CTNF 18/525,577 CTNF 84960 Notice of Pre-AIA or AIA Status 07-03-aia AIA 15-10-aia The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA. Claim Rejections - 35 USC § 102 07-07-aia AIA 07-07 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 – 07-08-aia AIA (a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention. 07-15-aia AIA Claim(s) 1-8, 13-20, and 25-27 is/are rejected under 35 U.S.C. 102 (a)(1) as being anticipated by Li, WO 2022/235525 Al . Claim Element Reference: Li Analysis 1. A method of wireless communication by a user equipment (UE), comprising: collaborating with a network device Abstract T his disclosure describes systems, methods, and devices related to collaboration between user equipment (UE) and network fo r machine learning in accordance with a selected network-UE collaboration level of a plurality of network-UE collaboration levels p. 9 lines 1-23; p. 13 line 1-p. 14 line 8 Examiner finds “serviceType” teaches levels of collaboration. One level is “default” service type another level is non-default service types. for machine learning operations, p. 5 lines 12-14 The present disclosure considers categories of UE capability: (1) Hardware capability; ( 2) Machine learning capability. at least one of the plurality of network-UE collaboration levels comprising a plurality of sub-categories. p. 9 lines 1-23 Examiner finds “federated learning” is a subcategory of service type and different use cases are subcategories of federated learning, for example. 2. The method of claim 1, further comprising: transmitting an indication of a level of support for the selected network-UE collaboration level, p. 9 lines 1-23 In one or more embodiments, the machine learning configuration may be sent via RRC signaling from the NG-RAN to the UE p. 13 lines 1-p. 14 line 9 Examiner finds serviceType is an indication of the selected network-UE collaboration level. the plurality of sub-categories corresponding to a life cycle management type p. 4 lines 23-39: After action is taken, UE may also be required to send the configuration update via RRC signaling. One example use case is beam management. Assuming distributed learning is 25 adopted at UE side, UE will predict the best receiving beam index or beamforming matrix according to the inference result. UE may also receive a policy guidance from the network. The guidance policy may include a tolerant range of how UE can tune their receiving beams. If the predicted result of receiving beam matrix is within the tolerant range, UE can thus tune its receiving beams. LCM type broadly includes use cases such as “positioning beam forming” when read in light of the specification. As such, Examiner finds use case of beam management teaches a life cycle management type. and a model transfer/delivery format; p. 5 lines 12-23 . . . 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 20 capability, e.g. supported SW library, etc. A Machine learning capability field is used to indicate NG-RAN the machine learning capability of UE .. . . p. 26 lines 29-p. 27 line 2; p. 11 lines 6-31; Examiner finds model delivery format includes type of model (e.g., CNN, RNN, etc.) and size, for example that the UE can receive. Examiner finds “model may be deployed to a UE” teaches model transfer capability. Capability also includes ML algorithms, loss function, hidden layer, weight, and model size, for example. and receiving a machine learning configuration based on the level of support. Fig. 1 step 114 p. 5 lines 7-14 I n one or more embodiments, before setting machine learning configuration, U E is required to send an indication to NG-RAN to indicate whether NG-RAN can support machine learning or not. A new field "ue-capabilityML-RequestFilter" in a UECapabilityEnquiry 10 message is introduced to request UE radio access capability of support machine learning. Moreover, 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. p. 9 lines 1-6: I n one or more embodiments, the machine learning configuration may be sent via RRC signaling from the NG-RAN to the UE. To support distributed/federated learning between the NG-RAN and the UE, the following content should be included in the RRC messages: A new message type "MachineLearningConfiguration" can be used to carry machine learning 5 configuration. The information from RAN to UE includes machine learning model, configuration, (and model parameter updates,) et c. 3. The method of claim 2, in which the level of support indicates at least one machine learning functionality, p. 5 lines 12-23; p. 8 lines 24-30 The present disclosure considers categories of UE capability: (1) Hardware capability; (2) Machine learning capability. For hardware capability, in general, it is used to indicate whether the hardware of 15 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. 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 20 capability, e.g. supported SW library, etc. A Machine learning capability field is used to indicate NG-RAN the machine learning capability of UE . The UECapabilityEnquiry message is used to request UE radio access capabilities for NR as well as for other RA Ts. Signaling radio bearer: SRB 1. RLC-SAP: AM. Logical channel: DCCH. Direction: Network to UE. Examiner finds learning capability, max model size, training, inference capability teaches ML functionality. and the method further comprises performing an inference with at least one machine learning model associated with the at least one machine learning model functionality. p. 4 lines 16-32 In one or more embodiments, three types of UE behavior should be considered under RAN-UE joint machine learning optimization: (1) RRC configuration per RAN configuration. In this mode, UE behavior is the same as the legacy mechanism. The difference is how NGRAN decides the resource allocation and configuration for a certain UE. (2) Policy guided. Per policy guidance, UE decide by itself according to the policy. In this mode, according to the outcome (inference output) from the machine learning model, UE can decide its action based on the policy (Policy Guidance configured via RRC configuration) configured by the network. After action is taken, UE may also be required to send the configuration update via RRC signaling. One example use case is beam management. Assuming distributed learning is adopted at UE side, UE will predict the best receiving beam index or beamforming matrix according to the inference result . UE may also receive a policy guidance from the network. The guidance policy may include a tolerant range of how UE can tune their receiving beams. If the predicted result of receiving beam matrix is within the tolerant range, UE can thus tune its receiving beams. (3) Action guided. UE sends action space to RAN, and perform action according to NG-RAN's feedback. In this mode, UE can either get the direct action from the machine learning model ( action in reinforcement learning) or UE decide to take certain action according to the prediction results. p. 5 lines 12-23 The present disclosure considers categories of UE capability: (1) Hardware capability; (2) Machine learning capability. For hardware capability, in general, it is used to indicate whether the hardware of 15 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. 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 20 capability, e.g. supported SW library, etc. A Machine learning capability field is used to indicate NG-RAN the machine learning capability of UE. The UECapabilityEnquiry message is used to request UE radio access capabilities for NR as well as for other RA Ts. Signaling radio bearer: SRB 1. RLC-SAP: AM. Logical channel: DCCH. Direction: Network to UE. Examiner finds learning capability, max model size, training, inference capability teaches ML functionality. 4. The method of claim 3, further comprising receiving the at least one machine learning model from a non-third generation partnership project (3GPP) entity. p. 25 lines 29-32 The network 300 may operate in a manner consistent with 3GPP technical 30 specifications for LTE or 5G/NR systems. However, the example embodiments are not limited in this regard and the described embodiments may apply to other networks that benefit from the principles described herein, such as future 3GPP systems, or the like p. 9 line 1-6 In one or more embodiments, the machine learning configuration may be sent via RRC signaling from the NG-RAN to the UE. To support distributed/federated learning between the NG-RAN and the UE, the following content should be included in the RRC messages: A new message type "MachineLearningConfiguration" can be used to carry machine learning 5 configuration. The information from RAN to UE includes machine learning model, configuration, (and model parameter updates,) etc. Examiner finds “not limited in this regard” teaches non-third party 3PP entities. 5. The method of claim 3, further comprising receiving, from the network device, instructions comprising at least one of: instructions for functionality activation, instructions for functionality deactivation, or instructions for performance monitoring. p. 29 lines 13-15 (activation/deactivation handovers); p. 3 lines 27-p. 4 line 6 (security activity ); p. 5 lines 14-23; p. 11 lines 13-16 Examiner handovers and security teaches “functionality activations” ML functionality also includes max model size, what type of ML model it can support, training capability, etc. 6. The method of claim 2, in which the level of support indicates at least one machine learning functionality p. 5 lines 14-23; p. 11 lines 13-16 In one or more embodiments, the Machine Learning Model field can also include other training related information, such as loss function, optimizer for training. The Machin e 15 Learning Model field is optional for distributed learning framework, mandatory for federated framework or model deployment from NG-RAN to UE. If the model size is larger than 8000bytes, RRC Segmentation is supported. ML functionality includes max model size, what type of ML model it can support, training capability, etc. and at least one machine learning model identification (ID) p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”); for each ML functionality, and the method further comprises receiving instructions for performing an inference with at least one machine learning model associated with the at least one machine learning model ID. p. 9 lines 12-18 Another scenario of distributed learning (scenario 2) is that NG-RAN serves as the machine learning training node and the UE as the inference node. UE can download the model 15 trained by RAN according to its interested machine learning service type. Up o n receiving such model, UE can start inference based on the input data. "service type" is used to indicate the service where the machine learning model is trained from. Examples include positioning, V2X, channel estimation, etc. p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”); p. 11 lines 16-20: 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 20 capability , e.g. supported SW library, etc. 7. The method of claim 6, further comprising receiving the at least one machine learning model from a non-third generation partnership project (3GPP) entity. p. 25 lines 29-32 The network 300 may operate in a manner consistent with 3GPP technical 30 specifications for LTE or 5G/NR systems. However, the example embodiments are not limited in this regard a nd the described embodiments may apply to other networks that benefit from the principles described herein, such as future 3GPP systems, or the like p. 9 line 1-6 In one or more embodiments, the machine learning configuration may be sent via RRC signaling from the NG-RAN to the UE. To support distributed/federated learning between the NG-RAN and the UE, the following content should be included in the RRC messages: A new message type "MachineLearningConfiguration" can be used to carry machine learning 5 configuration. The information from RAN to UE includes machine learning model, configuration, (and model parameter updates,) etc. 8. The method of claim 6, in which the instructions comprise at least one of: instructions for model selection, p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”) Identifying a model is selecting it. instructions for model activation, instructions for model deactivation p. 29 lines 13-15 Examiner finds that if the UE is deactivated, then the model is deactivated. instructions for model switching, instructions for model fallback, p. 5 lines 1-6 Examiner finds handover teaches switching/fallback. or instructions for model monitoring. p. 10 lines 9-22 Examiner finds periodicity of reporting model’s results teaches model monitoring. 13. The method of claim 2, in which the level of support indicates at least one machine learning functionality, p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity ”) Examiner finds models with smaller granularity teaches a level of support for ML model, for example. at least one machine learning model identification (ID) for each ML functionality, p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity ”) and a model transfer capability of the UE p. 5 lines 16-20. p. 40 lines 6-10; including a formula-based machine learning inference capability of the UE. p. 5 lines 16-20 and p. 40 lines 6-10 Formula based ML inference capability of the UE includes max model size when read in light of the specification. 14. A method of wireless communication by a network device, comprising: collaborating with a user equipment (UE) Abstract This disclosure describes systems, methods, and devices related to collaboration between user equipment (UE) and network fo r machine learning in accordance with a selected network-UE collaboration level of a plurality of network-UE collaboration levels p. 9 lines 1-23; p. 13 line 1-p. 14 line 8 Examiner finds “serviceType” teaches levels of collaboration. One level is “default” service type another level is non-default service types. for machine learning operations, p 5 lines 12-14 The present disclosure considers categories of UE capability: (1) Hardware capability; ( 2) Machine learning capability. at least one of the plurality of network-UE collaboration levels comprising a plurality of sub-categories. p. 9 lines 1-23 Examiner finds “federated learning” is a subcategory of service type and different use cases are subcategories of federated learning, for example. 15. The method of claim 14, further comprising: receiving an indication of a level of support for the selected network-UE collaboration level, p. 9 lines 1-23 In one or more embodiments, the machine learning configuration may be sent via RRC signaling from the NG-RAN to the UE p. 13 lines 1-p. 14 line 9 Examiner finds serviceType is an indication of the selected network-UE collaboration level. the plurality of sub-categories corresponding to a life cycle management type p. 4 lines 23-39: After action is taken, UE may also be required to send the configuration update via RRC signaling. One example use case is beam management. Assuming distributed learning is 25 adopted at UE side, UE will predict the best receiving beam index or beamforming matrix according to the inference result. UE may also receive a policy guidance from the network. The guidance policy may include a tolerant range of how UE can tune their receiving beams. If the predicted result of receiving beam matrix is within the tolerant range, UE can thus tune its receiving beams. LCM type broadly includes use cases such as “positioning beam forming” when read in light of the specification. As such, Examiner finds use case of beam management teaches a life cycle management type. and a model transfer/delivery format; p. 5 lines 12-23 . . . 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 20 capability, e.g. supported SW library, etc. A Machine learning capability field is used to indicate NG-RAN the machine learning capability of UE .. . . p. 26 lines 29-p. 27 line 2; p. 11 lines 6-31; Examiner finds model delivery format includes type of model (e.g., CNN, RNN, etc.) and size, for example that the UE can receive. Examiner finds “model may be deployed to a UE” teaches model transfer capability. Capability also includes ML algorithms, loss function, hidden layer, weight, and model size, for example. and transmitting a machine learning configuration based on the level of support. Fig. 1 step 114 p. 5 lines 7-14 I n one or more embodiments, before setting machine learning configuration, U E is required to send an indication to NG-RAN to indicate whether NG-RAN can support machine learning or not. A new field "ue-capabilityML-RequestFilter" in a UECapabilityEnquiry 10 message is introduced to request UE radio access capability of support machine learning. Moreover, 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. p. 9 lines 1-6: I n one or more embodiments, the machine learning configuration may be sent via RRC signaling from the NG-RAN to the UE. To support distributed/federated learning between the NG-RAN and the UE, the following content should be included in the RRC messages: A new message type "MachineLearningConfiguration" can be used to carry machine learning 5 configuration. The information from RAN to UE includes machine learning model, configuration, (and model parameter updates,) et c. 16. The method of claim 15, in which the level of support indicates at least one machine learning functionality, p. 5 lines 12-23; p. 8 lines 24-30 The present disclosure considers categories of UE capability: (1) Hardware capability; (2) Machine learning capability. For hardware capability, in general, it is used to indicate whether the hardware of 15 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. 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 20 capability, e.g. supported SW library, etc. A Machine learning capability field is used to indicate NG-RAN the machine learning capability of UE . The UECapabilityEnquiry message is used to request UE radio access capabilities for NR as well as for other RA Ts. Signaling radio bearer: SRB 1. RLC-SAP: AM. Logical channel: DCCH. Direction: Network to UE. Examiner finds learning capability, max model size, training, inference capability teaches ML functionality. and the method further comprises performing an inference with at least one machine learning model associated with the at least one machine learning model functionality. p. 4 lines 16-32 In one or more embodiments, three types of UE behavior should be considered under RAN-UE joint machine learning optimization: (1) RRC configuration per RAN configuration. In this mode, UE behavior is the same as the legacy mechanism. The difference is how NGRAN decides the resource allocation and configuration for a certain UE. (2) Policy guided. Per policy guidance, UE decide by itself according to the policy. In this mode, according to the outcome (inference output) from the machine learning model, UE can decide its action based on the policy (Policy Guidance configured via RRC configuration) configured by the network. After action is taken, UE may also be required to send the configuration update via RRC signaling. One example use case is beam management. Assuming distributed learning is adopted at UE side, UE will predict the best receiving beam index or beamforming matrix according to the inference result . UE may also receive a policy guidance from the network. The guidance policy may include a tolerant range of how UE can tune their receiving beams. If the predicted result of receiving beam matrix is within the tolerant range, UE can thus tune its receiving beams. (3) Action guided. UE sends action space to RAN, and perform action according to NG-RAN's feedback. In this mode, UE can either get the direct action from the machine learning model ( action in reinforcement learning) or UE decide to take certain action according to the prediction results. p. 5 lines 12-23 The present disclosure considers categories of UE capability: (1) Hardware capability; (2) Machine learning capability. For hardware capability, in general, it is used to indicate whether the hardware of 15 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. 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. The UECapabilityEnquiry message is used to request UE radio access capabilities for NR as well as for other RA Ts. Signaling radio bearer: SRB 1. RLC-SAP: AM. Logical channel: DCCH. Direction: Network to UE. Examiner finds learning capability, max model size, training, inference capability teaches ML functionality. 17. The method of claim 14, further comprising transmitting at least one machine learning model for the machine learning operations. p. 9 line 1-6 In one or more embodiments, the machine learning configuration may be sent via RRC signaling from the NG-RAN to the UE. To support distributed/federated learning between the NG-RAN and the UE, the following content should be included in the RRC messages: A new message type "MachineLearningConfiguration" can be used to carry machine learning 5 configuration. The information from RAN to UE includes machine learning model, configuration, (and model parameter updates,) etc. 18. The method of claim 16, further comprising transmitting instructions comprising at least one of: instructions for functionality activation, instructions for functionality deactivation, p. 29 lines 13-15 (activation/deactivation) or instructions for performance monitoring. p. 10 lines 9-22 Examiner finds periodicity of reporting model’s results teaches performance monitoring. 19. The method of claim 15, in which the level of support indicates at least one machine learning functionality and at least one machine learning model identification (ID) for each ML functionality, p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity ”) Examiner finds models with smaller granularity teaches a level of support for ML model, for example. and the method further comprises transmitting instructions for performing an inference with at least one machine learning model associated with the at least one machine learning model ID. p. 9 lines 12-18 Another scenario of distributed learning (scenario 2) is that NG-RAN serves as the machine learning training node and the UE as the inference node. UE can download the model 15 trained by RAN according to its interested machine learning service type. Up o n receiving such model, UE can start inference based on the input data. "service type" is used to indicate the service where the machine learning model is trained from. Examples include positioning, V2X, channel estimation, etc. p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”); p. 11 lines 16-20: 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 20 capability , e.g. supported SW library, etc. 20. The method of claim 19, in which the instructions comprise at least one of: instructions for model selection, p. 29 lines 13-15 (activation/deactivation handovers); p. 5 lines 14-23; p. 11 lines 13-16 Examiner handovers and security teaches “functionality activations” ML functionality also includes max model size, what type of ML model it can support, training capability, etc. instructions for model activation, instructions for model deactivation, p. 29 lines 13-15 (activation/deactivation handovers); instructions for model switching, p. 5 lines 1-6 Examiner finds handover teaches switching/fallback. instructions for model fallback, p. 5 lines 1-6 Examiner finds handover teaches switching/fallback. or instructions for model monitoring. p. 10 lines 9-22 Examiner finds periodicity of reporting model’s results teaches model monitoring. 25. The method of claim 15, in which the level of support indicates at least one machine learning functionality, at least one machine learning model identification (ID) for each ML functionality, p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity ”) Examiner finds models with smaller granularity teaches a level of support for ML model, for example. and a model transfer capability of the UE including a formula-based machine learning inference capability of the UE. p. 5 lines 16-20 and p. 40 lines 6-10 Formula based ML inference capability of the UE includes max model size when read in light of the specification. 26. An apparatus for wireless communication by a user equipment (UE), comprising: at least one memory; and at least one processor coupled to the at least one memory, the at least one processor configured to collaborate with a network device in accordance with a selected network-UE collaboration level of a plurality of network-UE collaboration levels p. 9 lines 1-23; p. 13 line 1-p. 14 line 8 Abstract This disclosure describes systems, methods, and devices related to collaboration between user equipment (UE) and network fo r machine learning Examiner finds “serviceType” teaches levels of collaboration. One level is “default” service type another level is non-default service types. for machine learning operations, p. 5 lines 12-14 The present disclosure considers categories of UE capability: (1) Hardware capability; ( 2) Machine learning capability. at least one of the plurality of network-UE collaboration levels comprising a plurality of sub-categories. p. 9 lines 1-23 Examiner finds “federated learning” is a subcategory of service type and different use cases are subcategories of federated learning, for example. 27. An apparatus for wireless communication by a network device, comprising: at least one memory; and at least one processor coupled to the at least one memory, the at least one processor configured to collaborate with a user equipment (UE) in accordance with a selected network-UE collaboration level of a plurality of network-UE collaboration levels for machine learning operations, Abstract This disclosure describes systems, methods, and devices related to collaboration between user equipment (UE) and network fo r machine learning p. 5 lines 12-14 The present disclosure considers categories of UE capability: (1) Hardware capability; ( 2) Machine learning capability. Examiner finds “serviceType” teaches levels of collaboration. One level is “default” service type another level is non-default service types. at least one of the plurality of network-UE collaboration levels comprising a plurality of sub-categories. p. 9 lines 1-23 Examiner finds “federated learning” is a subcategory of service type and different use cases are subcategories of federated learning, for example . Claim Rejections - 35 USC § 103 07-20-aia AIA 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 9-10 and 21-22 is/are rejected under 35 U.S.C. 103 as being obvious over Li in view of claim 2 above and further in view of Da Silva US 20240040461 A1. With respect to claim 9 , Li teaches the following: 9. The method of claim 2, in which: the level of support indicates at least one machine learning functionality, p. 5 lines 14-23; p. 11 lines 13-16 ML functionality includes max model size, what type of ML model it can support, training capability, etc. at least one machine learning model identification (ID) for each ML functionality, p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”) and a model transfer capability of the UE for reception p. 9 lines 19-23; p. 26 lines 29-p. 27 line 2; p. 11 lines 6-31; Examiner finds “model may be deployed to a UE” teaches model transfer capability. Capability also includes ML algorithms, loss function, hidden layer, weight, and model size, for example. from a third generation partnership project (3GPP) entity, p. 2 lines 3-7; p. 25 lines 29-32 a machine learning model being stored in a proprietary format, and the method further comprises receiving instructions for performing an inference with at least one machine learning model associated with the at least one machine learning model ID. p. 5 lines 18-20 . . . and for each machine learning model, it includes: maximum model size, training capability, e.g. supported SW library, inference 20 capability, p. 9 lines 4-18; p. 11 lines 10-12 (model ID) It appears Li fails to explicitly teach “a machine learning model being stored in a proprietary format.” However, Da Silva teaches “a machine learning model being stored in a proprietary format” in para. 232: [0232] The UE 403 may have stored the prediction model, e.g. UE proprietary prediction model, to be used when perform the prediction. Da Silva and Li are analogous art because they are from the same field of endeavor as the claimed invention. It would have been obvious to one skilled in the art before the effective filing date of the invention to modify “the method of claim 2, in which: the level of support indicates at least one machine learning functionality, at least one machine learning model identification (ID) for each ML functionality, and a model transfer capability of the UE for reception from a third generation partnership project (3GPP) entity” as taught by Li to include “a machine learning model being stored in a proprietary format” as taught by Da Silva. The motivation would have been to improve radio link robustness. See Da Silva para. 118. With respect to claim 10 , Li teaches the following: 10. The method of claim 9, in which the instructions comprise at least one of: instructions for model selection, p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”) Identifying a model is selecting it. instructions for model activation, instructions for model deactivation, p. 29 lines 13-15 (activation/deactivation handovers); instructions for model switching, instructions for model fallback, p. 5 lines 1-6 Examiner finds handover teaches switching/fallback. or instructions for model monitoring. p. 10 lines 9-22 Examiner finds periodicity of reporting model’s results teaches model monitoring. With respect to claim 21 , Li teaches the following: 21. The method of claim 15, in which: the level of support indicates at least one machine learning functionality, p. 5 lines 14-23; p. 11 lines 13-16 ML functionality includes max model size, what type of ML model it can support, training capability, etc. at least one machine learning model identification (ID) for each ML functionality, and a model transfer capability of the UE for reception from a third generation partnership project (3GPP) entity, p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”) p. 25 lines 29-32 The network 300 may operate in a manner consistent with 3GPP technical specifications for LTE or 5G/NR systems. However, the example embodiments are not limited in this regard and the described embodiments may apply to other networks that benefit from the principles described herein, such as future 3GPP systems, or the like a machine learning model being stored in a proprietary format, and the method further comprises transmitting instructions for performing an inference with at least one machine learning model associated with the at least one machine learning model ID. p. 9 lines 12-18 Another scenario of distributed learning (scenario 2) is that NG-RAN serves as the machine learning training node and the UE as the inference node. UE can download the model 15 trained by RAN according to its interested machine learning service type. Up o n receiving such model, UE can start inference based on the input data. "service type" is used to indicate the service where the machine learning model is trained from. Examples include positioning, V2X, channel estimation, etc. p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”); p. 11 lines 16-20: 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 20 capability , e.g. supported SW library, etc. It appears Li fails to explicitly teach “a machine learning model being stored in a proprietary format.” However, Da Silva teaches “the machine learning model being stored in a proprietary format” in para. 232: [0232] The UE 403 may have stored the prediction model, e.g. UE proprietary prediction model, to be used when perform the prediction. Da Silva and Li are analogous art because they are from the same field of endeavor as the claimed invention. It would have been obvious to one skilled in the art before the effective filing date of the invention to modify “21. The method of claim 15, in which: the level of support indicates at least one machine learning functionality, at least one machine learning model identification (ID) for each ML functionality, and a model transfer capability of the UE for reception from a third generation partnership project (3GPP) entity” as taught by Li to include “a machine learning model being stored in a proprietary format” as taught by Da Silva. The motivation would have been to improve radio link robustness. See Da Silva para. 118. With respect to claim 22, Li teaches the following: Claim Reference: Li Analysis 22. The method of claim 21, in which the instructions comprise at least one of: instructions for model selection, p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”) Identifying a model is selecting it. instructions for model activation, instructions for model deactivation, p. 29 lines 13-15 Examiner finds that if the UE is deactivated, then the model is deactivated. instructions for model switching, instructions for model fallback, p. 5 lines 1-6 Examiner finds handover teaches switching/fallback. or instructions for model monitoring. p. 10 lines 9-22 Examiner finds periodicity of reporting model’s results teaches model monitoring. Claim 11-12 and 23-24 is/are rejected under 35 U.S.C. 103 as being obvious over Li in view of claim 2 above and further in view of Ryden WO 2023274514 A1. With respect to claim 11 , Li teaches the following: 11. The method of claim 2, in which: the level of support indicates at least one machine learning functionality, p. 5 lines 12-23; p. Examiner finds learning capability, max model size, training, inference capability teaches ML functionality. at least one machine learning model identification (ID) for each ML functionality, p. 11 lines 10-12; p. 8 lines 24-30 Examiner finds model capability teaches functionality. and a model transfer capability of the UE including whether the UE supports receiving updated machine learning parameters from a third generation partnership project (3GPP) entity, p. 9 lines 1-6 In one or more embodiments, the machine learning configuration may be sent via RRC signaling from the NG-RAN to the UE. To support distributed/federated learning between the NG-RAN and the UE, the following content should be included in the RRC messages: A new message type " MachineLearningConfiguration" can be used to carry machine learning configuration. The information from RAN to UE includes machine learning model, configuration, (and model parameter updates,) etc. p. 9 lines 13-18 Another scenario of distributed learning (scenario 2) is that NG-RAN serves as the machine learning training node and the UE as the inference node. UE can download the model 15 trained by RAN according to its interested machine learning service type. Upon receiving such model, UE can start inference based on the input data. "service type" is used to indicate the service where the machine learning model is trained from. Examples include positioning, V2X, channel estimation, etc. a machine learning model being stored in an open format, and the method further comprises receiving instructions for performing an inference with at least one machine learning model associated with the at least one machine learning model ID. p. 9 lines 13-18 Another scenario of distributed learning (scenario 2) is that NG-RAN serves as the machine learning training node and the UE as the inference node. UE can download the model 15 trained by RAN according to its interested machine learning service type. Upon receiving such model, UE can start inference based on the input data . "service type" is used to indicate the service where the machine learning model is trained from. Examples include positioning, V2X, channel estimation, etc. p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”) It appears Li fails to explicitly teach “a machine learning model being stored in an open format/” However Ryden WO 2023274514 A1 teaches “a machine learning model being stored in an open format” on p. 24 lines 31-33: In some examples, the DAE forming part of the prediction information may be signaled to the UE using existing model formats, such as, the Open Neural Network Exchange (ONNX),1 or formats commonly used in ML and Al solutions such as Keras or Pytorch. And Li are analogous art because they are from the same field of endeavor as the claimed invention. It would have been obvious to one skilled in the art before the effective filing date of the invention to modify “the level of support indicates at least one machine learning functionality, at least one machine learning model identification (ID) for each ML functionality, and a model transfer capability of the UE including whether the UE supports receiving updated machine learning parameters from a third generation partnership project (3GPP) entity” in Li to include “a machine learning model being stored in an open format” as taught by Ryden. The motivation would have been portability across platforms. With respect to claim 12, Li teaches the following: Claim Reference: Li Analysis 12. The method of claim 11, in which the instructions comprise at least one of: instructions for model selection instructions for model selection, p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”) Identifying a model is selecting it. instructions for model activation, instructions for model deactivation, p. 29 lines 13-15 (activation/deactivation handovers); instructions for model switching, instructions for model fallback, p. 5 lines 1-6 Examiner finds handover teaches switching/fallback. or instructions for model monitoring. p. 10 lines 9-22 Examiner finds periodicity of reporting model’s results teaches model monitoring. With respect to claim 23 , Li teaches the following: Claim Reference: Li Analysis 23. The method of claim 15, in which: the level of support indicates at least one machine learning functionality, p. 5 lines 12-23; p. Examiner finds learning capability, max model size, training, inference capability teaches ML functionality. at least one machine learning model identification (ID) for each ML functionality, p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”) and a model transfer capability of the UE including whether the UE supports receiving updated machine learning parameters from a third generation partnership project (3GPP) entity, p. 10 lines 1-5 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. p. 9 lines 1-6 In one or more embodiments, the machine learning configuration may be sent via RRC signaling from the NG-RAN to the UE. To support distributed/federated learning between the NG-RAN and the UE, the following content should be included in the RRC messages: A new message type " MachineLearningConfiguration" can be used to carry machine learning configuration. The information from RAN to UE includes machine learning model, configuration, (and model parameter updates,) etc. p. 9 lines 13-18 Another scenario of distributed learning (scenario 2) is that NG-RAN serves as the machine learning training node and the UE as the inference node. UE can download the model 15 trained by RAN according to its interested machine learning service type. Upon receiving such model, UE can start inference based on the input data. "service type" is used to indicate the service where the machine learning model is trained from. Examples include positioning, V2X, channel estimation, etc. p. 25 lines 29-32 The network 300 may operate in a manner consistent with 3GPP technical specifications for LTE or 5G/NR systems. However, the example embodiments are not limited in this regard and the described embodiments may apply to other networks that benefit from the principles described herein, such as future 3GPP systems, or the like a machine learning model being stored in an open format, and the method further comprises transmitting instructions for performing an inference with at least one machine learning model associated with the at least one machine learning model ID. p. 9 lines 12-18 Another scenario of distributed learning (scenario 2) is that NG-RAN serves as the machine learning training node and the UE as the inference node. UE can download the model 15 trained by RAN according to its interested machine learning service type. Up o n receiving such model, UE can start inference based on the input data. "service type" is used to indicate the service where the machine learning model is trained from. Examples include positioning, V2X, channel estimation, etc. p. 11 lines 10-12 (“The ML models can also be in the form of identifiers for UE to download the models. According to different UE capability, NG-RAN may assign models with smaller granularity”); p. 11 lines 16-20: 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 20 capability , e.g. supported SW library, etc. It appears Li fails to explicitly teach “a machine learning model being stored in an open format” However Ryden WO 2023274514 A1 teaches “a machine learning model being stored in an open format” on p. 24 lines 31-33: In some examples, the DAE forming part of the prediction information may be signaled to the UE using existing model formats, such as, the Open Neural Network Exchange (ONNX),1 or formats commonly used in ML and Al solutions such as Keras or Pytorch. Ryden and Li are analogous art because they are from the same field of endeavor as the claimed invention. It would have been obvious to one skilled in the art before the effective filing date of the invention to modify “the level of support indicates at least one machine learning functionality, at least one machine learning model identification (ID) for each ML functionality, and a model transfer capability of the UE including whether the UE supports receiving updated machine learning parameters from a third generation partnership project (3GPP) entity” in Li to include “a machine learning model being stored in an open format” as taught by Ryden. The motivation would have been portability across platforms. Claim Reference: Li Analysis 24. The method of claim 23, in which the instructions comprise at least one of: instructions for model selection, instructions for model activation, instructions for model deactivation, p. 29 lines 13-15 (activation/deactivation); instructions for model switching, p. 5 lines 1-6 Examiner finds handover teaches switching/fallback. instructions for model fallback, p. 5 lines 1-6 Examiner finds handover teaches switching/fallback. or instructions for model monitoring. p. 10 lines 9-22 Examiner finds periodicity of reporting model’s results teaches model monitoring. Conclusion The following prior art is relevant to Applicant’s disclosure: Reference: ETSI TS 128 100 V17.0.0 Management and orchestration; Levels of autonomous network (3GPP TS 28.100 version 17.0.0 Release 17), 2022 p. 10 section 5: “ Human & System ” represents corresponding tasks are accomplished by collaboration of human and telecom system, the detailed collaboration pattern depends on the scenario, which is not addressed in the framework approach for evaluating autonomous network levels p. 12 section 6.2.1 p. 13 section 7.1.2; p. 14 section 7.1.2 and Fig. 7.12-1 p. 8 section 4.1 The network become complex due to large number of devices and diversity of services. Different autonomy mechanisms are introduced in the industry to reduce the complexity of network management and control. The ultimate goal for autonomous network is to enable telecommunication system (including management system and network) to be governed by itself with minimal to no human intervention by utilizing the autonomy mechanisms (including intelligence mechanism, e.g. AI/ML, and automation mechanism, e.g. rule-based automatic control, and other mechanisms to enable the autonomous network). p. 11 table 5-1; p. 15 Fig. 7.1.2-1 Any inquiry concerning this communication or earlier communications from the examiner should be directed to ALBERT M PHILLIPS, III whose telephone number is (571)270-3256. The examiner can normally be reached 10a-6:30pm EST M-F. 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, Ann J Lo can be reached at (571) 272-9767. 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. /ALBERT M PHILLIPS, III/Primary Examiner, Art Unit 2159 Application/Control Number: 18/525,577 Page 2 Art Unit: 2159 Application/Control Number: 18/525,577 Page 3 Art Unit: 2159 Application/Control Number: 18/525,577 Page 4 Art Unit: 2159 Application/Control Number: 18/525,577 Page 5 Art Unit: 2159 Application/Control Number: 18/525,577 Page 6 Art Unit: 2159 Application/Control Number: 18/525,577 Page 7 Art Unit: 2159 Application/Control Number: 18/525,577 Page 8 Art Unit: 2159 Application/Control Number: 18/525,577 Page 9 Art Unit: 2159 Application/Control Number: 18/525,577 Page 10 Art Unit: 2159 Application/Control Number: 18/525,577 Page 11 Art Unit: 2159 Application/Control Number: 18/525,577 Page 12 Art Unit: 2159 Application/Control Number: 18/525,577 Page 13 Art Unit: 2159 Application/Control Number: 18/525,577 Page 14 Art Unit: 2159 Application/Control Number: 18/525,577 Page 15 Art Unit: 2159 Application/Control Number: 18/525,577 Page 16 Art Unit: 2159 Application/Control Number: 18/525,577 Page 17 Art Unit: 2159 Application/Control Number: 18/525,577 Page 18 Art Unit: 2159 Application/Control Number: 18/525,577 Page 19 Art Unit: 2159 Application/Control Number: 18/525,577 Page 20 Art Unit: 2159 Application/Control Number: 18/525,577 Page 21 Art Unit: 2159 Application/Control Number: 18/525,577 Page 22 Art Unit: 2159 Application/Control Number: 18/525,577 Page 23 Art Unit: 2159 Application/Control Number: 18/525,577 Page 24 Art Unit: 2159 Application/Control Number: 18/525,577 Page 25 Art Unit: 2159 Application/Control Number: 18/525,577 Page 26 Art Unit: 2159 Application/Control Number: 18/525,577 Page 27 Art Unit: 2159 Application/Control Number: 18/525,577 Page 28 Art Unit: 2159 Application/Control Number: 18/525,577 Page 29 Art Unit: 2159 Application/Control Number: 18/525,577 Page 30 Art Unit: 2159 Application/Control Number: 18/525,577 Page 31 Art Unit: 2159 Application/Control Number: 18/525,577 Page 32 Art Unit: 2159 Application/Control Number: 18/525,577 Page 33 Art Unit: 2159 Application/Control Number: 18/525,577 Page 34 Art Unit: 2159 Application/Control Number: 18/525,577 Page 35 Art Unit: 2159 Application/Control Number: 18/525,577 Page 36 Art Unit: 2159 Application/Control Number: 18/525,577 Page 37 Art Unit: 2159 Application/Control Number: 18/525,577 Page 38 Art Unit: 2159 Application/Control Number: 18/525,577 Page 39 Art Unit: 2159 Application/Control Number: 18/525,577 Page 40 Art Unit: 2159 Application/Control Number: 18/525,577 Page 41 Art Unit: 2159 Application/Control Number: 18/525,577 Page 42 Art Unit: 2159 Application/Control Number: 18/525,577 Page 43 Art Unit: 2159 Application/Control Number: 18/525,577 Page 44 Art Unit: 2159 Application/Control Number: 18/525,577 Page 45 Art Unit: 2159 Application/Control Number: 18/525,577 Page 46 Art Unit: 2159 Application/Control Number: 18/525,577 Page 47 Art Unit: 2159 Application/Control Number: 18/525,577 Page 48 Art Unit: 2159 Application/Control Number: 18/525,577 Page 49 Art Unit: 2159 Application/Control Number: 18/525,577 Page 50 Art Unit: 2159 Application/Control Number: 18/525,577 Page 51 Art Unit: 2159 Application/Control Number: 18/525,577 Page 52 Art Unit: 2159 Application/Control Number: 18/525,577 Page 53 Art Unit: 2159 Application/Control Number: 18/525,577 Page 54 Art Unit: 2159 Application/Control Number: 18/525,577 Page 55 Art Unit: 2159 Application/Control Number: 18/525,577 Page 56 Art Unit: 2159 Application/Control Number: 18/525,577 Page 57 Art Unit: 2159 Application/Control Number: 18/525,577 Page 58 Art Unit: 2159 Application/Control Number: 18/525,577 Page 59 Art Unit: 2159 Application/Control Number: 18/525,577 Page 60 Art Unit: 2159 Application/Control Number: 18/525,577 Page 61 Art Unit: 2159 Application/Control Number: 18/525,577 Page 62 Art Unit: 2159 Application/Control Number: 18/525,577 Page 63 Art Unit: 2159 Application/Control Number: 18/525,577 Page 64 Art Unit: 2159 Application/Control Number: 18/525,577 Page 65 Art Unit: 2159 Application/Control Number: 18/525,577 Page 66 Art Unit: 2159 Application/Control Number: 18/525,577 Page 67 Art Unit: 2159 Application/Control Number: 18/525,577 Page 68 Art Unit: 2159 Application/Control Number: 18/525,577 Page 69 Art Unit: 2159 Application/Control Number: 18/525,577 Page 70 Art Unit: 2159 Application/Control Number: 18/525,577 Page 71 Art Unit: 2159 Application/Control Number: 18/525,577 Page 72 Art Unit: 2159 Application/Control Number: 18/525,577 Page 73 Art Unit: 2159 Application/Control Number: 18/525,577 Page 74 Art Unit: 2159 Application/Control Number: 18/525,577 Page 75 Art Unit: 2159
