The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
DETAILED ACTION
Claims 1-37 are presented for examination
Allowable Subject Matter
Claims 16, 30 and 37 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Response to Arguments
Applicant’s arguments, see pages 1 to 14, filed March 8, 2026, with respect to the rejection(s) of claim(s) 1, 24, and 31 under U.S.C. 102 have been fully considered and are not persuasive. Therefore, the rejection has been maintained.
Applicant argue that the reference generally mentioned the usage of the neural network, but not in the context of using the neural network to determine mode of operation. Application also fails to state in the claim how the neural network is used. Neural network is a broad term, so mentioning the word neural network is not enough to determine how it is use. This is why the other claims were objected, to show the need of adding something like reinforcement learning which is based on paragraph 0018 of the specification in order to add to the claim how the neural network is used to determine the port activity. To help with the prosecution of the application, an additional rejection is given below for the independent claims.
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 1, 24, and 31 are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Sanaullah (US Patent Application 20210243696).
As per claim 1, Sanaullah teaches a method [method shown in figure 56], comprising:
at a device [device shown in figure 3]:
processing, by a neural network, state information observed for one or more ports of one or more switches connected to a network to determine a mode of operation for at least one port of the one or more ports [0032, 0057, 0065, as pointed out the information handling system 100 include wireless utilization machine learning inference modulator which determines ports of activities. For example, the information handling system 100 may also include the wireless utilization machine learning inference modulator 132 that may be operably connected to the bus 108 as well as GPU utilization, port usage, energy consumption and the like. This data may be used by the wireless utilization machine learning inference modulator 132 in determining, based on the data, the predictive time and date based plural wireless control setting indices for the mobile information handling system 100. Where the power state is determined based on the result. For example, GPU usage, memory usage, battery usage or power state levels, and plurality of wireless network interface module settings. In other words, each new layer in the neural network may include a plurality of nodes representing a best guess of how each of these parameters from the iterative wireless utilization profile received may affect optimal plural wireless network interface module settings].
and causing the at least one port to operate in the mode of operation [0078, based on the determination above, specific port or interface can be disable as shown in figure 5, step 535. For example, the wireless utilization machine learning inference modulator may determine at block 535 in an example embodiment whether one or more wireless network interface modules are to be disabled in an example embodiment. In other embodiments, one or more control settings for control of the plural wireless interface modules may include various adjustments to set modules as enabled, disabled, cause some modules to enter one or more sleep modes that may save power].
As per claims 24 and 31, they do not teach or further define over the limitations recited in the rejected claims above. Therefore, claims 24 and 31 are also anticipated by Sanaullah for the same reasons set forth in the rejected claims above.
Claims 1-15 and 17-29 and 31-36 are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Kazimirsky (US Patent Application 20240073058).
As per claim 1, Kazimirsky teaches a method [500/600, fig. 5-6], comprising:
at a device [110, fig. 1A]:
processing, by a neural network [neural network switches: 0051], state information observed for one or more ports [machine learning prediction of signal conductor usage: 0071] of one or more switches connected to a network to determine a mode of operation for at least one port of the one or more ports [0055, 0057, 0059-0060, 0071, fig. 1A, fig. 2, as pointed out from the listed paragraphs and shown in figure 1A, the system is using neural network to monitor signal trough lines that connect to the ports such as port of ISC1C1-IS1N to signal lines 116, where the states of the signal determine the port states such as active and inactive states. For example, the signal conductors 140 connect the outbound switches OSW1-OSWN to the devices T1-TY. The outbound switches OSW1-OSWN may each include a different port that is connected by at least one of signal conductors 140 to at least one of the devices T1-TY].
causing the at least one port to operate in the mode of operation [0059, 0063, 0065, 00230, fig. 2, 20A, as pointed out the states of the signals to the port enable the port to operate at the states of the signal. For example, if the signal is active, then the port can be operating in active mode as well as if the signal is in inactive or idle states, then the port can be operated in that states as well as the devices connected to the port as shown in figures 2 and 20A. Below, each CPU/GPU can connect to a port of the switch: the switch(es) 206 may be implementations of the internal switches IS1C1-ISMCN and/or the outbound switches OSW1-OSWN].
PNG
media_image1.png
748
990
media_image1.png
Greyscale
[0071] The system 300 may save power and/or avoid performance loss, although there may be other benefits and/or uses of the system 300 in addition to saving power by predicting when the signal conductor 116B might be needed, and transitioning the signal conductor 116B from the inactive state L1 to the active state L0 (or waking up the signal conductor) in accordance with the predicted amount of inactive time. Because at least some workloads, such as machine learning workloads, may use the signal conductor 116B in a repeating usage pattern (e.g., over a number of iterations), the system 300 may determine this usage pattern and use the usage pattern to predict when the signal conductor 116B will next be used.
As per claim 2, Kazimirsky the state information is observed for a single port of a single switch connected to the network [0072-0073, fig. 1A, 2A, the hardware monitor 120 monitor signal of the lines 116 which is connected to the port which determines the active or inactive states of the ports].
As per claim 3, Kazimirsky teaches the state information is observed for multiple ports of a single switch connected to the network [0055, 0060, as pointed out multiple ports can be connected to the device where the signals 116 connected to the ports where the status of the signal is monitored].
As per claim 4, Kazimirsky teaches the state information is observed for a connected pair of ports respectively located on different switches connected to the network [0055, as pointed out and shown below, the signal is connected to different set of switches].
PNG
media_image2.png
399
799
media_image2.png
Greyscale
From above, signals 116 and 134 connected to multiple set of switches.
As per claim 5, Kazimirsky teaches the state information is observed for two connected pairs of ports respectively located on different switches connected to the network [055, as pointed out from the figure 1A above, the signals 134 and 116 connected to pair of different switches].
As per claim 6, Kazimirsky teaches a connection is established between the one or more ports [0051, 0055-0056, as pointed out and shown in fig. 1A, fig. 2, GPU and switch interconnect which connect to port of switches].
As per claim 7, Kazimirsky teaches the state information includes bandwidth [0114, fig. 8, as pointed out bandwidth information to perform states transition].
As per claim 8, Kazimirsky teaches the state information includes utilization [0400, processor utilization].
As per claim 9, Kazimirsky teaches the state information includes queue size [0238, queue such as memory queue utilization].
As per claim 10, Kazimirsky teaches the state information includes information associated with delayed packets in the network, the information including at least one of: a number of the delayed packets, or an average or maximum delay time among the delayed packets [0050, 0062, idle packet which is viewed as packet delay information].
As per claim 11, Kazimirsky teaches the mode of operation is selected between at least a first mode of operation and a second mode of operation [0077, 0098, select a transitional prediction between multiple modes of operation].
As per claim 12, Kazimirsky teaches the first mode of operation includes an active mode and wherein the second mode of operation includes an idle mode [0072-0073, 0077, idle period when there is not signal in conductor path 116].
As per claim 13, Kazimirsky teaches the active mode consumes more power than the idle mode [0050, 0083, fig. 5 idle mode not processing data therefore consume less power].
As per claim 14, Kazimirsky teaches the neural network is trained using reinforcement learning [0114, implementation by machine learning and training].
As per claim 15, Kazimirsky teaches the reinforcement learning is configured to maximize a cumulative reward over time [0134, 0244 machine learning include logic to perform algorithm to process data].
As per claim 17, Kazimirsky teaches the neural network is trained using a network simulator [0117, 0287, simulation for training and learning of neural network model].
As per claim 18, Kazimirsky teaches the neural network is trained on a remote datacenter [0287, 0298 network training inference such as data center application].
As per claim 19, Kazimirsky teaches causing the at least one port to operate in the mode of operation includes: causing the at least one port to toggle between a first mode of operation and a second mode of operation [0305, transition signal of the port from active to inactive mode].
As per claim 20, Kazimirsky teaches the first mode of operation includes an active mode and wherein the second mode of operation includes an idle mode [0050, switching form active mode to inactive mode or vice versa].
As per claim 21, Kazimirsky teaches receiving one or more reward signals, wherein the one or more reward signals are computed as a function of additional information observed for the one or more ports after causing the at least one port to operate in the mode of operation [0095, score information observed as shown in figure 7 which is related to power prediction].
As per claim 22, Kazimirsky teaches the neural network is dedicated for use in controlling operation of the one or more ports [0050-0051, neural network learning process control operation of the power management].
As per claim 23, Kazimirsky teaches the neural network is generalized for use in controlling operation of a plurality of ports of a plurality of switches [0050-0051, 0055, port and control via switching and signals as shown in figure 1A].
As per claims 24-29 and 31-36, they do not teach or further define over the limitations recited in the rejected claims above. Therefore, claims 24-29 and 31-36 are also anticipated by Kazimirsky for the same reasons set forth in the rejected claims above.
To help with the prosecution of this application additional rejection is given below for the independent claims.
Claims 1, 24, and 31 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Meirosu (US Patent Application 20130332762).
As per claim 1, Meirosu teaches a method [implemented by system shown in figure 2], comprising:
at a device [shown in figure 3]:
processing, by a neural network, state information observed for one or more ports of one or more switches connected to a network to determine a mode of operation for at least one port of the one or more ports [0059-0060, 0063, as pointed out an interface node can have multiple power states where the interface can be monitored to determine the power state. Where the management module can make a decision based on the power state of the interface.
causing the at least one port to operate in the mode of operation [0059-0060, 0063, as pointed out the management module can make a decision based on the power state of the interface where the interface can either be put on specific power state or remain in the power states it is already as stated].
As per claims 24 and 31, they do not teach or further define over the limitations recited in the rejected claims above. Therefore, claims 24 and 31 are also anticipated by Meirosu for the same reasons set forth in the rejected claims above.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Patronas (US 20240291738) teaches system for machine learning (ML) based network resilience and steering.
O’Toole (US 20190107877) teaches methods and systems for dynamic backup power management at a power node.
THIS ACTION IS MADE FINAL. 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 extension fee 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 VOLVICK DEROSE whose telephone number is (571)272-6260. The examiner can normally be reached on Monday-Friday 9AM-6PM.
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, Jaweed Abbaszadeh can be reached on 571.270.1640. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/VOLVICK DEROSE/Primary Examiner, Art Unit 2176
Prosecution Insights