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 a response to communications dated 09/30/2024. Claims 1-6, 8, 11-14, 19, 24-25, 27, 29, 31-33, 45, 53, 62, 71, and 74-75 are pending in the application.
Information Disclosure Statement
The information disclosure statement filed 09/30/2024 complies with the provisions of 37 CFR 1.97, 1.98 and MPEP § 609. It has been considered and placed in the application file.
Claim Objections
Claims 5, 25, and 32-33 are objected to because of the following informalities:
As per claim 5, line 10, “learning model;” should be changed to --learning model; and --.
As per claim 25, line 12, “associated report;” should be changed to --associated report; and--.
As per claim 32, line 12, “associated report;” should be changed to --associated report; and--.
As per claim 33, line 3, “signalling;” should be changed to --signalling; and--.
Appropriate correction is required.
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 1-6, 8, 13-14, 19, 24-25, 27, 29, 31-33, 53, 71, and 74-75 are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Pantelidou et al. (US 2023/0297882) (hereinafter “Pantelidou”).
Regarding claim 1, in accordance with Pantelidou reference entirety, Pantelidou teaches a method performed by a user equipment (UE), for resolving a performance problem in a machine learning model (FIG. 4 and para [0112]: “… the UE can activate the ML model if the UE measures that its throughput drops below a threshold or if the number of handover failures (at a certain location) exceeds a certain threshold … .”), the method comprising:
detecting one or more performance problems (State Change) in the machine learning model (execute and/or train the ML model) performed by the UE (para [9114]: “… UE detects a State Change that affects its ability to execute and/or train the ML model … .”); and
in response to the detecting (In this situation), performing one or more resolutions actions (declare to network/autonomously falls back) (para [0114]: “… In this situation, the UE can declare to the network it is not able to full ML processing (for instance using Option a or Option b), and UE autonomously falls back to Default Behavior.”).
Regarding claim 2, in addition to features recited in base claim 1 (see rationales discussed above), Pantelidou also teaches receiving a configuration (a message) from a network node (gNB), the configuration (a message) comprising one or more parameters for at least one of: monitoring performance in the machine learning model; detecting a performance problem in the machine learning model; and performing one or more resolution actions (Activate ML model) (para [0106]: “… network (gNB) sends to UE a message “Activate ML model” with which network activates a ML model (e.g., ML model m) to solve a certain problem pm. The model chosen by the network depends on the previously indicated ML capability of the UE … .” It is noted that the claim is drafted in an alternative format not requiring all recitations but one of the recitations).
Regarding claim 3, in addition to features recited in base claim 1 (see rationales discussed above), Pantelidou also teaches monitoring performance in the machine learning model performed (observed) by the UE (para [0112]: “Trigger based (active the ML model based on some event at the UE configured by the network). This Activation mode could be triggered if a certain/event measurement is observed by the UE … optimize internal parameters)”.).
Regarding claim 4, in addition to features recited in base claim 3 (see rationales discussed above), Pantelidou also teaches wherein the monitoring comprises generating one or more metrics (different value) related to an error (task/problem) or accuracy of the machine learning model (para [0114]: “… In this situation, the UE can declare to the network it is not able to full ML processing (for instance using Option a or Option b), and UE autonomously falls back to Default Behavior.” In Option A discussed in para [0095]: “… UE may send a different value of its ability to gNB … the UE may also update its Default Behavior for a given task/problem and inform the network thereabout … for the task.”).
Regarding claim 5, in addition to features recited in base claim 4 (see rationales discussed above), Pantelidou also teaches wherein the one or more metrics measure at least one of: an error calculated based on one or more outputs of the machine learning model and a corresponding parameter the machine learning model is configured to estimate and/or predict; an absolute or relative error; an accuracy calculated based on one or more outputs of the machine learning model and a corresponding parameter the machine learning model is configured to estimate and/or predict; an absolute or relative accuracy; a value in percentage that indicates a confidence level of an outcome of the machine learning model; a representation of a distribution that indicates a confidence level of an outcome of the machine learning model; a representation of a confidence interval; one or more statistics of data collected within a time window related to the machine learning model; and one or more key performance indicators indicating an ability, quality, power or accuracy of the machine learning model to estimate a parameter in comparison to an actual value of the parameter (para [0031]: “… functions of measurements, corresponding to a certain network model/behavior or proper, described as: “when the serving cell RSRP is in a certain range”, “how many times serving cell RSRP has fallen into predetermined range”, “when packet delay exceeds a certain threshold”, “when interference power received exceeds a certain threshold” to name a few.” It is noted that the claim is drafted in an alternative format not requiring all recitations but one of the recitations).
Regarding claim 6, in addition to features recited in base claim 4 (see rationales discussed above), Pantelidou also teaches wherein the one or more metrics are configured by a network node (network) (para [0030] and thereinafter: “… the network to trigger the UE to monitor through measurements or pre-configured functions of measurements’ the process of learning of the provided ML model, and directly use those measurements to train the ML model. The target output by the UE is the trained ML model.”).
Regarding claim 8, in addition to features recited in base claim 4 (see rationales discussed above), Pantelidou also teaches wherein the one or more metrics are associated with a capability which the UE reports to a network node (para [0091]: "There are different methods with which the UE may indicate its ML ability to the network. A mere reuse of the UE capability IE is insufficient to capture ML ability since it is a static field indicated once to the network during the registration process to inform all the details of the UE capabilities. UE capability IE can indicate whether the UE has the capability to execute (or even train itself) an ML algorithm, i.e., whether or not it is equipped with the necessary resources. In addition, according to some example embodiments, a UE is able to indicate its ML ability in the course of time. An ML capable UE may become unable to execute the trained model if its current state does not allow it, e.g., if its memory is getting full, or if its battery drops below a threshold, or ML performance overspends processing capabilities of the UE processor.").
Regarding claim 13, in addition to features recited in base claim 3 (see rationales discussed above), Pantelidou also teaches wherein the UE starts monitoring the performance of the machine learning model when the UE transitions to Radio Resource Control Connected, RRC_CONNECTED (para [0105]: “… the UE is initialized and has indicated to network its ML capabilities and Default Behavior … At initialization, it is assumed that the UE is able to execute and/or train the ML models.” The UE initialized is equated to “the UE transitions to Radio Resource Control Connected.”).
Regarding claim 14, in addition to features recited in base claim 13 (see rationales discussed above), Pantelidou also teaches wherein the UE monitors performance of the machine learning model while the UE is in RRC_CONNECTED, and wherein the UE stops monitoring the performance of the machine learning model when the UE transitions to Radio Resource Control Idle, RRC_IDLE, upon reception of an RRC Release message or when the UE transitions to Radio Resource Control Inactive, RRC_INACTIVE, upon reception of an RRC Release message with a suspend configuration (para [0116]: “… The UE receiving the De-Activate ML model message … acknowledges the deactivation with an “Accept” Response to the network.”).
Regarding claim 19, in addition to features recited in base claim 3 (see rationales discussed above), Pantelidou also teaches wherein the UE performs the monitoring periodically according to an assessment period, wherein the assessment period is configured by a network to the UE or the assessment period is derived by the UE based on one or more parameters (para [0090]: "UE indication at any time to the network of its (in-)ability to execute and/or train an ML model if its state does not allow (full) ML processing. The UE may indicate its inability to the network on its own (either periodically or trigger-based, i.e. when the UE becomes unable to execute and/or train the ML model), or the network may request the UE to provide an indication of its (in-)ability.").
Regarding claim 24, in addition to features recited in base claim 19 (see rationales discussed above), Pantelidou also teaches wherein the monitoring periodically is based on one or more parameters (para [0031]: For example, a UE can be configured by the network to monitor functions of measurements, corresponding to a certain network model/ behavior or property, described as: "when the serving cell RSRP is in a certain range", "how many times serving cell RSRP has fallen into predefined range", "when packet delay exceeds a certain threshold", "when interference power received exceeds a certain threshold" to name a few.").
Regarding claim 25, in addition to features recited in base claim 24 (see rationales discussed above), Pantelidou also teaches wherein the one or more parameters comprise one or more of: the machine learning model or features of the machine learning model to monitor (para [0031]: For example, a UE can be configured by the network to monitor functions of measurements, corresponding to a certain network model/ behavior or property, described as: "when the serving cell RSRP is in a certain range", "how many times serving cell RSRP has fallen into predefined range", "when packet delay exceeds a certain threshold", "when interference power received exceeds a certain threshold" to name a few.” It is also noted that the claim is drafted in an alternative format not requiring all recitations but one of the recitations); whether to monitor performance across multiple serving cells or only within a current serving cell; a time window for the monitoring; if the monitoring is to be stopped or suspended by the UE when the UE transitions to RRC IDLE and/or RRC INACTIVE state from RRC_CONNECTED state; if the monitoring is to be maintained by the UE when the UE transitions to RRC_CONNECTED; if the monitoring is performed both in discontinuous reception and non-discontinuous reception operation by the UE; if the monitoring is stopped when uplink, UL, timing alignment is lost; if the monitoring is stopped when the UE has lost UL synchronization; if the monitoring is stopped by the expiry of a timeAlignmentTimer; if the monitoring is stopped and then continues if an event is triggered; if the monitoring is stopped and the UE will try to achieve UL synchronization and, subsequently, transmit the associated report; information about a periodicity and/or time domain offset based on which the UE derives which time domain resources are allowed to be used for monitoring the performance of the machine learning model.
Regarding claim 27, in addition to features recited in base claim 2 (see rationales discussed above), Pantelidou also teaches wherein the UE performs the monitoring aperiodically (trigger-based) (para [0090]: "UE indication at any time to the network of its (in-)ability to execute and/or train an ML model if its state does not allow (full) ML processing. The UE may indicate its inability to the network on its own (either periodically or trigger-based, i.e. when the UE becomes unable to execute and/or train the ML model), or the network may request the UE to provide an indication of its (in-)ability.").
Regarding claim 29, in addition to features recited in base claim 4 (see rationales discussed above), Pantelidou also teaches receiving a request (input) from a network (the operator) to perform aperiodic monitoring (para [0119]: "Another trigger to de-activate the ML model in the UE may be an input from the operator. For example, the operator may have decided that the UE should not execute and/or train the ML model any more.").
Regarding claim 31, in addition to features recited in base claim 27 (see rationales discussed above), Pantelidou also teaches wherein the aperiodic monitoring is based on one or more parameters (para [0090]: "UE indication at any time to the network of its (in-)ability to execute and/or train an ML model if its state does not allow (full) ML processing. The UE may indicate its inability to the network on its own (either periodically or trigger-based, i.e. when the UE becomes unable to execute and/or train the ML model), or the network may request the UE to provide an indication of its (in-)ability.").
Regarding claim 32, in addition to features recited in base claim 31 (see rationales discussed above), Pantelidou also teaches wherein the one or more parameters comprise one or more of: the machine learning model or features of the machine learning model to monitor; whether to monitor performance across multiple serving cells or only within a current serving cell; a time window (time indication) for the monitoring (para [0111]: "Alternatively, a time indication (timer) in the message can tell the UE to activate a trained ML model for execution and/or training with some time delay after reception of the Activation message." It is also noted that the claim is drafted in an alternative format not requiring all recitations but one of the recitations). if the monitoring is to be stopped or suspended by the UE when the UE transitions to RRC IDLE and/or RRC INACTIVE state from RRC_CONNECTED state; if the monitoring is to be maintained by the UE when the UE transitions to RRC_CONNECTED; if the monitoring is performed both in discontinuous reception and non-discontinuous reception operation by the UE; if the monitoring is stopped when uplink, UL, timing alignment is lost; if the monitoring is stopped when the UE has lost UL synchronization; if the monitoring is stopped by the expiry of a timeAlignmentTimer; if the monitoring is stopped and then continues if an event is triggered; if the monitoring is stopped and the UE will try to achieve UL synchronization and, subsequently, transmit the associated report; and information about a periodicity and/or time domain offset based on which the UE derives which time domain resources are allowed to be used for monitoring the performance of the machine learning model.
Regarding claim 33, in addition to features recited in base claim 27 (see rationales discussed above), Pantelidou also teaches wherein the configuration of the aperiodic monitoring is done by one or more of: RRC (para [0108]: “RRC signaling may be used.” It is also noted that the claim is drafted in an alternative format not requiring all recitations but one of the recitations); Medium Access Control Control Element, MAC CE; layer one, L1, signalling; and Downlink Control Information, DCI, format.
Regarding claim 53, in addition to features recited in base claim 1 (see rationales discussed above), Pantelidou also teaches wherein the one or more resolution actions comprises the UE performing training or re-training of the machine learning model (para [0111]: "Alternatively, a time indication (timer) in the message can tell the UE to activate a trained ML model for execution and/or training with some time delay after reception of the Activation message.").
Regarding claim 71, in accordance with Pantelidou reference entirety, Pantelidou teaches a method performed by a network node for configuring a user equipment (UE) to monitor a machine learning model (FIG. 6; and paras [0116] to [0119]: “FIG. 6 shows an example where the network detects that the current ML model used by the UE is suboptimal.”), the method comprising:
transmitting to the UE a configuration (De-Activate ML model message) indicating the UE to perform detection of a performance problem in a machine learning model and to perform one or more resolution actions (para [0116]: "… if the network observes that the current ML model does not perform well, for example if the network conditions have changed. In this case, as shown in FIG. 6, the network upon detection of suboptimal operation of ML model m for a given problem pm, sends a De-Activate ML model message to the UE. The network may signal to the UE to De-Activate multiple ML models related to different problems. The UE receiving the De-Activate ML model message reverts to Default Behavior for all the indicated problems and acknowledges the deactivation with an “Accept” Response to the network.").
As per claim 74, the claim appears to call for a user equipment having functional limitations variously and essentially similar to method steps of method claim 1. Thus, it is anticipated by Pantelidou for the same rationales applied to method claim 1 as above discussed and in reference to apparatus having details of elements shown/inherent included in FIGs. 8-10 as discussed in paras [0122] and [0128].
As per claim 75, the claim appears to call for a network node having functional limitations variously and essentially similar to method steps of method claim 71. Thus, it is anticipated by Pantelidou for the same rationales applied to method claim 71 as above discussed and in reference to apparatus having details of elements shown/inherent included in FIGs. 12-17 as discussed in paras [0135] and [0142] and [0148].
Allowable Subject Matter
Claims 11-12, 45 and 62 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.
The following is a statement of reasons for the indication of allowable subject matter: The prior art of record, considered individually or in combination, appears to fail to fairly show or suggest a claim invention of base claim 1 and further limits with novel and unobvious limitations of “wherein the one or more metrics are compared to a reference value of the one or more metrics associated to a performance indicator of a feature associated to the machine learning model,” as recited in claims 11-12; “wherein if the UE detects a performance problem then the UE increments a failure counter, and if the number of performance problems reaches a determined maximum number for the failure counter, then a failure timer is started; and while the failure timer is running the UE counts a number of times that performance of the machine learning model becomes better than an unacceptable level; and if the counter number of times that performance of the machine learning model becomes better than an unacceptable level reaches a predetermined number, then the UE stops the failure timer and resets the failure counter,” as recited in claim 45; and “wherein the one or more resolution actions comprises: the UE resetting the machine learning model; deleting one or more outputs of the machine learning model; stopping one or more ongoing processes that use one or more outputs of the machine learning model; and the UE resetting at least one protocol layer or protocol entity where the machine learning model is being used,” as recited in claim 62, structurally and functionally interconnected in a manner as claimed.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Olgiati et al. (US 11,449,798).
Olgiati et al. (WO 2021/067385).
Faulhaber, Jr. et al. (US 2019/0156247).
Otterbach et al., CHAMELEON: A SEMI-AUTOML FRAMEWORK TARGETING QUICK AND SCALABLE DEVELOPMENT AND DEPLOYMENT OF PRODUCTION-READY ML SYSTEMS FOR SMES, arXiv, 6 pages, May 8, 2021.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to FRANK DUONG whose telephone number is (571)272-3164. The examiner can normally be reached 7:00AM-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, MICHAEL THIER can be reached at 571-272-2832. 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.
Applicant is encouraged to submit a written authorization for Internet communications (PTO/SB/439, http://www.uspto.gov/sites/default/files/documents/sb0439.pdf) in the instant patent application to authorize the examiner to communicate with the applicant via email. The authorization will allow the examiner to better practice compact prosecution. The written authorization can be submitted via one of the following methods only: (1) Central Fax which can be found in the Conclusion section of this Office action; (2) regular postal mail; (3) EFS WEB; or (4) the service window on the Alexandria campus. EFS web is the recommended way to submit the form since this allows the form to be entered into the file wrapper within the same day (system dependent). Written authorization submitted via other methods, such as direct fax to the examiner or email, will not be accepted. See MPEP § 502.03.
/FRANK DUONG/Primary Examiner, Art Unit 2474 September 4, 2026
Prosecution Insights