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 .
Response to Amendment
Claims 1-6 and 8-21 are pending for examination.
Claims 1, 8 and 16 are independent Claims.
Claims 1-6 and 8-21 are rejected under 35 U.S.C. §112(a) and §103.
Allowable Subject Matter
Claims 1-6 are allowed.
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.
Claim(s) 8 and 16-18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lee et al. (U.S. 2024/0141746 hereinafter Lee) in view of Kim et al. (U.S. 2025/0097158 hereinafter Kim) in further view of Vandikas et al. (U.S. 2022/0052925 hereinafter Vandikas) in further view of Gao et al. (U.S. 2015/0245360 hereinafter Gao).
As Claim 8, Lee teaches a method, comprising:
determining, from the first group of RICs, a second group of RICs that satisfy a dissimilarity criterion (Lee (¶0229, fig. 26 step S2602), server groups first UEs corresponding to some of the plurality of UEs into a first UE group based on first location information), wherein the dissimilarity criterion identifies a dissimilarity between respective first datasets of respective first RICs of the first group of RICs (Lee (¶0229, fig. 26 step S2602), server groups first UEs corresponding to some of the plurality of UEs into a first UE group based on first location information) and respective selected datasets of selected RICs of the RICs (Lee (¶0230, fig. 26 step S2603), server groups second UEs, that correspond to some of the plurality of UEs and do not belong to the first UE group into a second UE group based on second location information), wherein respective selected RICs of the selected RICs correspond to respective regions of the open radio access network, wherein the dissimilarity criterion measures a lack of overlap between the respective selected datasets that corresponds to the respective regions (Lee (¶0009 line 5-12, ¶0238), second group of UEs do not belong to the first group UEs because they are associated with second location (non-overlap between second position and first position) and second channel information), and wherein the selected RICs are selected for a current round of federated learning of a machine learning model (Lee (¶¶ 0234-0235, fig. 26 step S2607, S2608), system receives local model parameters);
instructing the second group of RICs to perform federated learning of the machine learning model on respective second datasets (Lee (¶0231, fig. 26 item S2604, ¶0233 fig. 26 item S2606), server transmits resource allocation for making local machine learning model(s)), to produce respective local machine learning models (Lee (¶0234, fig. 26 item S2607), server receives local parameter form the first UEs);
based on receiving respective indications of the respective local machine learning models, generating a global machine learning model based on the respective indications of the respective local machine learning models (Lee (¶0236), server performs post-processing in order to learn a global model from local updates); and
sending an indication of the global machine learning model to the RICs (Lee (¶0237, fig. 26 item S2610, ¶0214 last 7 lines), system transmits global mode to each of the edge device)
Lee may not explicitly disclose:
Determining, by a system comprising at least one processor, a first group of radio access network intelligent controllers (RICs) that satisfy a near real-time criterion of RICs of an open radio access network that satisfy a performance capability criterion, wherein the open radio access network is configured to facilitate broadband cellular communications with user equipment;
Kim teaches:
determining a first group of radio access network intelligent controllers (RICs) that satisfy a near real-time criterion of RICs (Kim (¶0145 line 2-end), “for each iteration, the candidate training devices report their computation resource available for the training task of the FL server”. “for each iteration” is construed as near real-time criterion) of an open radio access network (Kim (¶0037 line 1-4), wireless devices perform communication using radio access technology) that satisfy a performance capability criterion (Kim (¶0145 line 3-9, fig. 7), training devices reports their computation resource availability to the FL server. The FL server makes the training device selection based on the reports), wherein the open radio access network is configured to facilitate broadband cellular communications with user equipment (Kim (¶0039 line 5-7), the network may be configured using 3G, 4G, 5G and beyond 5G network);
Lee teaches the grouping of selected machine learning devices. Kim teaches the selection of selected machine learning device based on device capabilities. 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 grouping of selected machine learning devices of Lee with a mechanism to selected machine learning devices based on device capabilities, with a reasonable expectation of success. The motivation would be to allow “the consumed the computation, communication and energy resources over the AI/ML endpoints are optimized” (Kim (¶0143 last 3 lines)).
Lee in view of Kim may not explicitly disclose:
wherein the RICs use the global machine learning model to predict a network performance metric of the open radio access network, and to adjust the broadband cellular communications based on the network performance metric.
Vandikas teaches:
wherein the RICs use the global machine learning model to predict a network performance metric of the open radio access network (Vandikas (¶0083 last 3 lines, fig. 3 item 300B), local models are updated with a global model), and to adjust the broadband cellular communications based on the network performance metric (Vandikas (¶0084, fig. 3 item 310 and 320), edge node predicts network communication performance and implement remedial or preventive measure to account for the network communication performance at the edge node being predicted to decrease).
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 grouping of selected machine learning devices of Lee in view of Kim with local model prediction and remedial/preventive measure of Vandikas, with a reasonable expectation of success. The motivation would be to allow the system to perform “one or more remedial or preventative measures to account for the network communication performance at the edge node 10-1 being predicted to decrease” (Vandikas (¶0083 last 4 lines).
Lee in view of Kim in further view of Vandikas may not explicitly disclose:
wherein the dissimilarity criterion is based on respective physical resource block utilizations of the respective first RICs and the respective selected RICs, wherein the dissimilarity criterion is further based on respective average modulation and coding scheme values for respective scheduled user equipment in respective distributed units that are connected to the respective first RICs and the respective selected RICs, and
Gao teaches:
wherein the dissimilarity criterion (Gao (¶0083 line 1-5), “If there are multiple candidate MU groups, the multiple candidate MU groups are scored (916) (e.g., in accordance with a group of scoring parameters). A candidate MU group with a best (e.g., highest) score is selected (916) as the MU group.”) is based on respective physical resource block utilizations of the respective first RICs and the respective selected RICs, wherein the dissimilarity criterion is further based on respective average modulation and coding scheme values for respective scheduled user equipment in respective distributed units that are connected to the respective first RICs and the respective selected RICs (Gao (¶0072 line 2-9), “it is determined (904) that the first wireless device satisfies one or more criteria for inclusion in the MU group. Examples of possible criteria include, but are not limited to, a minimum queue length (resource block utilization), a minimum signal-to-interference-and-noise ratio (SINR), a minimum modulation-and-coding scheme (MCS) (e.g., a minimum MCS data rate) (modulation and coding scheme), and a maximum Doppler profile”), and
Lee teaches a system/method to group devices based on distance. Gao teaches additional parameters for grouping devices. 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 grouping of selected machine learning devices of Lee in view of Kim in further view of Vandikas with grouping parameters of Gao, with a reasonable expectation of success. The motivation would be to allow the system to “to determine channel statistics in order to allow improved operation of the system” (Gao (¶0004 last 2 lines)).
As Claim 16, the Claim is rejected for the same reasons as Claim 8.
As Claim 17, besides Claim 16, Lee and Kim in view of Vandikas in further view of Gao teaches wherein the operations further comprise:
deploying the machine learning model to a first agent of the agents, wherein the first agent is separate from the second group of agents (Lee (¶0237, fig. 26 item S2610, ¶0214 last 7 lines), system transmits global mode to each of the edge device).
As Claim 18, besides Claim 16, Lee and Kim in view of Vandikas in further view of Gao teaches wherein an input to the machine learning model comprises an indication of network utilization metrics (Lee (¶0231), server transmit first resource allocation information to the first UEs).
Claim(s) 9-15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lee, Kim and Vandikas in view of Gao in further view of Larsson et al. (U.S. 2022/0230062 hereinafter Larsson).
As Claim 9, besides Claim 8, Lee and Kim in view of Vandikas in further view of Gao does not explicitly disclose:
wherein the federated learning is first federated learning, and further comprising:
in response to detecting that an inference accuracy by the global machine learning model is below a defined inference accuracy specified by a performance criterion,
selecting, by the system, a third group of RICs of the RICs with which to perform second federated learning for an update of the global machine learning model.
Larsson teaches:
wherein the federated learning is first federated learning, and further comprising:
in response to detecting that an inference accuracy by the global machine learning model is below a defined inference accuracy specified by a performance criterion (Larsson (¶0048 line 4-9, fig. 2 item 206), when the evaluation metric crosses a threshold, number of client computing devices are increased),
selecting, by the system, a third group of RICs of the RICs with which to perform second federated learning for an update of the global machine learning model (Larsson (¶0048 line 4-9, fig. 2 item 206), when the evaluation metric crosses a threshold, number of client computing devices are increased).
Lee and Kim in view of Vandikas in further view of Gao disclose a selection of multiple learning devices. Larsson teaches the possibility to select a different group of client computing devices in order to improve the device accuracy. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify machine learning group of Lee and Kim in view of Vandikas in further view of Gao with the learning groups taught by Larsson. The motivation would be to allow “the number of client computing devices to be adjusted to a suitable number, i.e., the minimum number of client computing devices required to achieve adequate performance” (Larsson (¶0045 line 3-7)).
As Claim 10, besides Claim 9, Lee, Kim and Vandikas in view of Gao in further view of Larsson teaches wherein the respective indications are respective first indications, and wherein detecting that the inference accuracy by the global machine learning model is below the defined inference accuracy specified by the performance criterion comprises:
receiving respective second indications of inference accuracies from respective second RICs of the second RICs (Larsson (¶0048 line 4-9, fig. 2 item 206), when the evaluation metric crosses a threshold, number of client computing devices are increased).
As Claim 11, besides Claim 10, Lee, Kim and Vandikas in view of Gao in further view of Larsson teaches wherein the respective inference accuracies identify a root mean square error associated with operating the global machine learning model (Larsson (¶0044 line 5-10), mean squared error).
As Claim 12, besides Claim 9, Lee, Kim and Vandikas in view of Gao in further view of Larsson teaches wherein selecting the third group of RICs of RICs with which to perform the federated learning for the update of the global machine learning model comprises:
performing, by the system, a metadata similarity between a first RIC that is outside of the second group of RICs and a second RIC of the second group of RICs (Kim (¶0145 line 3-9, fig. 7), training devices reports their computation resource availability to the FL server. The FL server makes the training device selection based on the reports), wherein the first RIC indicates an inference accuracy that is less than the defined inference accuracy specified by the performance criterion (Larsson (¶0048 line 4-9, fig. 2 item 206), when the evaluation metric crosses a threshold, number of client computing devices are increased).
As Claim 13, besides Claim 9, Lee, Kim and Vandikas in view of Gao in further view of Larsson teaches further comprising:
modifying, by the system, a value of the dissimilarity criterion for selection of the third group of RICs of the RICs with which to perform the second federated learning for the update of the global machine learning model (Lee (¶0229, fig. 26 step S2602), server groups first UEs corresponding to some of the plurality of UEs into a first UE group based on first location information).
As Claim 14, besides Claim 9, Lee, Kim and Vandikas in view of Gao in further view of Larsson teaches further comprising: determining, by the system, a fourth group of RICs that satisfy the performance capability criterion, wherein selecting the third group of RICs is based on the fourth group of RICs (Larsson (¶0048 line 4-9, fig. 2 item 206), when the evaluation metric crosses a threshold, number of client computing devices are increased).
As Claim 15, besides Claim 9, Lee, Kim and Vandikas in view of Gao in further view of Larsson teaches further comprising:
in response to determining that the third group of RICs matches the second group of RICs, adjusting, by the system, a value of the dissimilarity criterion to produce an adjusted dissimilarity criterion (Lee (¶0229, fig. 26 step S2602), server groups first UEs corresponding to some of the plurality of UEs into a first UE group based on first location information); and
selecting, by the system and based on the adjusted dissimilarity criterion, a fourth group of RICs of the RICs with which to perform the second federated learning for the update of the global machine learning model (Lee (¶0230, fig. 26 step S2603), server groups second UEs, that correspond to some of the plurality of UEs and do not belong to the first UE group into a second UE group based on second location information).
Claim(s) 19-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lee, Kim and Vandikas in view of Gao in further view of Shanmugaraju et al. (U.S. 2020/0187021 hereinafter Shanmugaraju).
As Claim 19, besides Claim 16, Lee and Kim in view of Vandikas in further view of Gao does not explicitly disclose:
wherein an output of the machine learning model comprises an indication of mean user equipment throughput.
Larsson teaches:
wherein an output of the machine learning model comprises an indication of mean user equipment throughput (Shanmugaraju (¶0077), system predicts average throughput put of selected devices).
Lee and Kim in view of Vandikas in further view of Gao disclose a prediction. Shanmugaraju teaches the output as mean user equipment throughput. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine prediction output Lee and Kim in view of Vandikas in further view of Gao with the mean equipment output taught by Shanmugaraju. The motivation would be to allow the system to “dynamically select QSVs for the wireless devices” (Shanmugaraju (¶0068 line 4-6)).
As Claim 20, besides Claim 16, Lee, Kim and Vandikas in view of Gao in further view of Shanmugaraju teaches wherein the respective metadata of respective agents of the agents comprises statistics about respective datasets of the respective agents, the statistics comprising at least one of a throughput, a retainability, or an accessibility (Shanmugaraju (¶0077), system predicts average throughput put of selected devices).
Claim(s) 21 is/are rejected under 35 U.S.C. 103 as being unpatentable over Lee, Kim and Vandikas in view of Gao in further view of Wang et al. (U.S. 2020/0187021 hereinafter Shanmugaraju).
As Claim 21, besides Claim 8, Lee and Kim in view of Vandikas in further view of Gao may not explicitly disclose:
wherein the dissimilarity criterion is based on respective Euclidian distances between the respective first datasets the respective selected datasets. wherein a first Euclidian distance of the Euclidian distances satisfies the dissimilarity criterion, wherein a second Euclidian distance of the Euclidian distances fails to satisfy the dissimilarity criterion, and wherein the first Euclidian distance is greater than the second Euclidian distance.
Wang teaches:
wherein the dissimilarity criterion is based on respective Euclidian distances between the respective first datasets the respective selected datasets. wherein a first Euclidian distance of the Euclidian distances satisfies the dissimilarity criterion, wherein a second Euclidian distance of the Euclidian distances fails to satisfy the dissimilarity criterion, and wherein the first Euclidian distance is greater than the second Euclidian distance (Wang (¶0084 line 4-6), “selecting the subset of VEs based on common hardware capabilities, commensurate signal and/or link quality parameters (e.g., within a threshold value or range to one another)”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine prediction output Lee and Kim in view of Vandikas in further view of Gao with the distance condition taught by Wang. The motivation would be to allow the system to “identify one or more baseline ML configurations for the subset of VEs, such as by analyzing a neural network table based on the VE parameters and/or characteristics” (Wang (¶0089 line 12-14)).
Response to Arguments
I. Rejection of Claims 1-15 under 35 U.S.C. §112(a)
Applicants amended the Claims; therefore, 35 U.S.C. §112(a) rejections are respectfully withdrawn.
II. Rejection of Claims 1-8 and 16-18 under 35 U.S.C. §103:
Applicant argues that Lee does not disclose current amended limitation in Claim 1 (last paragraph of page 13 in the remarks).
PNG
media_image1.png
91
636
media_image1.png
Greyscale
Applicants’ arguments are not persuasive because new reference Gao teaches resource blocks or modulation and coding schemes. However, Claims 1-6 incorporate additional limitation(s) and are currently allowable.
All independent/dependent Claims are not allowable for the same reasons above.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to NHAT HUY T NGUYEN whose telephone number is (571)270-7333. The examiner can normally be reached M-F: 12:00-8:00 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, Viker Lamardo can be reached at 571-270-5871. 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.
/NHAT HUY T NGUYEN/Primary Examiner, Art Unit 2147
Prosecution Insights