured in a control plane of the communications network 10 for monitoring the various PM metrics and alarms from the various components and/or sensors …”; para. [0062] “… The CTRL device 26 may be configured to utilize various processes (e.g., ML processes) for performing Root Cause Identification (RCI) with network time-series object detection to determine the root cause of one or more issues in the communications network 10 … The embodiments of the present disclosure may rely on Machine Learning (ML) for performing RCI with the input time-series data obtained from the communications network 10 …”; para. [0074] “… Software applications may use Big Data and ML techniques to identify and classify network issues automatically and help with pro-active remediation process. With a ML model trained, the software applications can then utilize current data to predict when an issue is imminent. In the example of FIG. 3B, two issues (e.g., faults, failures, line breaks, etc.) 120, 122 are shown …”) , and the second portion of the machine learning model generates the contextual embedding vector based on the second embedding vector (Lu teaches the traffic prediction model generating a feature vector/contextual representation of the real-time collected traffic data, the embedding representations of the network traffic data as taught in Hollender may be the input to the traffic prediction model; FIG. 3; para. [0057] “… Step 301, inputting the real-time collected traffic data sequence into the convolutional neural network, to obtain a feature vector of the real-time collected traffic data sequence …”; para. [0063] “… Step 302, inputting the feature vector of the real-time collected traffic data sequence into the residual network, to obtain a fused feature vector of the feature vector of the real-time collected traffic data sequence …”; as discussed above, Cote teaches receiving the traffic data such as PM metrics and alarms from layered composite structures of an overlay network and the traffic data may be input to a layer/portion of the traffic prediction model/neural network model). See further motivation to combine for claim 11. For Claim 26, the claim is substantially similar to claim 20 and therefore is rejected for the same reasoning set forth above . Claim Rejections - 35 USC § 103 07-21-aia AIA Claim s 21-22 and 24-25 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 20210028973 A1 (hereinafter Cote), in view of US 20200036611 A1 (hereinafter Lu), in view of US 20220014422 A1 (hereinafter Hyde), in view of US 20220044178 A1 (hereinafter Hollender), and in further view of “Reasoning under Uncertainty for Overlay Fault Diagnosis” (hereinafter Tang) . For Claim 21, Cote-Lu-Hyde teaches the non-transitory machine-readable storage medium of claim 11, wherein the third portion of the machine learning model identifies the fault (Hyde teaches an AI model generating an anomaly impact score for an overlay network supported application by correlating anomaly predictions for each layer of a plurality of network layers; FIG. 15; para. [0126] - para. [0129]). Cote-Lu-Hyde does not explicitly teach, but Tang teaches identifies the fault as being present in the underlay network based on the contextual representation indicating that no fault is present in the overlay network (Tang teaches a network fault diagnosis framework DigOver solves the overlay versus underlay fault determination by quantifying a fault likelihood for each component of overlay and underlay networks and locating faulty components including overlay nodes and underlay routers; Section III. System Overview A. Evidential Fault Diagnosis Overview page 37 “… the fault likelihoods and evaluation uncertainty of all the investigated components are fed to the plausible reasoning (PR) component. By choosing an appropriate uncertainty handling strategy specified by a risk model, a total belief metric for each investigated component (denoted as Ψ c) can be calculated. Then, a plausible fault graph (PFC) is dynamically constructed to represent the correlation between each potential faulty component c and its related evidences e i (\/e i € E c , c € e i ). Finally, based on the total belief metric values, PR locates the most likely faulty components that explain the end-user observations. This is proven to be an NP-complete problem …”). Tang and Cote-Lu-Hyde are analogous art because they are both related to collecting network metrics. Before the effective filing date of the claimed invention it would have been obvious to one of ordinary skill in the art to use the quantifying fault likelihood of each component of overlay and underlay network techniques of Tang with the system of Cote-Lu-Hyde to overcome the overlay network fault diagnosis challenges (Tang, Section I. Introduction) . For Claim 22, Cote-Lu-Hyde teaches the non-transitory machine-readable storage medium of claim 11, wherein the third portion of the machine learning model identifies the fault (Hyde teaches an AI model generating an anomaly impact score for an overlay network supported application by correlating anomaly predictions for each layer of a plurality of network layers; FIG. 15; para. [0126] - para. [0129]). Cote-Lu-Hyde does not explicitly teach, but Tang teaches identifies the fault as being present in the overlay network based on the contextual representation indicating that a fault is present in the overlay network (Tang teaches a network fault diagnosis framework DigOver solves the overlay versus underlay fault determination by quantifying a fault likelihood for each component of overlay and underlay networks and locating faulty components including overlay nodes and underlay routers; Section III. System Overview A. Evidential Fault Diagnosis Overview page 37 “… the fault likelihoods and evaluation uncertainty of all the investigated components are fed to the plausible reasoning (PR) component. By choosing an appropriate uncertainty handling strategy specified by a risk model, a total belief metric for each investigated component (denoted as Ψ c) can be calculated. Then, a plausible fault graph (PFC) is dynamically constructed to represent the correlation between each potential faulty component c and its related evidences e i (\/e i € E c , c € e i ). Finally, based on the total belief metric values, PR locates the most likely faulty components that explain the end-user observations. This is proven to be an NP-complete problem …”). See motivation to combine for claim 21. For Claim 24, Cote-Lu-Hyde teaches the non-transitory machine-readable storage medium of claim 23, wherein the third portion of the machine learning model identifies the fault (Hyde teaches an AI model generating an anomaly impact score for an overlay network supported application by correlating anomaly predictions for each layer of a plurality of network layers; FIG. 15; para. [0126] - para. [0129]). Cote-Lu-Hyde does not explicitly teach, but Tang teaches identifies the fault as being present in the underlay network based on the contextual representation indicating that no fault is present in the overlay network and the service (Tang teaches a network fault diagnosis framework DigOver solves the overlay versus underlay fault determination by quantifying a fault likelihood for each component of overlay and underlay networks and locating faulty components including overlay nodes and underlay routers; Section III. System Overview A. Evidential Fault Diagnosis Overview page 37 “… the fault likelihoods and evaluation uncertainty of all the investigated components are fed to the plausible reasoning (PR) component. By choosing an appropriate uncertainty handling strategy specified by a risk model, a total belief metric for each investigated component (denoted as Ψ c) can be calculated. Then, a plausible fault graph (PFC) is dynamically constructed to represent the correlation between each potential faulty component c and its related evidences e i (\/e i € E c , c € e i ). Finally, based on the total belief metric values, PR locates the most likely faulty components that explain the end-user observations. This is proven to be an NP-complete problem …”). See motivation to combine for claim 21. For Claim 25, Cote-Lu-Hyde teaches the non-transitory machine-readable storage medium of claim 23, wherein the third portion of the machine learning model identifies the fault (Hyde teaches an AI model generating an anomaly impact score for an overlay network supported application by correlating anomaly predictions for each layer of a plurality of network layers; FIG. 15; para. [0126] - para. [0129]). Cote-Lu-Hyde does not explicitly teach, but Tang teaches identifies the fault as being present in the overlay network or the service based on the contextual representation indicating that a fault is present in the overlay network or the service (Tang teaches a network fault diagnosis framework DigOver solves the overlay versus underlay fault determination by quantifying a fault likelihood for each component of overlay and underlay networks and locating faulty components including overlay nodes and underlay routers; Section III. System Overview A. Evidential Fault Diagnosis Overview page 37 “… the fault likelihoods and evaluation uncertainty of all the investigated components are fed to the plausible reasoning (PR) component. By choosing an appropriate uncertainty handling strategy specified by a risk model, a total belief metric for each investigated component (denoted as Ψ c) can be calculated. Then, a plausible fault graph (PFC) is dynamically constructed to represent the correlation between each potential faulty component c and its related evidences e i (\/e i € E c , c € e i ). Finally, based on the total belief metric values, PR locates the most likely faulty components that explain the end-user observations. This is proven to be an NP-complete problem …”). See motivation to combine for claim 21 . Claim Rejections - 35 USC § 103 07-21-aia AIA Claim 31 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 20210028973 A1 (hereinafter Cote), in view of US 20220044178 A1 (hereinafter Hollender), in view of US 20210083926 A1 (hereinafter Costa), in view of US 20200036611 A1 (hereinafter Lu), in view of US 20220014422 A1 (hereinafter Hyde), and in further view of US 20200076841 A1 (hereinafter Hajimirsadeghi) . For Claim 31, Cote-Hollender-Costa-Lu-Hyde teaches the system of claim 17 . Cote-Hollender-Costa-Lu-Hyde does not explicitly teach, but Hajimirsadeghi teaches wherein the contextual representation comprises a contextual embedding representation of likelihoods of faults occurring in the overlay network (Hajimirsadeghi teaches contextual embedding vector being applied for network traffic anomaly detection and analysis; FIG. 11; para. [0097] “… Techniques are provided herein for contextual embedding of features of operational logs or network traffic for anomaly detection based on sequence prediction. These techniques may detect malicious activity by anomaly analysis of context-aware feature embeddings of network packet flows, log messages, and/or log traces …”; para. [0208] “… dense sequence 1140 may be applied as stimulus input into a predictive RNN (not shown) to predict a next log message in the sequence. For example, dense feature vector 1141 may be applied to a first recurrent step of the predictive RNN to generate a next predicted dense feature vector (not shown) that is expected to match/approximate next dense feature vector 1142, which may be used to detect whether or not dense feature vector 1142 is contextually anomalous …”). Hajimirsadeghi and Cote-Hollender-Costa-Lu-Hyde are analogous art because they are both related to identifying network issues. Before the effective filing date of the claimed invention it would have been obvious to one of ordinary skill in the art to use the contextual embedding vector techniques of Hajimirsadeghi with the system of Cote-Hollender-Costa-Lu-Hyde to provide a dense, reduce and optimized vector that has more semantics to facilitate processing large volume of network traffic logs (Hajimirsadeghi, para. [0020]) .). Citation of Pertinent Prior Art 07-96 AIA The prior art made of record and not relied upon is considered pertinent to applicant's disclosure is listed below, thank you: i. US 2023/0153680 A1(hereinafter Rohrkemper) teaches using machine learning model validated sensor data to generate recommendations for remediating issues in a monitored system. A machine learning model is trained to identify correlations among sensors for a monitored system. Upon receiving current sensor data, the machine learning model identifies a subset of the current sensor data that cannot be validated. The system generates estimated values for the sensor data that cannot be validated based on the learned correlations among the sensor values. The system generates the recommendations for remediating the issues in the monitored system based on validated sensor values and the estimated sensor values (Abstract). ii. US 2023/0015709 A1 (hereinafter Bisht) teaches that A method of managing a controller of a software defined networking (SDN) network is implemented by a computing device in the SDN network. The method includes receiving status information for the controller, receiving usage information for the operating environment, generating at least one failure prediction for the controller based on the received status information, and outputting prediction information for the at least one failure prediction (Abstract). Conclusion 07-40 AIA 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 ZONGHUA DU whose telephone number is (408)918-7596. The examiner can normally be reached Monday - Friday 8 AM - 5 PM PST. 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, John Follansbee can be reached on (571) 272-3964. 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. /Z.D./Examiner, Art Unit 2444 /SCOTT B CHRISTENSEN/ Primary Examiner, Art Unit 2444 Application/Control Number: 18/661,928 Page 2 Art Unit: 2444 Application/Control Number: 18/661,928 Page 3 Art Unit: 2444 Application/Control Number: 18/661,928 Page 4 Art Unit: 2444 Application/Control Number: 18/661,928 Page 5 Art Unit: 2444 Application/Control Number: 18/661,928 Page 6 Art Unit: 2444 Application/Control Number: 18/661,928 Page 7 Art Unit: 2444 Application/Control Number: 18/661,928 Page 8 Art Unit: 2444 Application/Control Number: 18/661,928 Page 9 Art Unit: 2444 Application/Control Number: 18/661,928 Page 10 Art Unit: 2444 Application/Control Number: 18/661,928 Page 11 Art Unit: 2444 Application/Control Number: 18/661,928 Page 12 Art Unit: 2444 Application/Control Number: 18/661,928 Page 13 Art Unit: 2444 Application/Control Number: 18/661,928 Page 14 Art Unit: 2444 Application/Control Number: 18/661,928 Page 15 Art Unit: 2444 Application/Control Number: 18/661,928 Page 16 Art Unit: 2444 Application/Control Number: 18/661,928 Page 17 Art Unit: 2444 Application/Control Number: 18/661,928 Page 18 Art Unit: 2444 Application/Control Number: 18/661,928 Page 19 Art Unit: 2444 Application/Control Number: 18/661,928 Page 20 Art Unit: 2444 Application/Control Number: 18/661,928 Page 21 Art Unit: 2444 Application/Control Number: 18/661,928 Page 22 Art Unit: 2444 Application/Control Number: 18/661,928 Page 23 Art Unit: 2444 Application/Control Number: 18/661,928 Page 24 Art Unit: 2444 Application/Control Number: 18/661,928 Page 25 Art Unit: 2444 Application/Control Number: 18/661,928 Page 26 Art Unit: 2444 Application/Control Number: 18/661,928 Page 27 Art Unit: 2444 Application/Control Number: 18/661,928 Page 28 Art Unit: 2444 Application/Control Number: 18/661,928 Page 29 Art Unit: 2444
Prosecution Insights