DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This Office Action is in response to amendments filed on May 29, 2026.
Claims 1-2, 4, 6-9, 11, 13-16, 18 and 21-27 are pending.
Claims 1, 4, 6, 8, 11, 13 and 15 have been amended.
Claims 5, 12, 19 and 20 have been canceled.
Claims 21-27 have been added.
Response to Amendment
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 1-2, 4, 6-9, 11, 13-16, 18 and 23-27 are rejected under 35 U.S.C. 103 as being unpatentable over John Cable (“AI Powers Windows 10 April 2018 Update Rollout”, Jun 14, 2018) in view of Wei et al. (US 2009/0070756).
With respect to Claim 1, John Cable discloses:
obtaining, by at least one processing device, input data comprising a set of parameters associated with a computing environment; (retrieving characteristics of devices (set of parameters) to help identify target device (computing environment), Page 1-2, lines 1-19 and 1-8 respectively; continuously collecting feedback/telemetry data, Page 1, lines 4-9)
determining, by the at least one processing device using at least one machine learning (ML) model based on the input data (update experience/feedback/telemetry data is used to retrain the model, Page 1, lines 7-8), whether a device of the computing environment is due for an operating system (OS) upgrade, (Our AI approach (at least one ML model) intelligently selects devices (due for an upgrade) that our feedback data indicate would have a great update experience and offers the April 2018 Update to these devices first, Page 1, lines 4-7) wherein the OS upgrade for the device comprises a sequence of OS upgrade steps customized for a device type of the device; (2018 Update (version 1803) is fully available for all compatible devices (device type) wherein the update includes adjusting and preventing devices (device type) that are affected by an issue, throttle the update rollout and continuing again once issues have been resolved (sequence of OS upgrade steps), Page 2, AI means both safe AND fast, lines 6-10 and Page 4, Windows 10 April 2018 Update (1803) is now fully available, lines 1-6; OS upgrade can include work arounds and fixes that prevent issues that arise during an OS upgrade (sequence of OS upgrade steps customized for a device type), Pages 2-3, AI means both safe AND fast, lines 11-18 and 1-4 respectively)
in response to determining that the device is due for the OS upgrade, initiating, by the at least one processing device, the OS upgrade for the device; (using AI models to select devices that would have a great update experience (due for an OS upgrade) and offering/rolling out the update, Page 1, lines 4-12)
identifying, by the at least one processing device, a staged OS for the device; (commencing rollout of the April 2018 update (a staged OS) to devices that would have a great experience, Page 1, lines 4-10)
determining, by the at least one processing device and based on the staged OS and the set of parameters, whether to continue the OS upgrade for the device, (When our AI model, feedback or telemetry data (set of parameters) indicate that there may be an issue, we quickly adjust and prevent affected devices (the device) from being offered the update until we thoroughly investigate (determine whether to continue). Once issues are resolved we proceed again with confidence. This allows us to throttle the update rollout to customers without them needing to take any action., Page 2, AI means both safe AND fast, lines 6-10), wherein determining whether to continue the OS upgrade for the device comprises:
determining, by using the at least one ML model, whether at least one issue with implementing the OS upgrade exists, and wherein the at least one issue comprises a deviation from normal device activity for the device type; (When our AI model (at least one ML model), feedback or telemetry data indicate that there may be an issue (deviation from normal activity), we quickly adjust and prevent affected devices (device type) from being offered the update until we thoroughly investigate, Page 2, AI means both safe AND fast, lines 6-10)
and determining, by using the at least one ML model, [an issue]. (When our AI model (at least one ML model), feedback or telemetry data indicate that there may be an issue, we quickly adjust and prevent affected devices from being offered the update until we thoroughly investigate. Page 2, AI means both safe AND fast, lines 6-10)
in response to determining to complete the OS upgrade for the device, continuing, by the at least one processing device, the OS upgrade for the device. (Once issues are resolved (determining to complete the OS upgrade) we proceed again with confidence. (continuing the OS upgrade) This allows us to throttle the update rollout to customers without them needing to take any action., Page 2, AI means both safe AND fast, lines 6-10)
John Cable does not disclose:
[an issue] includes whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device;
However, Wei et al. disclose:
[an issue] includes whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device; (determining whether resources are available (resource consumption satisfies a threshold condition associated with a normal resource consumption pattern) to continue an update process and if not, suspend a software update until resources are available, Paragraph 64; resources may not be available due to network congestion, large pending document processing jobs, errors experienced by the document processing device or the like (normal/abnormal resource consumption pattern), Paragraph 64)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of Wei et al. into the teaching of John Cable to include [an issue] includes whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device in order to be able to throttle software updates based on resource utilization which can help complete software updates in a way that minimizes competition with other resources associated with device operation. (Wei et al., Abstract and Paragraph 1, lines 1-2 and 2-5 respectively)
With respect to Claim 2, all the limitations of Claim 1 have been addressed above; and John Cable further disclose:
wherein determining whether the device is due for the OS upgrade further comprises determining whether an upgrade flag for the device is set. (Based on the update quality and reliability we are seeing through our AI approach, we are now expanding the release broadly to make the April 2018 Update (version 1803) fully available for all compatible devices running Windows 10 worldwide. Full availability is the final phase of our rollout process. You don’t have to do anything to get the update; it will rollout automatically (upgrade flag for the device is set) to you through Windows Update, Page 4, Windows 10 April 2018 (1803) is now fully available, lines 1-6)
With respect to Claim 4, all the limitations of Claim 1 have been addressed above; and John Cable further disclose:
wherein identifying the staged OS comprises:
selecting, by the at least one processing device, an approved OS for the device; (using AI to select devices that would have a great update experience and offering the April 2018 update to these devices (an approved OS), Page 1, lines 4-7)
and performing, by the at least one processing device, a staging of the approved OS to obtain the staged OS; (commencing rollout of the April 2018 update (approved OS/staged OS), Page 1, lines 4-10)
With respect to Claim 6, all the limitations of Claim 1 have been addressed above; and John Cable further disclose:
wherein continuing the OS upgrade for the device further comprises:
reloading the device with the staged OS; (once issues are resolved, we proceed again (reloading the device) with the update with confidence, Page 2, AI means both safe AND fast, lines 6-10; continue safe rollout of the April 2018 update to devices using a fix/updated update (reloading the device with the staged OS), Pages 2-3, AI means both safe AND fast, lines 11-18 and 1-4 respectively)
and executing a post-check process to determine whether an issue with the OS upgrade for the device exists. (As our rollout progresses, we continuously collect update experience data (post-check process) and retrain our models to learn which devices will have a positive experience and where we may need to wait until we have higher confidence in a great experience (determine whether an issue with the OS upgrade for the device exists), Page 1, lines 7-10)
With respect to Claim 7, all the limitations of Claim 1 have been addressed above; and John Cable further disclose:
further comprising using, by the at least one processing device, the at least one ML model to make a prediction associated with [an issue]. (When our AI model, feedback or telemetry data indicate that there may be an issue (prediction associated with computing environment), we quickly adjust and prevent affected devices from being offered the update until we thoroughly investigate. Once issues are resolved we proceed again with confidence. This allows us to throttle the update rollout to customers without them needing to take any action., Page 2, AI means both safe AND fast, lines 6-10)
John Cable does not disclose:
[an issue] includes computing environment resource consumption
However, Wei et al. disclose:
[an issue] includes computing environment resource consumption (determining whether resources are available (computing environment resource consumption) to continue an update process and if not, suspend a software update until resources are available, Paragraph 64; resources may not be available due to network congestion, large pending document processing jobs, errors experienced by the document processing device or the like (computing environment resource consumption), Paragraph 64)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of Wei et al. into the teaching of John Cable to include [an issue] includes computing environment resource consumption in order to be able to throttle software updates based on resource utilization which can help complete software updates in a way that minimizes competition with other resources associated with device operation. (Wei et al., Abstract and Paragraph 1, lines 1-2 and 2-5 respectively)
Claims 8-9, 11 and 13-14 are system claims corresponding to the method claims above (Claims 1-2, 4 and 6-7) and, therefore, are rejected for the same reasons set forth in the rejections of Claims 1-2, 4 and 6-7.
With respect to Claim 15, John Cable discloses:
obtaining, by a processing device, input data for training at least one machine learning (ML) model to manage an operating system (OS) upgrade for a device of a computing environment, (continuously collect update experience data (input data) and retrain our models to learn which devices will have a positive update experience and whether we may need to wait until we have higher confidence in a great experience (manage an OS upgrade for a device of a computing environment), Page 1, lines 4-10) wherein the input data comprises a set of training parameters, (update experience data (set of training parameters), Page 1, lines 7-8) and wherein the OS upgrade for the device comprises a sequence of OS upgrade steps customized for a device type of the device; (2018 Update (version 1803) is fully available for all compatible devices (device type) wherein the update includes adjusting and preventing devices (device type) that are affected by an issue, throttle the update rollout and continuing again once issues have been resolved (sequence of OS upgrade steps), Page 2, AI means both safe AND fast, lines 6-10 and Page 4, Windows 10 April 2018 Update (1803) is now fully available, lines 1-6; OS upgrade can include work arounds and fixes that prevent issues that arise during an OS upgrade (sequence of OS upgrade steps customized for a device type), Pages 2-3, AI means both safe AND fast, lines 11-18 and 1-4 respectively)
and training, by the processing device based on the input data, the at least one ML model to manage the OS upgrade for the device, (continuously collect update experience data (input data) and retrain our models to learn which devices will have a positive update experience and whether we may need to wait until we have higher confidence in a great experience (manage an OS upgrade for a device of a computing environment), Page 1, lines 4-10) wherein training the at least one ML model to manage the OS upgrade for the device comprises training the at least one ML model to determine, based on the set of parameters and a staged OS for the device, whether to continue the OS upgrade for the device, (When our AI model, feedback or telemetry data (the set of parameters) indicate that there may be an issue, we quickly adjust and prevent affected devices from being offered the update (staged OS) (determine whether to continue the OS upgrade/mange the OS upgrade) until we thoroughly investigate. Once issues are resolved we proceed again with confidence (continue with the OS upgrade when there are no longer any issues). This allows us to throttle the update rollout to customers without them needing to take any action., Page 2, AI means both safe AND fast, lines 6-10) and wherein training the ML model to determine whether to continue the OS upgrade for the device comprises:
training the at least one ML model to determine whether at least one issue with implementing the OS upgrade exists, and wherein the at least one issue comprises a deviation from normal device activity for the device type; (When our AI model (at least one ML model), feedback or telemetry data indicate that there may be an issue (deviation from normal activity), we quickly adjust and prevent affected devices (device type) from being offered the update until we thoroughly investigate, Page 2, AI means both safe AND fast, lines 6-10)
and training the ML model to determine, using the at least one ML model, [an issue]. (When our AI model (ML model), feedback or telemetry data indicate that there may be an issue, we quickly adjust and prevent affected devices from being offered the update until we thoroughly investigate (determine whether to continue). Once issues are resolved we proceed again with confidence. This allows us to throttle the update rollout to customers without them needing to take any action., Page 2, AI means both safe AND fast, lines 6-10)
John Cable does not disclose:
[an issue] includes whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device;
However, Wei et al. disclose:
[an issue] includes whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device; (determining whether resources are available (resource consumption satisfies a threshold condition associated with a normal resource consumption pattern) to continue an update process and if not, suspend a software update until resources are available, Paragraph 64; resources may not be available due to network congestion, large pending document processing jobs, errors experienced by the document processing device or the like (normal/abnormal resource consumption pattern), Paragraph 64)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of Wei et al. into the teaching of John Cable to include [an issue] includes whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device in order to be able to throttle software updates based on resource utilization which can help complete software updates in a way that minimizes competition with other resources associated with device operation. (Wei et al., Abstract and Paragraph 1, lines 1-2 and 2-5 respectively)
With respect to Claim 16, all the limitations of Claim 15 have been addressed above; and John Cable further disclose:
wherein training the ML model the at least one ML model to manage the OS upgrade further comprises training the ML model to make a prediction associated with [an issue]. (When our AI model (trained ML model), feedback or telemetry data indicate that there may be an issue (prediction associated with computing environment), we quickly adjust and prevent affected devices from being offered the update until we thoroughly investigate. Once issues are resolved we proceed again with confidence. This allows us to throttle the update rollout to customers without them needing to take any action., Page 2, AI means both safe AND fast, lines 6-10)
John Cable does not disclose:
[an issue] includes computing environment resource consumption
However, Wei et al. disclose:
[an issue] includes computing environment resource consumption (determining whether resources are available (computing environment resource consumption) to continue an update process and if not, suspend a software update until resources are available, Paragraph 64; resources may not be available due to network congestion, large pending document processing jobs, errors experienced by the document processing device or the like (computing environment resource consumption), Paragraph 64)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of Wei et al. into the teaching of John Cable to include [an issue] includes computing environment resource consumption in order to be able to throttle software updates based on resource utilization which can help complete software updates in a way that minimizes competition with other resources associated with device operation. (Wei et al., Abstract and Paragraph 1, lines 1-2 and 2-5 respectively)
With respect to Claim 18, all the limitations of Claim 15 have been addressed above; and John Cable further disclose:
wherein training the ML model to manage the OS upgrade for the device further comprises training the ML model to cause the OS upgrade for the device to be halted in response to determining to discontinue the OS upgrade for the device. (When our AI model (trained ML model), feedback or telemetry data indicate that there may be an issue, we quickly adjust and prevent affected devices (halt the OS upgrade) from being offered the update until we thoroughly investigate (discontinue the OS upgrade). Once issues are resolved we proceed again with confidence. This allows us to throttle the update rollout to customers without them needing to take any action., Page 2, AI means both safe AND fast, lines 6-10; in cases where devices already offered the update may see issues, we immediately block all PCs that could be impacted by the issue from being updated (halt the OS upgrade/discontinue the OS upgrade), Page 2, AI means both safe AND fast, lines 11-16)
With respect to Claim 23, all the limitations of Claim 1 have been addressed above; and John Cable further disclose:
further comprising, by the at least one processing device in response to determining to discontinue the OS upgrade for the device, initiating at least one remedial action, wherein the at least one remedial action comprises at least one of: cancellation of the OS upgrade, postponement of the OS upgrade, or an OS downgrade. (When our AI model, feedback or telemetry data indicate that there may be an issue, (determine to discontinue the OS upgrade) we quickly adjust and prevent affected devices (initiating at least one remedial action) from being offered the update until we thoroughly investigate (postponement/cancellation of the OS upgrade). Once issues are resolved we proceed again with confidence. This allows us to throttle the update rollout to customers without them needing to take any action., Page 2, AI means both safe AND fast, lines 6-10
With respect to Claim 24, all the limitations of Claim 1 have been addressed above; and John Cable further disclose:
further comprising, by the at least one processing device in response to determining to discontinue the OS upgrade for the device, sending an alert to at least one administrator device indicating that the OS upgrade should be discontinued. (When our AI model, feedback or telemetry data indicate that there may be an issue, (determine to discontinue the OS upgrade) we quickly adjust and prevent (sending an alert that the OS upgrade should be discontinued) affected devices from being offered the update (the system (one administrator device) no longer offers the update to affected devices) until we thoroughly investigate. Page 2, AI means both safe AND fast, lines 6-10)
With respect to Claim 25, all the limitations of Claim 1 have been addressed above; and John Cable further disclose:
further comprising, by the at least one processing device in response to detecting an anomaly after the OS upgrade has been completed for the device, causing an OS upgrade for at least one other device of the device type to be halted. (In cases where devices already offered the update may see issues (an anomaly after the OS upgrade has been completed), we immediately blocked all PCs that could be impacted by this issue (causing an OS upgrade for at least one other device of the device type to be halted), Page 2, AI means both safe AND fast, lines 11-17)
Claims 26 and 27 are system claims corresponding to the method claims above (Claims 23 and 25) and, therefore, are rejected for the same reasons set forth in the rejections of Claims 23 and 25.
Claims 21-22 are rejected under 35 U.S.C. 103 as being unpatentable over John Cable (“AI Powers Windows 10 April 2018 Update Rollout”, Jun 14, 2018) in view of Wei et al. (US 2009/0070756) and in further view of Martin Brinkmann ("Portable Update: Search, Download and Install All (Missing) Windows Update”, Jun 2, 2013).
With respect to Claim 21, all the limitations of Claim 1 have been addressed above; and John Cable and Wei et al. do not disclose:
wherein determining whether the device is due for the OS upgrade using the at least one ML model comprises determining whether a deviation from an OS upgrade history for the device exists.
However, Martin Brinkmann discloses:
wherein determining whether the device is due for the OS upgrade using the at least one ML model comprises determining whether a deviation from an OS upgrade history for the device exists. (search Microsoft’s update repository (OS upgrade history) and determine any missing updates (deviation) and download/install those only on the PC in question, Page 1, lines 10-13)
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teaching of Martin Brinkmann into the teaching of John Cable and Wei et al. to include determining whether a deviation from an OS upgrade history for the device exists in order to determine, download and install any missing updates on specific computers. (Martin Brinkmann, Page 1, lines 10-13)
Claim 22 is a system claim corresponding to the method claim above (Claim 21) and, therefore, is rejected for the same reasons set forth in the rejection of Claim 21.
Response to Arguments
Applicant's arguments filed May 29, 2026 have been fully considered but they are not persuasive.
In the Remarks, Applicant argues:
Applicant respectfully submits that the combination of Cable and Wei fails to disclose, teach, or suggest every element of claim 1, either as previously presented or as currently amended. As set forth below, neither Cable nor Wei, alone or in combination, discloses or suggests the use of at least one ML model to (1) determine "whether at least one issue with implementing the OS upgrade exists, and wherein the at least one issue comprises a deviation from normal device activity for the device type," or (2) determine "whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device."
Examiner’s Response:
The Examiner respectfully disagrees. As can be seen in the updated rejection to Claim 1 above, it is the Examiner’s position that Cable discloses the “use of at least one ML model to (1) determine whether at least one issue with implementing the OS upgrade exists, and wherein the at least one issue comprises a deviation from normal device activity for the device type”. Specifically, Cable discloses “When our AI model, feedback or telemetry data indicate that there may be an issue, we quickly adjust and prevent affected devices from being offered the update until we thoroughly investigate.” (see Page 2, AI means both safe AND fast, lines 6-10) The “AI model” in Cable can be reasonably interpreted as the Applicant’s “ML model”. This “AI model” is used to indicate (determine) if there is an issue. Any “issue” can be reasonably interpreted as a “deviation from normal device activity”. The affected devices or the number of PCs that are affected by the issue can be reasonably considered the “device type”. The claims do no provide further detail on either how “normal device activity” has deviated or what a “device type” is.
Further, it is the Examiner’s position that through the combination of Cable and Wei disclose the “use of at least one ML model to “(2) determine whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device.” Specifically, Cable discloses using an ML model to determine an “issue” as explained above. Cable is silent that the specific issue can be “whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device.” Wei was then used to disclose determining this specific issue. Wei discloses “determining whether resources are available to continue an update process and if not, suspend a software update until resources are available.” (see Paragraph 64) Further, Wei discloses that resources may not be available due to network congestion, large pending document processing jobs, errors experienced by the document processing device or the like. (see Paragraph 64) Therefore, due to reasons such as network congestion (normal resource consumption pattern is above a threshold (satisfies a threshold condition)), determining that there is not enough available resources and suspending a software update. This citation can be reasonably interpreted as the Applicant’s “(2) determine whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device.”
In the Remarks, Applicant argues:
With respect to Cable, the Office Action at pages 4-6 relies on the following passage to teach determining whether to continue the OS upgrade: "When our Al model, feedback or telemetry data indicate that there may be an issue, we quickly adjust and prevent affected devices from being offered the update until we thoroughly investigate. Once issues are resolved we proceed again with confidence. This allows us to throttle the update rollout to customers without them needing to take any action." Cable describes using Al to determine whether to offer an update to a device based on "feedback or telemetry data."
However, Cable's Al model operates before the update is initiated. It "prevent[s] affected devices from being offered the update" based on issues identified from other devices' experiences. Cable does not disclose or suggest using a machine learning model to determine, during an ongoing OS upgrade for a particular device, "whether at least one issue with implementing the OS upgrade exists, and wherein the at least one issue comprises a deviation from normal device activity for the device type," as recited in amended claim 1. Nor does Cable disclose or suggest using a machine learning model to determine "whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device."
Examiner’s Response:
The Examiner respectfully disagrees. Applicant argues that “Cable's Al model operates before the update is initiated.” However, Cable discloses “Our AI approach has enabled us to quickly spot issues during deployment of a feature update.” (see Page 2, AI means both safe AND fast, lines 1-2) Therefore, Cable’s AI model operates after an updated is initiated (i.e. during deployment). Further, Cable disclose that “[i]n cases where devices already offered the update may see issues…we immediately block all PCs that could be impacted by the issue from being updated..” (see Page 2, AI means both safe AND fast, lines 11-17). In this scenario, the AI model has identified an issue after an update has been offered (initiated) on devices.
Further, Applicant argues that “Cable does not disclose or suggest using a machine learning model to determine, during an ongoing OS upgrade for a particular device, “whether at least one issue with implementing the OS upgrade exists, and wherein the at least one issue comprises a deviation from normal device activity for the device type”. As responded to in the previous arguments above, Cable discloses the “use of at least one ML model to determine whether at least one issue with implementing the OS upgrade exists, and wherein the at least one issue comprises a deviation from normal device activity for the device type”. Specifically, Cable discloses “When our AI model, feedback or telemetry data indicate that there may be an issue, we quickly adjust and prevent affected devices from being offered the update until we thoroughly investigate.” (see Page 2, AI means both safe AND fast, lines 6-10) The “AI model” in Cable can be reasonably interpreted as the Applicant’s “ML model”. This “AI model” is used to indicate (determine) if there is an issue. Any “issue” can be reasonably interpreted as a “deviation from normal device activity”. The affected devices or number of PCs can be reasonably considered the “device type”. The claims do no provide further detail on either how “normal device activity” has deviated or what a “device type” is. Further still, Cable discloses “Our AI approach has enabled us to quickly spot issues during deployment of a feature update.” (see Page 2, AI means both safe AND fast, lines 1-2) Therefore, Cable’s AI model operates “during an ongoing OS upgrade”.
In the Remarks, Applicant argues:
The Office Action at pages 4-5 acknowledges that Cable does not disclose "[an issue] includes whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device," and cites paragraph [0064] of Wei to cure this deficiency of Cable. Wei likewise fails to cure the deficiencies of Cable.
For example, the Abstract of Wei states:
The subject application is directed to a system and method for resource utilization-based throttling of software updates. A transfer of updated software is first commenced to a document processing device so as to perform an update of operational software associated with the device. Resource utilization associated with the operation of the document processing device is then sensed. In accordance with an output of the sensed resource utilization, an update process is throttled associated with the document processing device and the updated software.
Paragraph [0064] of Wei states:
A determination is then made at step 622 based upon the sensed resource
utilization, whether resources are available to continue the update process. When resources are not available, e.g., network congestion, large pending document processing jobs, errors experienced by the document processing device 104, or the like, the operation terminates. When resources are available, flow proceeds to step 624, whereupon the controller 108 or other suitable component associated with the document processing device 104 determines if the update corresponds to a previously interrupted or suspended software update. When a previously suspended update is detected, flow proceeds to step 626, whereupon the controller 108 retrieves, from the data storage device 110, suspension data corresponding to the suspended software update, e.g., software downloaded, where in the update process the operation was suspended, and the like. Thereafter, the transfer of the incomplete software update is re-initiated at step 628.
As the above passages make clear, Wei merely describes sensing "whether resources are available to continue the update process." Specifically, Wei describes determining that "[w]hen resources are not available, e.g., network congestion, large pending document processing jobs, errors experienced by the document processing device 104, or the like, the operation terminates." Wei's disclosure is fundamentally different from the claimed invention in at least the following respects:
Wei does not disclose or suggest "determining, by using the at least one ML model, whether at least one issue with implementing the OS upgrade exists, and wherein the at least one issue comprises a deviation from normal device activity for the device type," as recited in amended claim 1. Wei's resource utilization sensing is not directed to detecting "a deviation from normal device activity for the device type." Rather, Wei checks whether resources such as "network congestion, large pending document processing jobs, [or] errors" are present at a given moment. Wei does not teach or suggest any baseline of "normal device activity" against which current device activity could be compared to identify deviations generally.
Wei does not disclose or suggest "determining, by using the at least one ML model, whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device," as recited in amended claim 1. Wei discloses sensing "network congestion, large pending document processing jobs, errors experienced by the document processing device 104, or the like." This is a reactive, rule-based resource availability check to determine whether resources happen to be available for a device at a given moment, not a pattern-based analysis based on "a normal resource consumption pattern of the device."
In summary, Wei does not use at least one ML model to analyze resource consumption or detect issues. Wei's resource sensing does not involve any machine learning, or predictive or pattern-recognition capability in general.
In contrast, amended claim 1 recites "determining, by using the at least one ML model, whether at least one issue with implementing the OS upgrade exists, and wherein the at least one issue comprises a deviation from normal device activity for the device type" and "determining, by using the at least one ML model, whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device." Wei's reactive, non-ML-based resource availability check provides no teaching or suggestion of using at least one ML model to detect deviations from normal device activity or to determine whether resource consumption during an OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device.
Examiner’s Response:
The Examiner respectfully disagrees. Applicant argues that Wei does not disclose "determining, by using the at least one ML model, whether at least one issue with implementing the OS upgrade exists, and wherein the at least one issue comprises a deviation from normal device activity for the device type”. As can be seen in the updated §103 rejection to claim 1, the Examiner has not relied on Wei do disclose this limitation.
Further, Applicant argues that Wei does not disclose “determining, by using the at least one ML model, whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device”. The Examiner would first like to note that Wei was not solely used to disclose the above limitation. It is through the combination of Cable and Wei disclose “determining, by using the at least one ML model, whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device.” As stated above, Cable discloses using an ML model to determine an “issue” as explained above. Cable is silent that the specific issue can be “whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device.” Wei was then used to disclose determining this specific issue. Wei discloses “determining whether resources are available to continue an update process and if not, suspend a software update until resources are available.” (see Paragraph 64) Further, Wei discloses that resources may not be available due to network congestion, large pending document processing jobs, errors experienced by the document processing device or the like. (see Paragraph 64) Therefore, due to reasons such as network congestion (normal resource consumption pattern is above a threshold (satisfies a threshold condition)), determining that there is not enough available resources and suspending a software update.
Further still, Applicant argues that Wei does not disclose a “pattern-based analysis based on a normal resource consumption pattern of the device." In response to applicant's argument that the references fail to show certain features of the invention, it is noted that the features upon which applicant relies (i.e., pattern-based analysis) are not recited in the rejected claim(s). Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993). As responded to above, it is the Examiner’s position that Wei discloses determining “whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device.”
In the Remarks, Applicant argues:
Moreover, the Office Action fails to articulate a sufficient rationale for combining Cable and Wei. An obviousness determination requires "some articulated reasoning with some rational underpinning to support the legal conclusion of obviousness." The Office Action's statement on page 5 that Wei's teachings would "help complete software updates in a way that minimizes competition with other resources associated with device operation" only provides a motivation for utilizing the non-ML-based resource availability check described by Wei. It does not explain why one of ordinary skill in the art would have been motivated to fundamentally transform Wei's non-ML-based resource availability check into a sophisticated ML-based analysis that (1) determines "whether at least one issue with implementing the OS upgrade exists, and wherein the at least one issue comprises a deviation from normal device activity for the device type," and (2) determines "whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device."
Examiner’s Response:
The Examiner respectfully disagrees. Applicant argues that the Examiner has not explained why one of ordinary skill in the art would be motivated to transform “Wei's non-ML-based resource availability check into a sophisticated ML-based analysis that (1) determines whether at least one issue with implementing the OS upgrade exists, and wherein the at least one issue comprises a deviation from normal device activity for the device type, and (2) determines whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device.” However, as can be seen in the updated §103 rejection, Wei’s non-ML based resource availability check is not being transformed but Cable’s disclosure of an issue is being “transformed” and/or modified to be a specific issue that is identified. As stated above, Wei was used to disclose the determination of the specific issue of “whether a resource consumption of the device during the OS upgrade satisfies a threshold condition associated with a normal resource consumption pattern of the device.” One skilled in the art would want to identify this issue in order to help complete software updates in a way that minimizes competition with other resources associated with device operation. (see Wei, Abstract and Paragraph 1, lines 1-2 and 2-5 respectively)
In the Remarks, Applicant argues:
Additionally, Wei provides no teaching or suggestion of establishing baseline patterns, training at least one ML model to recognize deviations, or applying any form of machine learning analysis. Such a modification would require a wholesale replacement of Wei's technical approach, not a mere combination of known elements.
Examiner’s Response:
The Examiner respectfully disagrees. Applicant argues that Wei “provides no teaching or suggestion of establishing baseline patterns, training at least one ML model to recognize deviations, or applying any form of machine learning analysis.” However, Wei was not used to disclose any of these features.
In the Remarks, Applicant argues:
In view of the foregoing, the proposed combination of Cable and Wei constitutes
impermissible hindsight reconstruction. The Office Action has relied on Applicant's own claim language as a template to selectively extract disparate teachings from the cited references, supplying an after-the-fact rationale to justify the proposed combination. See In re McLaughlin, 443 F.2d 1392, 1395 (CCPA 1971) ("Any judgment on obviousness is in a sense necessarily a reconstruction based upon hindsight reasoning, but so long as it takes into account only knowledge which was within the level of ordinary skill at the time the claimed invention was made and does not include knowledge gleaned only from applicant's disclosure, such a reconstruction is proper."). Here, the Office Action's reasoning improperly draws from Applicant's disclosure to bridge the substantial gap between the teachings of Cable and Wei and the claimed invention. Accordingly, the Office Action has not established a prima facie case of obviousness, and one of ordinary skill in the art would not have found it obvious to combine the teachings of Cable and Wei to arrive at the claimed invention.
Examiner’s Response:
In response to applicant's argument that the examiner's conclusion of obviousness is based upon improper hindsight reasoning, it must be recognized that any judgment on obviousness is in a sense necessarily a reconstruction based upon hindsight reasoning. But so long as it takes into account only knowledge which was within the level of ordinary skill at the time the claimed invention was made, and does not include knowledge gleaned only from the applicant's disclosure, such a reconstruction is proper. See In re McLaughlin, 443 F.2d 1392, 170 USPQ 209 (CCPA 1971).
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 LANNY N UNG whose telephone number is (571)270-7708. The examiner can normally be reached Mon-Thurs 6:30am-3:30pm.
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, Bradley Teets can be reached at 571-272-3338. 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.
/LANNY N UNG/Primary Examiner, Art Unit 2197
Prosecution Insights