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 .
Priority
Acknowledgment is made of applicant’s claim for foreign priority under 35 U.S.C. 119 (a)-(d).
Claim Rejections - 35 USC § 102
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 (i.e., changing from AIA to pre-AIA ) 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.
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 61-64, 67-74, 77-78, and 80 are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Li et al. (US 2024/0349082), hereinafter Li.
Regarding Claim 61, Li teaches: A method performed by a user equipment (UE) for reporting a performance of at least one machine-learning (ML) model to a network, the method comprising: utilizing at least one ML model; generating one or more reports or reportable information of a performance of the at least one ML model; and reporting the one or more reports or reportable information to a network: “The present disclosure provides embodiments/implementations to support machine learning configuration/reporting between NG-RAN and UE over radio resource control (RRC) signaling. Embodiments include new radio bearer(s) and/or a new system information block (SIB) to carry new messages including ML configurations, ML reporting and information update, ML requests, and the like. According to a ML report from a UE, the NG-RAN can also update and provide a ML model to different UEs according to the confidence level/model bias/variance of ML results in the reports from UEs” (Li ¶ 0015).
Regarding Claim 62, Li teaches: The method of claim 61, wherein the reporting is based on one or more rules or trigger events that trigger the UE to indicate to the network that the at least one ML model is not functioning within performance bounds: “In one or more embodiments, the Model Bias Threshold field sets a model bias threshold to UE when to trigger MachineLearningModelUpdateRequest. In one or more embodiments, the Model Variance Threshold field sets a model variance threshold to UE when to trigger MachineLearningModelUpdateRequest” (Li ¶ 0039-0040).
Regarding Claim 63, Li teaches: The method of claim 61, further comprising: receiving one or more configurations from a network node: “The ML model and/or configuration may be trained and/or updated by the network device 102 and/or the UE device 104” (Li ¶ 0068); wherein the reporting is based on the one or more configurations: “In one or more embodiments, the Model Bias Threshold field sets a model bias threshold to UE when to trigger MachineLearningModelUpdateRequest. In one or more embodiments, the Model Variance Threshold field sets a model variance threshold to UE when to trigger MachineLearningModelUpdateRequest” (Li ¶ 0039-0040).
Regarding Claim 64, Li teaches: The method of claim 63, wherein the one or more configurations comprise configuring the UE to perform the reporting in a periodic manner: “The ML model and/or configuration may be trained and/or updated by the network device 102 and/or the UE device 104. When the ML model and/or configuration is updated by the network device 102, the UE device 104 periodically and optionally may generate and send a ML model update request 120 to the network device 102 to request the update” (Li ¶ 0068).
Regarding Claim 67, Li teaches: The method of claim 61, wherein the one or more reports comprise at least one or more of: a value indicating a confidence level associated with output of the at least one ML model; a confidence interval: “In one or more embodiments, there may be a UE selective training/ML model update. “ConfidenceLevel” is introduced to show how network can trust the model updated by UE or the prediction results get from UE. According to the “confidenceLevel” reported by UE, NG-RAN can selectively update machine learning model to different UEs. NG-RAN can prioritize to update machine learning models to UEs with lower rate of confidenceLevel” (Li ¶ 0061); an uncertainty level associated with the output of the at least one ML model; a statistic associated with data collected within a time window; an indication that the output of the at least one ML model should be, or should no longer be, trusted or valid; an identification associating the one or more reports with the at least one ML model or with an ML feature; and an indication that input data to the at least one ML model is currently out-of-distribution.
Regarding Claim 68, Li teaches: The method of claim 61, wherein the at least one ML model facilities one or more of: channel state information (CSI) prediction, beam management, and positioning: “In communication systems, machine learning models can be applied to channel estimation, channel prediction, channel state information feedback and the like, it can be seen that the model performance of the machine learning models affects the performance of the communication system” (Li ¶ 0051).
Regarding Claim 70, Li teaches: The method of claim 61, wherein the one or more reports are sent to the network with at least one of the following: output of the at least one ML model: “. Once the UE device 104 has executed the ML model and generated corresponding outputs, the UE device 104 may generate and send a ML report 118 to the network device 102. The ML report 118 may indicate predictions prior to the ML execution 116, outcomes of the ML execution 116 (e.g., actual outputs of the ML execution 116), and requested or selected actions (e.g., an action space) for the UE device 104 to perform based on the ML execution 116” (Li ¶ 0068); or a channel state information (CSI) report.
Regarding Claim 71, Li teaches: The method of claim 61, wherein the performance of the at least one ML model is monitored for different functionalities that trigger different actions: “the processing circuitry is further configured to: determine at least one of a machine learning model bias, a machine learning model variance, a machine learning model confidence level, or feedback associated with use of the machine learning model; generate, based on the at least one of the machine learning model bias, the machine learning model variance, the machine learning model confidence level, or the feedback, an update to the machine learning configuration for use by the UE device; and cause the node B device to transmit the update to the machine learning configuration to the UE device” (Li ¶ 0160).
Regarding Claim 72, Li teaches: The method of claim 61, further comprising: indicating, to the network, a capability of the UE to analyze the performance of the at least one ML model: “A new field “ue-capabilityML-Information” in UEInformationRequest message is also used to send the machine learning capability of UE to the network. The present disclosure considers categories of UE capability: (1) Hardware capability; (2) Machine learning capability . . . For machine learning capability, it is used to indicate: what type of machine learning model UE can support (e.g. CNN, RNN, RL, Classification, regression, etc.) and for each machine learning model, it includes: maximum model size, training capability, e.g. supported SW library, inference capability, e.g. supported SW library, etc. A Machine learning capability field is used to indicate NG-RAN the machine learning capability of UE” (Li ¶ 0021).
Regarding Claim 73, Li teaches: The method of claim 61, further comprising, indicating, to the network, one or more reporting methods that the UE supports for reporting the performance of the at least one ML model: “The NG-RAN may generate and send a UE capability inquiry to the UE to request ML and hardware capabilities of the UE, and the UE may respond by providing the NG-RAN with its ML and hardware capabilities. Based on the ML and hardware capabilities of the UE, the NG-RAN may generate a ML configuration for a ML model of a service requested by the UE, and may send the ML model and ML configuration to the UE for use by the UE. Once the UE has implemented the ML model and generated results, the UE may send a ML report to the NG-RAN to report the predictions, outcomes, and action space (e.g., actions that the UE will or is requesting to perform as a result of the ML outcomes)” (Li ¶ 0017).
Regarding Claim 74, Li teaches: The method of claim 61, wherein the at least one ML model is configured by the network in at least one of the following aspects: model architecture; a parameter used to control a learning process of the at least one ML model; and an optimization objective for the at least one ML model: “In one or more embodiments, the NG-RAN (e.g., a gNB device) may generate and send a ML capability indication to UEs to indicate that the NG-RAN supports ML (e.g., and may facilitate ML operations at the UE). A UE may respond with an interest indication (e.g., service registration) that indicates for which service a UE requests a ML model for use by the UE. The NG-RAN may generate and send a UE capability inquiry to the UE to request ML and hardware capabilities of the UE, and the UE may respond by providing the NG-RAN with its ML and hardware capabilities. Based on the ML and hardware capabilities of the UE, the NG-RAN may generate a ML configuration for a ML model of a service requested by the UE, and may send the ML model and ML configuration to the UE for use by the UE. Once the UE has implemented the ML model and generated results, the UE may send a ML report to the NG-RAN to report the predictions, outcomes, and action space (e.g., actions that the UE will or is requesting to perform as a result of the ML outcomes)” (Li ¶ 0017).
Regarding Claim 77, Li teaches: The method of claim 61, further comprising, in response to a command from the network, train or retrain the at least one ML model: “the RRC message further comprises a service type indicator indicating that the UE device is not required to synchronize the machine learning configuration with the node B device and that the UE device is permitted to train a machine learning model associated with an action space of the machine learning configuration” (Li ¶ 0155).
Regarding Claim 78, Li teaches: The method of claim 61, further comprising, in response to a command from the network, update the at least one ML model: “As for federated learning, UE will update and iterate local machine learning model based on its own environment and input/output, UE can also report the updated machine learning parameters generated by local nodes to NG-RAN, so that NG-RAN can update the centralized model accordingly. This type of reporting can be called a model parameter update” (Li ¶ 0033).
Regarding Claim 80, Li teaches: A user equipment (UE) for performing a machine learning (ML) model and reporting to a network, comprising: processing circuitry: “The UE device 104 may include any suitable processor-driven device including, but not limited to, a mobile device or a non-mobile, e.g., a static device” (Li ¶ 0072) configured to perform operations comprising: utilizing at least one ML model; generating one or more reports or reportable information of a performance of the at least one ML model; and reporting the one or more reports or reportable information to a network: “The present disclosure provides embodiments/implementations to support machine learning configuration/reporting between NG-RAN and UE over radio resource control (RRC) signaling. Embodiments include new radio bearer(s) and/or a new system information block (SIB) to carry new messages including ML configurations, ML reporting and information update, ML requests, and the like. According to a ML report from a UE, the NG-RAN can also update and provide a ML model to different UEs according to the confidence level/model bias/variance of ML results in the reports from UEs” (Li ¶ 0015); and power supply circuitry configured to supply power to the processing circuitry: “For hardware capability, in general, it is used to indicate whether the hardware of UE chip can/want support machine learning or not. Details may also include: Chip type, max battery capacity, UE's current battery status, batching data size, etc” (Li ¶ 0021).
Claim Rejections - 35 USC § 103
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 (i.e., changing from AIA to pre-AIA ) 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.
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.
The factual inquiries 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.
Claims 65-67 are rejected under 35 U.S.C. 103 as being unpatentable over Li as applied to claim 61 above in further view of Wang et al. (US 2021/0326726), hereinafter Wang.
Regarding Claim 65, Li teaches: The method of claim 63.
Li does not teach: the one or more configurations comprise configuring the UE to perform the reporting in an aperiodic manner.
Regarding Claim 65, Wang teaches: the one or more configurations comprise configuring the UE to perform the reporting in an aperiodic manner: “The ML algorithms may be trained with training datasets that are produced through periodic and/or aperiodic data collection at one or more nodes. In various aspects, measurement data collection serves as input to the ML modules” (Wang ¶ 0031).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the disclosure of Li with Wang for the purpose of reducing data sent between UE and Base station. According to Wang: “For example, by feeding back sampled data that corresponds to a prediction error and/or a subset of the sampled data that corresponds to a correct prediction, the ML-based system can reduce the amount of sampled data that needs to be fed back between the UE and the BS” (Wang ¶ 0034).
Regarding Claim 66, Li teaches: The method of claim 63.
Li does not teach: the one or more configurations comprise configuring the UE to perform the reporting in response to a detected performance drift of the at least one ML model.
Regarding Claim 67, Wang teaches: the one or more configurations comprise configuring the UE to perform the reporting in response to a detected performance drift of the at least one ML model: “The present disclosure provides techniques for the UE to feed back sampled data of the channel properties that correspond to an incorrect prediction (or generally referred to as “a prediction error” herein) to the BS to serve as more insightful data that helps improve the ML algorithms. To update the neural network(s), it may be more valuable to feed back the sampled data when a prediction error occurs. In various aspects, the UE is configured to teed back a subset of the sampled data that corresponds to a correct prediction to the BS since this sampled data may not be as indicative as to how to improve the ML algorithms other than to reassure that the ML algorithms are performing as expected” (Wang ¶ 0033).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the disclosure of Li with Wang for the purpose of reducing data sent between UE and Base station. According to Wang: “For example, by feeding back sampled data that corresponds to a prediction error and/or a subset of the sampled data that corresponds to a correct prediction, the ML-based system can reduce the amount of sampled data that needs to be fed back between the UE and the BS” (Wang ¶ 0034).
Regarding Claim 69, Li teaches: The method of claim 61.
Li does not teach: the one or more reports are reported for one or more granularities, the one or more granularities comprising one or more of: per frequency range; per sub-band; per set of sub-bands; per Bandwidth Part; per cell; per reference signal beam; and per UE speed.
Regarding Claim 69, Wang teaches: the one or more reports are reported for one or more granularities, the one or more granularities comprising one or more of: per frequency range; per sub-band: “In some aspects, the transceiver configured to communicate the report may be further configured to transmit, to the BS in a first subband of a plurality of subbands, sampled data for updating the machine learning-based network. In some aspects, the first subband includes a plurality of physical uplink shared channels (e.g., PUSCHs) multiplexed in at least one of time or frequency in a first portion of a first time period, and the transceiver configured to transmit the sampled data may be further configured to transmit the sampled data in one or more PUSCHs of the plurality of PUSCHs. In some aspects, the sampled data includes a feedback pairing of at least one signal measurement in the set of received signal measurements and the output of the machine learning-based network associated with the at least one signal measurement” (Wang ¶ 0036); per set of sub-bands; per Bandwidth Part; per cell; per reference signal beam; and per UE speed.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the disclosure of Li with Wang for the purpose of reducing data sent between UE and Base station. According to Wang: “For example, by feeding back sampled data that corresponds to a prediction error and/or a subset of the sampled data that corresponds to a correct prediction, the ML-based system can reduce the amount of sampled data that needs to be fed back between the UE and the BS” (Wang ¶ 0034).
Claims 75-76 and 79 are rejected under 35 U.S.C. 103 as being unpatentable Li as applied to claim 61 above in further view of Pezeshki et al (US 2022/0150727), hereinafter Pezeshki.
Regarding Claim 75, Li teaches: The method of claim 61.
Li does not teach: in response to a command from the network, stop using the at least one ML model.
Regarding Claim 75, Pezeshki teaches: in response to a command from the network, stop using the at least one ML model: “The UEs may apply the appropriate machine learning model (or updated machine learning) based on the current serving TRP. For instance, when there is a change in a serving TRP (e.g., from a first TRP to a second TRP) for the UE 1 (or UE 2), the BS may notify the UE 1 (or UE 2) to disable a previous machine learning model (associated with the first TRP) or enable a new machine learning model (associated with the second TRP)” (Pezeshki ¶ 0151).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the disclosure of Li with Pezeshki for the purpose of allowing a UE to change between different ML models for different TRPs. According to Pezeshki: “The computer-readable medium may include codes executable to provide one or more machine learning models; generate an indication of one or more TRPs for which the one or more machine learning models are applicable; and at least one of output for transmission to a UE or obtaining from the UE the one or more machine learning models and the indication of the one or more TRPs for which the one or more machine learning models are applicable” (Pezeshki ¶ 0032).
Regarding Claim 76, Li teaches: The method of claim 61.
Li does not teach: in response to a command from the network, start or re-start using a non-ML model.
Regarding Claim 76, Pezeshki teaches: in response to a command from the network, start or re-start using a non-ML model: “The UEs may apply the appropriate machine learning model (or updated machine learning) based on the current serving TRP. For instance, when there is a change in a serving TRP (e.g., from a first TRP to a second TRP) for the UE 1 (or UE 2), the BS may notify the UE 1 (or UE 2) to disable a previous machine learning model (associated with the first TRP) or enable a new machine learning model (associated with the second TRP)” (Pezeshki ¶ 0151).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the disclosure of Li with Pezeshki for the purpose of allowing a UE to change between different ML models for different TRPs. According to Pezeshki: “The computer-readable medium may include codes executable to provide one or more machine learning models; generate an indication of one or more TRPs for which the one or more machine learning models are applicable; and at least one of output for transmission to a UE or obtaining from the UE the one or more machine learning models and the indication of the one or more TRPs for which the one or more machine learning models are applicable” (Pezeshki ¶ 0032).
Regarding Claim 79, Li teaches: The method of claim 61.
Li does not teach: in response to a command from the network, switch from the at least one ML model to a different at least one ML model.
Regarding Claim 79, Pezeshki teaches: in response to a command from the network, switch from the at least one ML model to a different at least one ML model: “The UEs may apply the appropriate machine learning model (or updated machine learning) based on the current serving TRP. For instance, when there is a change in a serving TRP (e.g., from a first TRP to a second TRP) for the UE 1 (or UE 2), the BS may notify the UE 1 (or UE 2) to disable a previous machine learning model (associated with the first TRP) or enable a new machine learning model (associated with the second TRP)” (Pezeshki ¶ 0151).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the disclosure of Li with Pezeshki for the purpose of allowing a UE to change between different ML models for different TRPs. According to Pezeshki: “The computer-readable medium may include codes executable to provide one or more machine learning models; generate an indication of one or more TRPs for which the one or more machine learning models are applicable; and at least one of output for transmission to a UE or obtaining from the UE the one or more machine learning models and the indication of the one or more TRPs for which the one or more machine learning models are applicable” (Pezeshki ¶ 0032).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BRADLEY DAVIS LYTLE whose telephone number is (703)756-4593. The examiner can normally be reached M-F 8:00 AM - 4:00 PM EST.
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, Kwang bin Yao can be reached at 571-272-3182. 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.
/B.D.L./Examiner, Art Unit 2473
/BRADLEY D LYTLE JR./Examiner, Art Unit 2473
/KWANG B YAO/Supervisory Patent Examiner, Art Unit 2473
Prosecution Insights