Artificial intelligence (AI)-machine learning (ML) operational workflow
A standardized generic ML operational workflow for 5G networks addresses the challenges of ML management across different domains, enhancing efficiency and reliability by structuring ML operations into training, emulation, deployment, and inference phases, thereby improving network performance and user experience.
Patent Information
- Application Number
- PCT/US2024/028274
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-09
- Filing Date
- 2024-05-08
- Publication Date
- 2024-11-14
AI Technical Summary
Current wireless communication systems face challenges in efficiently managing machine learning (ML) workflows across various domains in 5G networks, including training, testing, emulation, deployment, and inference, which can impact network performance and user experience due to the lack of standardized and generic operational workflows.
A generic high-level ML operational workflow is introduced, comprising phases such as training, emulation, deployment, and inference, managed by ML training, emulation, loading, and inference modules, respectively, with a focus on standardized management capabilities to ensure seamless operation across different domains and functions within 5G systems.
This approach standardizes ML management across 5G networks, enhancing the efficiency and reliability of ML operations, reducing the risk of negative impacts on the network and users, and enabling more effective deployment and performance evaluation of ML entities within the 5G ecosystem.
Smart Images

Figure US2024028274_14112024_PF_FP_ABST
Abstract
Description
[0001] ARTIFICIAL INTELLIGENCE (Al) -MACHINE LEARNING (ML) OPERATIONAL WORKFLOW
[0002] CROSS REFERENCE TO RELATED APPLICATION
[0003] The present application claims priority to U.S. Provisional Patent Application No. 63 / 501,085, which was filed May 9, 2023.
[0004] BACKGROUND
[0005] Various embodiments generally may relate to the field of wireless communications.
[0006] BRIEF DESCRIPTION OF THE DRAWINGS
[0007] Embodiments will be readily understood by the following detailed description in conjunction with the accompanying drawings. To facilitate this description, like reference numerals designate like structural elements. Embodiments are illustrated by way of example and not by way of limitation in the figures of the accompanying drawings.
[0008] Figure 1 illustrates an example machine learning (ML) -related operational workflow, in accordance with various embodiments.
[0009] Figure 2 illustrates a network in accordance with various embodiments.
[0010] Figure 3 schematically illustrates a wireless network 300 in accordance with various embodiments.
[0011] Figure 4 is a block diagram illustrating components, according to some example embodiments, able to read instructions from a machine-readable or computer-readable medium (e.g., a non-transitory machine-readable storage medium) and perform any one or more of the methodologies discussed herein.
[0012] Figure 5 illustrates a network in accordance with various embodiments.
[0013] Figure 6 illustrates a simplified block diagram of (ML) -assisted communication between a UE and a RAN, in accordance with various embodiments
[0014] Figure 7 depicts an example procedure based on the ML-related operation workflow described herein.
[0015] DETAILED DESCRIPTION
[0016] The following detailed description refers to the accompanying drawings. The same reference numbers may be used in different drawings to identify the same or similar elements. In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular structures, architectures, interfaces, techniques, etc. in order to provide a thorough understanding of the various aspects of various embodiments. However, it will be apparent to those skilled in the art having the benefit of the present disclosure that the various aspects of the various embodiments may be practiced in other examples that depart from these specific details. In certain instances, descriptions of well-known devices, circuits, and methods are omitted so as not to obscure the description of the various embodiments with unnecessary detail. For the purposes of the present document, the phrases “A or B” and “A / B” mean (A), (B), or (A and B).
[0017] As used herein, the term “Al” may refer to “artificial intelligence.” The term “ML” may refer to “machine learning.” “Al / ML” may be used herein as an umbrella term for both Al and ML. Generally, the terms maybe used interchangeably for the purpose of this discussion, unless indicated otherwise. The term “5G” may refer to “fifth generation.”
[0018] ML may be used in 5G systems (5GS) (including the 5G core (5GC), next generation (NG) -radio access network (RAN) components, the 5G management system, etc.). ML management capabilities may be related to, or may operate, one or more of the following example operations: ML training, ML testing, ML emulation, ML entity deployment / loading, and / or ML inference.
[0019] The ML workflows involving or related to these ML management capabilities may be generic for different domains (e.g., the 5GC, NG-RAN(s), the 5G management system, etc.). Additionally or alternatively, the ML workflows involving or related to these ML management capabilities may be generic for different ML-related functions such as the following non-limiting examples: a distributed mobility optimization function, a distributed load balancing function, or a distributed energy saving function running on a gNB, or Analytics Logical function (AnLF) on a network data analytics function (NWDAF), a Management Data Analytics Function (MDAF), etc. In other words, the high level ML workflow described herein may be applicable to various of the domains described above, or ML-related functions that are performed within those given domains. It will be understood that the term “generic” is used herein with respect to the ML workflow for the purpose of indicating that the workflow is applicable to the various domains and / or functions. The term “generic” is not intended to imply, nor should it be interpreted as implying, that the ML workflow described herein is known in the prior art.
[0020] Description of a generic high-level ML workflow, for example as may be present in or related to one or more technical specifications (TSs) or technical reports (TRs) of a standards body such as the third generation partnership project (3GPP), may be desirable to illustrate where and what ML management / operations may be used, and how they may systematically work together to bring ML into the 5GS.
[0021] Embodiments herein relate to example ML operational workflows that may generically support various ML or ML-related functions in cellular networks such as 5GS (including, for example, the 5GC, NG-RAN(s), the 5G management system, etc.). Some aspects or embodiments may be standardized or otherwise incorporated into the 3GPP 5G standards, for example in the 3GPP TS 28.105.
[0022] Example Generic ML Workflow
[0023] An example generic ML operational workflow in the lifecycle of an ML entity is depicted in Figure 1. The example of Figure 1 depicts actions taken by a ML training module 105, a ML testing module 110, a ML emulation module 115, a ML loading module 120, and an ML inference module 125. It will be understood that the modules may be implemented in hardware, software, firmware, and / or some combination thereof. In some embodiments, two or more of the various modules may be implemented by the same element and / or system (e.g., by the same hardware / software / firmware / etc.) while, in other embodiments, each of the various modules may be implemented by different elements and / or systems.
[0024] In embodiments, the workflow of Figure 1 may be implemented by system that is generically referred to herein as a “ML Management System” of a cellular network. In some embodiments, the ML Management System may be or include one or both of the MLT MnS producer and MLT MnS consumer. In some embodiments the ML Management System may be or include one or more of the ML-related modules depicted in Figure 1. In some embodiments, the ML Management System may be separate from, but configured to control the operation of, one or more of the MLT MnS producer, MLT MnS consumer, and ML-related entities depicted in Figure 1. To put it another way, in some embodiments the ML Management System may at least partially perform one or more of the various actions described herein (e.g., the training, testing, emulation, loading, deployment, etc.). In some additional or alternative embodiments, the ML Management System may facilitate performance of such an action. An example of such facilitation may include providing control signals to a different entity such as a MLT MnS producer, an MLT MnS consumer, a ML- related module of Figure 1, etc., which in turn perform the described action. In some embodiments, the ML Management System may partially perform the action and partially facilitate performance of the action. As an example, the ML Management System may provide a control signal to a MLT MnS producer / consumer or ML-related module of Figure 1, and then analyze data that results from such an action to identify, for example, whether to re-train the ML entity or forward the ML entity for emulation (or some other decision or analysis). In some embodiments the ML Management System may be distributed across a plurality of electronic devices (e.g., an electronic system), or may be implemented in a single electronic device. It will be understood that the above is intended as a non-limiting discussion of how an ML Management System may function and / or be implemented, and other embodiments may include additional or alternative variations in one or more respects.
[0025] As may be seen in Figure 1, the example workflow may involve 4 main phases: training, emulation, deployment, and inference. Generally, one or more of the phases may be performed by a machine learning training (MLT) management services (MnS) producer, which may be responsible for generating an ML entity. The MnS producer may be responsible, for example, for the training and / or emulation phases. The MnS producer may then provide the ML entity to a MnS consumer, which may be responsible for implementation of the model. For example, the MnS consumer may be responsible for the deployment and / or inference phases. In some embodiments, the deployment phase may be performed at least partially be both the MnS producer and the MnS consumer. In some embodiments, initiation of the workflow may be done based on a request from the MnS consumer to the MnS producer for a ML entity (e.g., consumer-initiated). In some embodiments, initiation of the workflow may be done based on identification by the MnS producer that such an ML entity is needed by the MnS consumer (e.g., producer-initiated). In some embodiments, the MnS producer and the MnS consumer may be implemented in a same electronic device, a same piece of hardware, a same piece of firmware, a same piece of software, or some combination thereof. In other embodiments, the MnS producer and the MnS consumer may be implemented in different devices or electronic systems.
[0026] Non-limiting example tasks for each phase may include one or more of the following:
[0027] Training phase:
[0028] ML training: ML training may be performed by the ML training module 105. Training may include or relate to initial training and re-training of an ML entity or a group of ML entities. As used herein, the term “ML entity” may refer to a ML model that is configured to perform a given function or functions. ML training may also include validation of the trained ML entity to evaluate the performance variance when the ML entity performs on the training data and validation data. If the validation result does not meet the expectation (e.g., the variance is not acceptable or some other value is not within a predetermined range), the ML entity may need to be re-trained. The ML training may be considered to be the initial phase of the workflow.
[0029] ML testing: ML testing may be performed by the ML testing module 110. Testing of the validated ML entity may be performed to evaluate the performance of the trained ML entity when it performs on testing data. If the testing result meets one or more expectation, the ML entity may proceed with the next phase, otherwise the ML entity may need to be re-trained.
[0030] Emulation phase:
[0031] ML emulation: ML emulation may be performed by the ML emulation module 115. Emulation may refer to running an ML entity or Al / ML inference function for inference in an emulation environment. In some embodiments, the emulation environment may be the target network or system environment in which inference is to be performed. In other embodiments, the emulation environment may be different than the target network or system environment, but may simulate the target network / system environment. In embodiments, the ML emulation module 115 may be configured to evaluate the inference performance in the emulation environment prior to applying the ML entity to the target network or system. To put it another way, the ML emulation may be performed on test data that simulates the real-world data (e.g., data from a real-world environment, which may also be referred to as a “live” environment) which will be provided to the ML entity in the inference environment. In this way, the ML entity (e.g., the ML model) may be evaluated in an environment that simulates the real-world inference environment, but does not impact or systems associated with the real-world inference environment.
[0032] As a specific example, an ML entity may be related to a 5GC system. If the ML entity or ML model had an error, operating the 5GC system based on the ML entity / model may negatively impact the 5G network and / or users of the 5G network. By performing ML emulation, the results of inference of the ML entity / model in the simulated environment may be reviewed, which may significantly reduce the possibility of negative impact to the 5G network and / or users thereof.
[0033] It will be understood that the emulation phase may not be needed in some cases, for example, a re-trained ML entity with some small changes may not have to be evaluated in the emulation environment.
[0034] Loading phase:
[0035] ML entity loading: ML loading may be performed by ML loading module 120. In some embodiments, loading may also be referred to as “deployment.” Loading may refer to loading of a trained ML entity to the target ML inference function that will use the ML entity for inference in a real-world environment (e.g., on real-world data). It will be understood that the ML entity loading phase may not be needed in some cases, for example when the training function and inference function are co-located. Inference phase:
[0036] AI / ML inference: Inference may be performed by ML inference module 125. As used herein, inference may refer to application of the ML entity in a real-world setting (e.g., a setting that is not related to test data as may be used during validation or emulation data as may be used during the emulation phase). The ML inference module 125 may collect real-world data from the 5GS which is provided to the ML entity / ML model for processing to perform one or more of the 5GS-related functions described above.
[0037] Example Management Capabilities
[0038] Each phase of the workflow described above with respect to Figure 1 may be managed by the corresponding management capabilities. These capabilities may be performed, managed by, or otherwise associated with one or more of the ML-related modules depicted in Figure 1. Some examples of such capabilities may include or relate to:
[0039] Management capabilities for ML training phase (as may be performed by the ML training module 105 and / or the ML testing module 110): including the management capabilities for consumer-requested ML entity training, control of producer-initiated ML entity training, control and monitoring of the training process, ML entity validation performance report, consumer-requested ML entity testing, control of producer-initiated ML entity testing, etc.
[0040] Management capabilities for ML emulation phase (as may be performed by the ML emulation module 115): including the management capabilities for consumer- requested ML emulation, control of producer-initiated ML emulation, control and monitoring of the emulation process, ML emulation performance report, etc.
[0041] Management capabilities for loading phase (as may be performed by the ML loading module 120): including the management capabilities for consumer- requested ML entity loading, control of producer-initiated ML entity loading, control and monitoring of the ML entity loading process.
[0042] Management capabilities for AI / ML inference phase (as may be performed by the ML inference module 125): including the management capabilities for configuration (e.g., control) of the ML inference function and / or ML entity, inference performance evaluation of the ML inference function or ML entity, etc. SYSTEMS AND IMPLEMENTATIONS
[0043] Figures 2-6 illustrate various systems, devices, and components that may implement aspects of disclosed embodiments.
[0044] Figure 2 illustrates a network 200 in accordance with various embodiments. The network 200 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.
[0045] The network 200 may include a UE 202, which may include any mobile or non-mobile computing device designed to communicate with a RAN 204 via an over-the-air connection. The UE 202 may be communicatively coupled with the RAN 204 by a Uu interface. The UE 202 may be, but is not limited to, a smartphone, tablet computer, wearable computer device, desktop computer, laptop computer, in-vehicle infotainment, in-car entertainment device, instrument cluster, head-up display device, onboard diagnostic device, dashtop mobile equipment, mobile data terminal, electronic engine management system, electronic / engine control unit, electronic / engine control module, embedded system, sensor, microcontroller, control module, engine management system, networked appliance, machine-type communication device, M2M or D2D device, loT device, etc.
[0046] In some embodiments, the network 200 may include a plurality of UEs coupled directly with one another via a sidelink interface. The UEs may be M2M / D2D devices that communicate using physical sidelink channels such as, but not limited to, PSBCH, PSDCH, PSSCH, PSCCH, PSFCH, etc.
[0047] In some embodiments, the UE 202 may additionally communicate with an AP 206 via an over-the-air connection. The AP 206 may manage a WLAN connection, which may serve to offload some / all network traffic from the RAN 204. The connection between the UE 202 and the AP 206 maybe consistent with any IEEE 802.11 protocol, wherein the AP 206 could be a wireless fidelity (Wi-Fi®) router. In some embodiments, the UE 202, RAN 204, and AP 206 may utilize cellular-WLAN aggregation (for example, LWA / LW1P). Cellular-WLAN aggregation may involve the UE 202 being configured by the RAN 204 to utilize both cellular radio resources and WLAN resources.
[0048] The RAN 204 may include one or more access nodes, for example, AN 208. AN 208 may terminate air-interface protocols for the UE 202 by providing access stratum protocols including RRC, PDCP, RLC, MAC, and LI protocols. In this manner, the AN 208 may enable data / voice connectivity between CN 220 and the UE 202. In some embodiments, the AN 208 may be implemented in a discrete device or as one or more software entities running on server computers as part of, for example, a virtual network, which may be referred to as a CRAN or virtual baseband unit pool. The AN 208 be referred to as a BS, gNB, RAN node, eNB, ng-eNB, NodeB, RSU, TRxP, TRP, etc. The AN 208 may be a macrocell base station or a low power base station for providing femtocells, picocells or other like cells having smaller coverage areas, smaller user capacity, or higher bandwidth compared to macrocells.
[0049] In embodiments in which the RAN 204 includes a plurality of ANs, they may be coupled with one another via an X2 interface (if the RAN 204 is an LTE RAN) or an Xn interface (if the RAN 204 is a 5G RAN). The X2 / Xn interfaces, which may be separated into control / user plane interfaces in some embodiments, may allow the ANs to communicate information related to handovers, data / context transfers, mobility, load management, interference coordination, etc.
[0050] The ANs of the RAN 204 may each manage one or more cells, cell groups, component carriers, etc. to provide the UE 202 with an air interface for network access. The UE 202 may be simultaneously connected with a plurality of cells provided by the same or different ANs of the RAN 204. For example, the UE 202 and RAN 204 may use carrier aggregation to allow the UE 202 to connect with a plurality of component carriers, each corresponding to a Pcell or ScelL In dual connectivity scenarios, a first AN may be a master node that provides an MCG and a second AN may be secondary node that provides an SCG. The first / second ANs may be any combination of eNB, gNB, ng-eNB, etc.
[0051] The RAN 204 may provide the air interface over a licensed spectrum or an unlicensed spectrum. To operate in the unlicensed spectrum, the nodes may use LAA, eLAA, and / or feLAA mechanisms based on CA technology with PCells / Scells. Prior to accessing the unlicensed spectrum, the nodes may perform medium / carrier-sensing operations based on, for example, a listen-before-talk (LBT) protocol.
[0052] In V2X scenarios the UE 202 or AN 208 may be or act as a RSU, which may refer to any transportation infrastructure entity used for V2X communications. An RSU may be implemented in or by a suitable AN or a stationary (or relatively stationary) UE. An RSU implemented in or by: a UE may be referred to as a “UE-type RSU”; an eNB may be referred to as an “eNB-type RSU”; a gNB may be referred to as a “gNB-type RSU”; and the like. In one example, an RSU is a computing device coupled with radio frequency circuitry located on a roadside that provides connectivity support to passing vehicle UEs. The RSU may also include internal data storage circuitry to store intersection map geometry, traffic statistics, media, as well as applications / software to sense and control ongoing vehicular and pedestrian traffic. The RSU may provide very low latency communications required for high speed events, such as crash avoidance, traffic warnings, and the like. Additionally or alternatively, the RSU may provide other cellular / WLAN communications services. The components of the RSU may be packaged in a weatherproof enclosure suitable for outdoor installation, and may include a network interface controller to provide a wired connection (e.g., Ethernet) to a traffic signal controller or a backhaul network.
[0053] In some embodiments, the RAN 204 maybe an LTE RAN 210 with eNBs, for example, eNB 212. The LTE RAN 210 may provide an LTE air interface with the following characteristics: SCS of 15 kHz; CP-OFDM waveform for DL and SC-FDMA waveform for UL; turbo codes for data and TBCC for control; etc. The LTE air interface may rely on CS1-RS for CS1 acquisition and beam management; PDSCH / PDCCH DMRS for PDSCH / PDCCH demodulation; and CRS for cell search and initial acquisition, channel quality measurements, and channel estimation for coherent demodulation / detection at the UE. The LTE air interface may operating on sub-6 GHz bands.
[0054] In some embodiments, the RAN 204 may be an NG-RAN 214 with gNBs, for example, gNB 216, or ng-eNBs, for example, ng-eNB 218. The gNB 216 may connect with 5G-enabled UEs using a 5G NR interface. The gNB 216 may connect with a 5G core through an NG interface, which may include an N2 interface or an N3 interface. The ng-eNB 218 may also connect with the 5G core through an NG interface, but may connect with a UE via an LTE air interface. The gNB 216 and the ng-eNB 218 may connect with each other over an Xn interface.
[0055] In some embodiments, the NG interface may be split into two parts, an NG user plane (NG-U) interface, which carries traffic data between the nodes of the NG-RAN 214 and a UPF 248 (e.g., N3 interface), and an NG control plane (NG-C) interface, which is a signaling interface between the nodes of the NG-RAN214 and an AMF 244 (e.g., N2 interface).
[0056] The NG-RAN 214 may provide a 5G-NR air interface with the following characteristics: variable SCS; CP-OFDM for DL, CP-OFDM and DFT-s-OFDM for UL; polar, repetition, simplex, and Reed-Muller codes for control and LDPC for data. The 5G-NR air interface may rely on CS1-RS, PDSCH / PDCCH DMRS similar to the LTE air interface. The 5G- NR air interface may not use a CRS, but may use PBCH DMRS for PBCH demodulation; PTRS for phase tracking for PDSCH; and tracking reference signal for time tracking. The 5G-NR air interface may operating on FR1 bands that include sub-6 GHz bands or FR2 bands that include bands from 24.25 GHz to 52.6 GHz. The 5G-NR air interface may include an SSB that is an area of a downlink resource grid that includes PSS / SSS / PBCH. In some embodiments, the 5G-NR air interface may utilize BWPs for various purposes. For example, BWP can be used for dynamic adaptation of the SCS. For example, the UE 202 can be configured with multiple BWPs where each BWP configuration has a different SCS. When a BWP change is indicated to the UE 202, the SCS of the transmission is changed as well. Another use case example of BWP is related to power saving. In particular, multiple BWPs can be configured for the UE 202 with different amount of frequency resources (for example, PRBs) to support data transmission under different traffic loading scenarios. A BWP containing a smaller number of PRBs can be used for data transmission with small traffic load while allowing power saving at the UE 202 and in some cases at the gNB 216. A BWP containing a larger number of PRBs can be used for scenarios with higher traffic load.
[0057] The RAN 204 is communicatively coupled to CN 220 that includes network elements to provide various functions to support data and telecommunications services to customers / subscribers (for example, users of UE 202). The components of the CN 220 may be implemented in one physical node or separate physical nodes. In some embodiments, NFV may be utilized to virtualize any or all of the functions provided by the network elements of the CN 220 onto physical compute / storage resources in servers, switches, etc. A logical instantiation of the CN 220 may be referred to as a network slice, and a logical instantiation of a portion of the CN 220 may be referred to as a network sub-slice.
[0058] In some embodiments, the CN 220 may be an LTE CN 222, which may also be referred to as an EPC. The LTE CN 222 may include MME 224, SGW 226, SGSN 228, HSS 230, PGW 232, and PCRF 234 coupled with one another over interfaces (or “reference points”) as shown. Functions of the elements of the LTE CN 222 may be briefly introduced as follows.
[0059] The MME 224 may implement mobility management functions to track a current location of the UE 202 to facilitate paging, bearer activation / deactivation, handovers, gateway selection, authentication, etc.
[0060] The SGW 226 may terminate an SI interface toward the RAN and route data packets between the RAN and the LTE CN 222. The SGW 226 may be a local mobility anchor point for inter-RAN node handovers and also may provide an anchor for inter-3GPP mobility. Other responsibilities may include lawful intercept, charging, and some policy enforcement.
[0061] The SGSN 228 may track a location of the UE 202 and perform security functions and access control. In addition, the SGSN 228 may perform inter-EPC node signaling for mobility between different RAT networks; PDN and S-GW selection as specified by MME 224; MME selection for handovers; etc. The S3 reference point between the MME 224 and the SGSN 228 may enable user and bearer information exchange for inter-3GPP access network mobility in idle / active states.
[0062] The HSS 230 may include a database for network users, including subscription- related information to support the network entities’ handling of communication sessions. The HSS 230 can provide support for routing / roaming, authentication, authorization, naming / addressing resolution, location dependencies, etc. An S6a reference point between the HSS 230 and the MME 224 may enable transfer of subscription and authentication data for authenticating / authorizing user access to the LTE CN 220.
[0063] The PGW 232 may terminate an SGi interface toward a data network (DN) 236 that may include an application / content server 238. The PGW 232 may route data packets between the LTE CN 222 and the data network 236. The PGW 232 may be coupled with the SGW 226 by an S5 reference point to facilitate user plane tunneling and tunnel management. The PGW 232 may further include a node for policy enforcement and charging data collection (for example, PCEF). Additionally, the SGi reference point between the PGW 232 and the data network 2 36 may be an operator external public, a private PDN, or an intraoperator packet data network, for example, for provision of IMS services. The PGW 232 may be coupled with a PCRF 234 via a Gx reference point.
[0064] The PCRF 234 is the policy and charging control element of the LTE CN 222. The PCRF 234 may be communicatively coupled to the app / content server 238 to determine appropriate QoS and charging parameters for service flows. The PCRF 232 may provision associated rules into a PCEF (via Gx reference point) with appropriate TFT and QC1.
[0065] In some embodiments, the CN 220 may be a 5GC 240. The 5GC 240 may include an AUSF 242, AMF 244, SMF 246, UPF 248, NSSF 250, NEF 252, NRF 254, PCF 256, UDM 258, and AF 260 coupled with one another over interfaces (or “reference points”) as shown. Functions of the elements of the 5GC 240 may be briefly introduced as follows.
[0066] The AUSF 242 may store data for authentication of UE 202 and handle authentication-related functionality. The AUSF 242 may facilitate a common authentication framework for various access types. In addition to communicating with other elements of the 5GC 240 over reference points as shown, the AUSF 242 may exhibit an Nausf servicebased interface.
[0067] The AMF 244 may allow other functions of the 5GC 240 to communicate with the UE 202 and the RAN 204 and to subscribe to notifications about mobility events with respect to the UE 202. The AMF 244 may be responsible for registration management (for example, for registering UE 202), connection management, reachability management, mobility management, lawful interception of AMF-related events, and access authentication and authorization. The AMF 244 may provide transport for SM messages between the UE 202 and the SMF 246, and act as a transparent proxy for routing SM messages. AMF 244 may also provide transport for SMS messages between UE 202 and an SMSF. AMF 244 may interact with the AUSF 242 and the UE 202 to perform various security anchor and context management functions. Furthermore, AMF 244 may be a termination point of a RAN CP interface, which may include or be an N2 reference point between the RAN 204 and the AMF 244; and the AMF 244 may be a termination point of NAS (Nl) signaling, and perform NAS ciphering and integrity protection. AMF 244 may also support NAS signaling with the UE 202 over an N3 1WF interface.
[0068] The SMF 246 maybe responsible for SM (for example, session establishment, tunnel management between UPF 248 and AN 208); UE IP address allocation and management (including optional authorization); selection and control of UP function; configuring traffic steering at UPF 248 to route traffic to proper destination; termination of interfaces toward policy control functions; controlling part of policy enforcement, charging, and QoS; lawful intercept (for SM events and interface to LI system); termination of SM parts of NAS messages; downlink data notification; initiating AN specific SM information, sent via AMF 244 over N2 to AN 208; and determining SSC mode of a session. SM may refer to management of a PDU session, and a PDU session or “session” may refer to a PDU connectivity service that provides or enables the exchange of PDUs between the UE 202 and the data network 236.
[0069] The UPF 248 may act as an anchor point for intra-RAT and inter-RAT mobility, an external PDU session point of interconnect to data network 236, and a branching point to support multi-homed PDU session. The UPF 248 may also perform packet routing and forwarding, perform packet inspection, enforce the user plane part of policy rules, lawfully intercept packets (UP collection), perform traffic usage reporting, perform QoS handling for a user plane (e.g., packet filtering, gating, UL / DL rate enforcement), perform uplink traffic verification (e.g., SDF-to-QoS flow mapping), transport level packet marking in the uplink and downlink, and perform downlink packet buffering and downlink data notification triggering. UPF 248 may include an uplink classifier to support routing traffic flows to a data network.
[0070] The NSSF 250 may select a set of network slice instances serving the UE 202. The NSSF 250 may also determine allowed NSSA1 and the mapping to the subscribed S-NSSAls, if needed. The NSSF 250 may also determine the AMF set to be used to serve the UE 202, or a list of candidate AMFs based on a suitable configuration and possibly by querying the NRF 254. The selection of a set of network slice instances for the UE 202 may be triggered by the AMF 244 with which the UE 202 is registered by interacting with the NSSF 250, which may lead to a change of AMF. The NSSF 250 may interact with the AMF 244 via an N22 reference point; and may communicate with another NSSF in a visited network via an N31 reference point (not shown). Additionally, the NSSF 250 may exhibit an Nnssf service-based interface.
[0071] The NEF 252 may securely expose services and capabilities provided by 3GPP network functions for third party, internal exposure / re-exposure, AFs (e.g., AF 260), edge computing or fog computing systems, etc. In such embodiments, the NEF 252 may authenticate, authorize, or throttle the AFs. NEF 252 may also translate information exchanged with the AF 260 and information exchanged with internal network functions. For example, the NEF 252 may translate between an AF-Service-ldentifier and an internal 5GC information. NEF 252 may also receive information from other NFs based on exposed capabilities of other NFs. This information may be stored at the NEF 252 as structured data, or at a data storage NF using standardized interfaces. The stored information can then be re-exposed by the NEF 252 to other NFs and AFs, or used for other purposes such as analytics. Additionally, the NEF 252 may exhibit an Nnef service-based interface.
[0072] The NRF 254 may support service discovery functions, receive NF discovery requests from NF instances, and provide the information of the discovered NF instances to the NF instances. NRF 254 also maintains information of available NF instances and their supported services. As used herein, the terms “instantiate,” “instantiation,” and the like may refer to the creation of an instance, and an “instance” may refer to a concrete occurrence of an object, which may occur, for example, during execution of program code. Additionally, the NRF 254 may exhibit the Nnrf service-based interface.
[0073] The PCF 256 may provide policy rules to control plane functions to enforce them, and may also support unified policy framework to govern network behavior. The PCF 256 may also implement a front end to access subscription information relevant for policy decisions in a UDR of the UDM 258. In addition to communicating with functions over reference points as shown, the PCF 256 exhibit an Npcf service-based interface.
[0074] The UDM 258 may handle subscription-related information to support the network entities’ handling of communication sessions, and may store subscription data of UE 202. For example, subscription data may be communicated via an N8 reference point between the UDM 258 and the AMF 244. The UDM 258 may include two parts, an application front end and a UDR. The UDR may store subscription data and policy data for the UDM 258 and the PCF 256, and / or structured data for exposure and application data (including PFDs for application detection, application request information for multiple UEs 202) for the NEF 252. The Nudr service-based interface may be exhibited by the UDR 221 to allow the UDM 258, PCF 256, and NEF 252 to access a particular set of the stored data, as well as to read, update (e.g., add, modify), delete, and subscribe to notification of relevant data changes in the UDR. The UDM may include a UDM-FE, which is in charge of processing credentials, location management, subscription management and so on. Several different front ends may serve the same user in different transactions. The UDM-FE accesses subscription information stored in the UDR and performs authentication credential processing, user identification handling, access authorization, registration / mobility management, and subscription management. In addition to communicating with other NFs over reference points as shown, the UDM 258 may exhibit the Nudm service-based interface.
[0075] The AF 260 may provide application influence on traffic routing, provide access to NEF, and interact with the policy framework for policy control.
[0076] In some embodiments, the 5GC 240 may enable edge computing by selecting operator / 3rdparty services to be geographically close to a point that the UE 202 is attached to the network. This may reduce latency and load on the network. To provide edgecomputing implementations, the 5GC 240 may select a UPF 248 close to the UE 202 and execute traffic steering from the UPF 248 to data network 236 via the N6 interface. This may be based on the UE subscription data, UE location, and information provided by the AF 260. In this way, the AF 260 may influence UPF (re)selection and traffic routing. Based on operator deployment, when AF 260 is considered to be a trusted entity, the network operator may permit AF 260 to interact directly with relevant NFs. Additionally, the AF 260 may exhibit an Naf service-based interface.
[0077] The data network 236 may represent various network operator services, Internet access, or third party services that may be provided by one or more servers including, for example, application / content server 238.
[0078] Figure 3 schematically illustrates a wireless network 300 in accordance with various embodiments. The wireless network 300 may include a UE 302 in wireless communication with an AN 304. The UE 302 and AN 304 may be similar to, and substantially interchangeable with, like-named components described elsewhere herein.
[0079] The UE 302 may be communicatively coupled with the AN 304 via connection 306.
[0080] The connection 306 is illustrated as an air interface to enable communicative coupling, and can be consistent with cellular communications protocols such as an LTE protocol or a 5G NR protocol operating at mmWave or sub-6GHz frequencies.
[0081] The UE 302 may include a host platform 308 coupled with a modem platform 310. The host platform 308 may include application processing circuitry 312, which may be coupled with protocol processing circuitry 314 of the modem platform 310. The application processing circuitry 312 may run various applications for the UE 302 that source / sink application data. The application processing circuitry 312 may further implement one or more layer operations to transmit / receive application data to / from a data network. These layer operations may include transport (for example UDP) and Internet (for example, IP) operations
[0082] The protocol processing circuitry 314 may implement one or more of layer operations to facilitate transmission or reception of data over the connection 306. The layer operations implemented by the protocol processing circuitry 314 may include, for example, MAC, RLC, PDCP, RRC and NAS operations.
[0083] The modem platform 310 may further include digital baseband circuitry 316 that may implement one or more layer operations that are “below” layer operations performed by the protocol processing circuitry 314 in a network protocol stack. These operations may include, for example, PHY operations including one or more of HARQ-ACK functions, scrambling / descrambling, encoding / decoding, layer mapping / de-mapping, modulation symbol mapping, received symbol / bit metric determination, multi-antenna port precoding / decoding, which may include one or more of space-time, space-frequency or spatial coding, reference signal generation / detection, preamble sequence generation and / or decoding, synchronization sequence generation / detection, control channel signal blind decoding, and other related functions.
[0084] The modem platform 310 may further include transmit circuitry 318, receive circuitry 320, RF circuitry 322, and RF front end (RFFE) 324, which may include or connect to one or more antenna panels 326. Briefly, the transmit circuitry 318 may include a digital-to-analog converter, mixer, intermediate frequency (IF) components, etc.; the receive circuitry 320 may include an analog-to-digital converter, mixer, IF components, etc.; the RF circuitry 322 may include a low-noise amplifier, a power amplifier, power tracking components, etc.; RFFE 324 may include filters (for example, surface / bulk acoustic wave filters), switches, antenna tuners, beamforming components (for example, phase-array antenna components), etc. The selection and arrangement of the components of the transmit circuitry 318, receive circuitry 320, RF circuitry 322, RFFE 324, and antenna panels 326 (referred generically as “transmit / receive components”) may be specific to details of a specific implementation such as, for example, whether communication is TDM or FDM, in mmWave or sub-6 gHz frequencies, etc. In some embodiments, the transmit / receive components may be arranged in multiple parallel transmit / receive chains, may be disposed in the same or different chips / modules, etc.
[0085] In some embodiments, the protocol processing circuitry 314 may include one or more instances of control circuitry (not shown) to provide control functions for the transmit / receive components.
[0086] A UE reception may be established by and via the antenna panels 326, RFFE 324, RF circuitry 322, receive circuitry 320, digital baseband circuitry 316, and protocol processing circuitry 314. In some embodiments, the antenna panels 326 may receive a transmission from the AN 304 by receive-beamforming signals received by a plurality of antennas / antenna elements of the one or more antenna panels 326.
[0087] A UE transmission may be established by and via the protocol processing circuitry 314, digital baseband circuitry 316, transmit circuitry 318, RF circuitry 322, RFFE 324, and antenna panels 326. In some embodiments, the transmit components of the UE 304 may apply a spatial filter to the data to be transmitted to form a transmit beam emitted by the antenna elements of the antenna panels 326.
[0088] Similar to the UE 302, the AN 304 may include a host platform 328 coupled with a modem platform 330. The host platform 328 may include application processing circuitry 332 coupled with protocol processing circuitry 334 of the modem platform 330. The modem platform may further include digital baseband circuitry 336, transmit circuitry 338, receive circuitry 340, RF circuitry 342, RFFE circuitry 344, and antenna panels 346. The components of the AN 304 may be similar to and substantially interchangeable with like-named components of the UE 302. In addition to performing data transmission / reception as described above, the components of the AN 308 may perform various logical functions that include, for example, RNC functions such as radio bearer management, uplink and downlink dynamic radio resource management, and data packet scheduling.
[0089] Figure 4 is a block diagram illustrating components, according to some example embodiments, able to read instructions from a machine-readable or computer-readable medium (e.g., a non-transitory machine-readable storage medium) and perform any one or more of the methodologies discussed herein. Specifically, Figure 4 shows a diagrammatic representation of hardware resources 400 including one or more processors (or processor cores) 410, one or more memory / storage devices 420, and one or more communication resources 430, each of which may be communicatively coupled via a bus 440 or other interface circuitry. For embodiments where node virtualization (e.g., NFV) is utilized, a hypervisor 402 may be executed to provide an execution environment for one or more network slices / sub-slices to utilize the hardware resources 400.
[0090] The processors 410 may include, for example, a processor 412 and a processor 414. The processors 410 may be, for example, a central processing unit (CPU), a reduced instruction set computing (RISC) processor, a complex instruction set computing (CISC) processor, a graphics processing unit (GPU), a DSP such as a baseband processor, an ASIC, an FPGA, a radio-frequency integrated circuit (RF1C), another processor (including those discussed herein), or any suitable combination thereof.
[0091] The memory / storage devices 420 may include main memory, disk storage, or any suitable combination thereof. The memory / storage devices 420 may include, but are not limited to, any type of volatile, non-volatile, or semi-volatile memory such as dynamic random access memory (DRAM), static random access memory (SRAM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), Flash memory, solid-state storage, etc.
[0092] The communication resources 430 may include interconnection or network interface controllers, components, or other suitable devices to communicate with one or more peripheral devices 404 or one or more databases 406 or other network elements via a network 408. For example, the communication resources 430 may include wired communication components (e.g., for coupling via USB, Ethernet, etc.), cellular communication components, NFC components, Bluetooth® (or Bluetooth® Low Energy) components, Wi-Fi® components, and other communication components.
[0093] Instructions 450 may comprise software, a program, an application, an applet, an app, or other executable code for causing at least any of the processors 410 to perform any one or more of the methodologies discussed herein. The instructions 450 may reside, completely or partially, within at least one of the processors 410 (e.g., within the processor’s cache memory), the memory / storage devices 420, or any suitable combination thereof. Furthermore, any portion of the instructions 450 may be transferred to the hardware resources 400 from any combination of the peripheral devices 404 or the databases 406. Accordingly, the memory of processors 410, the memory / storage devices 420, the peripheral devices 404, and the databases 406 are examples of computer-readable and machine-readable media. Figure 5 illustrates a network 500 in accordance with various embodiments. The network 500 may operate in a matter consistent with 3GPP technical specifications or technical reports for 6G systems. In some embodiments, the network 500 may operate concurrently with network 200. For example, in some embodiments, the network 500 may share one or more frequency or bandwidth resources with network 200. As one specific example, a UE (e.g., UE 502) maybe configured to operate in both network 500 and network 200. Such configuration may be based on a UE including circuitry configured for communication with frequency and bandwidth resources of both networks 200 and 500. In general, several elements of network 500 may share one or more characteristics with elements of network 200. For the sake of brevity and clarity, such elements may not be repeated in the description of network 500.
[0094] The network 500 may include a UE 502, which may include any mobile or non-mobile computing device designed to communicate with a RAN 508 via an over-the-air connection. The UE 502 may be similar to, for example, UE 202. The UE 502 may be, but is not limited to, a smartphone, tablet computer, wearable computer device, desktop computer, laptop computer, in-vehicle infotainment, in-car entertainment device, instrument cluster, head-up display device, onboard diagnostic device, dashtop mobile equipment, mobile data terminal, electronic engine management system, electronic / engine control unit, electronic / engine control module, embedded system, sensor, microcontroller, control module, engine management system, networked appliance, machine-type communication device, M2M or D2D device, loT device, etc.
[0095] Although not specifically shown in Figure 5, in some embodiments the network 500 may include a plurality ofUEs coupled directly with one another via a sidelink interface. The UEs may be M2M / D2D devices that communicate using physical sidelink channels such as, but not limited to, PSBCH, PSDCH, PSSCH, PSCCH, PSFCH, etc. Similarly, although not specifically shown in Figure 5, the UE 502 may be communicatively coupled with an AP such as AP 206 as described with respect to Figure 2. Additionally, although not specifically shown in Figure 5, in some embodiments the RAN 508 may include one or more ANss such as AN 208 as described with respect to Figure 2. The RAN 508 and / or the AN of the RAN 508 may be referred to as a base station (BS), a RAN node, or using some other term or name.
[0096] The UE 502 and the RAN 508 may be configured to communicate via an air interface that may be referred to as a sixth generation (6G) air interface. The 6G air interface may include one or more features such as communication in a terahertz (THz) or sub-THz bandwidth, or joint communication and sensing. As used herein, the term “joint communication and sensing” may refer to a system that allows for wireless communication as well as radar-based sensing via various types of multiplexing. As used herein, THz or sub- THz bandwidths may refer to communication in the 80 GHz and above frequency ranges. Such frequency ranges may additionally or alternatively be referred to as “millimeter wave” or “mmWave” frequency ranges.
[0097] The RAN 508 may allow for communication between the UE 502 and a 6G core network (CN) 510. Specifically, the RAN 508 may facilitate the transmission and reception of data between the UE 502 and the 6G CN 510. The 6G CN 510 may include various functions such as NSSF 250, NEF 252, NRF 254, PCF 256, UDM 258, AF 260, SMF 246, and AUSF 242. The 6G CN 510 may additional include UPF 248 and DN 236 as shown in Figure 5.
[0098] Additionally, the RAN 508 may include various additional functions that are in addition to, or alternative to, functions of a legacy cellular network such as a 4G or 5G network. Two such functions may include a Compute Control Function (Comp CF) 524 and a Compute Service Function (Comp SF) 536. The Comp CF 524 and the Comp SF 536 may be parts or functions of the Computing Service Plane. Comp CF 524 may be a control plane function that provides functionalities such as management of the Comp SF 536, computing task context generation and management (e.g., create, read, modify, delete), interaction with the underlying computing infrastructure for computing resource management, etc.. Comp SF 536 may be a user plane function that serves as the gateway to interface computing service users (such as UE 502) and computing nodes behind a Comp SF instance. Some functionalities of the Comp SF 536 may include: parse computing service data received from users to compute tasks executable by computing nodes; hold service mesh ingress gateway or service API gateway; service and charging policies enforcement; performance monitoring and telemetry collection, etc. In some embodiments, a Comp SF 536 instance may serve as the user plane gateway for a cluster of computing nodes. A Comp CF 524 instance may control one or more Comp SF 536 instances.
[0099] Two other such functions may include a Communication Control Function (Comm CF) 528 and a Communication Service Function (Comm SF) 538, which may be parts of the Communication Service Plane. The Comm CF 528 may be the control plane function for managingthe Comm SF 538, communication sessions creation / configuration / releasing, and managing communication session context. The Comm SF 538 may be a user plane function for data transport. Comm CF 528 and Comm SF 538 may be considered as upgrades of SMF 246 and UPF 248, which were described with respect to a 5G system in Figure 2. The upgrades provided by the Comm CF 528 and the Comm SF 538 may enable service-aware transport. For legacy (e.g., 4G or 5G) data transport, SMF 246 and UPF 248 may still be used.
[0100] Two other such functions may include a Data Control Function (Data CF) 522 and Data Service Function (Data SF) 532 may be parts of the Data Service Plane. Data CF 522 may be a control plane function and provides functionalities such as Data SF 532 management, Data service creation / configuration / releasing, Data service context management, etc. Data SF 532 may be a user plane function and serve as the gateway between data service users (such as UE 502 and the various functions of the 6G CN 510) and data service endpoints behind the gateway. Specific functionalities may include include: parse data service user data and forward to corresponding data service endpoints, generate charging data, report data service status.
[0101] Another such function may be the Service Orchestration and Chaining Function (SOCF) 520, which may discover, orchestrate and chain up communication / computing / data services provided by functions in the network. Upon receiving service requests from users, SOCF 520 may interact with one or more of Comp CF 524, Comm CF 528, and Data CF 522 to identify Comp SF 536, Comm SF 538, and Data SF 532 instances, configure service resources, and generate the service chain, which could contain multiple Comp SF 536, Comm SF 538, and Data SF 532 instances and their associated computing endpoints. Workload processing and data movement may then be conducted within the generated service chain. The SOCF 520 may also responsible for maintaining, updating, and releasing a created service chain.
[0102] Another such function maybe the service registration function (SRF) 514, which may act as a registry for system services provided in the user plane such as services provided by service endpoints behind Comp SF 536 and Data SF 532 gateways and services provided by the UE 502. The SRF 514 may be considered a counterpart of NRF 254, which may act as the registry for network functions.
[0103] Other such functions may include an evolved service communication proxy (eSCP) and service infrastructure control function (S1CF) 526, which may provide service communication infrastructure for control plane services and user plane services. The eSCP may be related to the service communication proxy (SCP) of 5G with user plane service communication proxy capabilities being added. The eSCP is therefore expressed in two parts: eCSP-C 512 and eSCP-U 534, for control plane service communication proxy and user plane service communication proxy, respectively. The S1CF 526 may control and configure eCSP instances in terms of service traffic routing policies, access rules, load balancing configurations, performance monitoring, etc. Another such function is the AMF 544. The AMF 544 may be similar to 244, but with additional functionality. Specifically, the AMF 544 may include potential functional repartition, such as move the message forwarding functionality from the AMF 544 to the RAN 508.
[0104] Another such function is the service orchestration exposure function (SOEF) 518. The SOEF may be configured to expose service orchestration and chaining services to external users such as applications.
[0105] The UE 502 may include an additional function that is referred to as a computing client service function (comp CSF) 504. The comp CSF 504 may have both the control plane functionalities and user plane functionalities, and may interact with corresponding network side functions such as SOCF 520, Comp CF 524, Comp SF 536, Data CF 522, and / or Data SF 532 for service discovery, request / response, compute task workload exchange, etc. The Comp CSF 504 may also work with network side functions to decide on whether a computing task should be run on the UE 502, the RAN 508, and / or an element of the 6G CN 510.
[0106] The UE 502 and / or the Comp CSF 504 may include a service mesh proxy 506. The service mesh proxy 506 may act as a proxy for service-to-service communication in the user plane. Capabilities of the service mesh proxy 506 may include one or more of addressing, security, load balancing, etc.
[0107] Figure 6 illustrates a simplified block diagram of ML-assisted communication between a UE 605 and a RAN 610, in accordance with various embodiments. More specifically, as described in further detail herein, ML models may be used or leveraged to facilitate over-the-air communication between UE 605 and RAN 610. In some embodiments, the various elements of the UE 605 and / or the RAN 610 may be similar to elements of the ML Management System as described herein.
[0108] One or both of the UE 605 and the RAN 610 may operate in a matter consistent with 3GPP technical specifications or technical reports for 6G systems. In some embodiments, the wireless cellular communication between the UE 605 and the RAN 610 may be part of, or operate concurrently with, networks 500, 200, and / or some other network described herein.
[0109] The UE 605 may be similar to, and share one or more features with, UE 502, UE 202, and / or some other UE described herein. The UE 605 may be, but is not limited to, a smartphone, tablet computer, wearable computer device, desktop computer, laptop computer, in-vehicle infotainment, in-car entertainment device, instrument cluster, head-up display device, onboard diagnostic device, dashtop mobile equipment, mobile data terminal, electronic engine management system, electronic / engine control unit, electronic / engine control module, embedded system, sensor, microcontroller, control module, engine management system, networked appliance, machine-type communication device, M2M or D2D device, loT device, etc. The RAN 610 may be similar to, and share one or more features with, RAN 214, RAN 508, and / or some other RAN described herein.
[0110] As may be seen in Figure 6, the ML-related elements of UE 605 may be similar to the ML-related elements of RAN 610. For the sake of discussion herein, description of the various elements will be provided from the point of view of the UE 605, however it will be understood that such discussion or description will apply to equally named / numbered elements of RAN 610, unless explicitly stated otherwise.
[0111] As previously noted, the UE 605 may include various elements or functions that are related to ML. Such elements may be implemented as hardware, software, firmware, and / or some combination thereof. In embodiments, one or more of the elements may be implemented as part of the same hardware (e.g., chip or multi-processor chip), software (e.g., a computing program), or firmware as another element.
[0112] One such element may be a data repository 615. The data repository 615 may be responsible for data collection and storage. Specifically, the data repository 615 may collect and store RAN configuration parameters, measurement data, performance key performance indicators (KPIs), model performance metrics, etc., for model training, update, and inference. More generally, collected data is stored into the repository. Stored data can be discovered and extracted by other elements from the data repository 615. For example, as maybe seen, the inference data selection / filter element 650 may retrieve data from the data repository 615. In various embodiments, the UE 605 may be configured to discover and request data from the data repository 610 in the RAN, and vice versa. More generally, the data repository 615 of the UE 605 maybe communicatively coupled with the data repository 615 of the RAN 610 such that the respective data repositories of the UE and the RAN may share collected data with one another.
[0113] Another such element may be a training data selection / filtering functional block 620. The training data selection / filter functional block 620 may be configured to generate training, validation, and testing datasets for model training. Training data may be extracted from the data repository 615. Data may be selected / filtered based on the specific ML entity / ML model to be trained. Data may optionally be transformed / augmented / pre- processed (e.g., normalized) before being loaded into datasets (e.g., by ML loading module 120). The training data selection / filter functional block 620 may label data in datasets for supervised learning. The produced datasets may then be fed into model training the model training functional block 625. In some embodiments, functional blocks 620 / 625 may be similar to the ML training module 105 and / or the ML testing module 110.
[0114] As noted above, another such element may be the model training functional block 625. This functional block may be responsible for training and updating (re-training) ML models. The selected model may be trained using the fed-in datasets (including training, validation, testing) from the training data selection / filtering functional block. The model training functional block 625 may produce trained and tested ML models which are ready for deployment. The produced trained and tested models can be stored in a model repository 635.
[0115] The model repository 635 may be responsible for ML models’ (both trained and untrained) storage and exposure. Trained / updated model(s) may be stored into the model repository 635. Model and model parameters may be discovered and requested by other functional blocks (e.g., the training data selection / filter functional block 620 and / or the model training functional block 625). In some embodiments, the UE 605 may discover and request ML models from the model repository 635 of the RAN 610. Similarly, the RAN 610 may be able to discover and / or request ML models from the model repository 635 of the UE 605. In some embodiments, the RAN 610 may configure models and / or model parameters in the model repository 635 of the UE 605.
[0116] Another such element may be a model management functional block 640. The model management functional block 640, which may be or may be similar to the ML Loading Module 120, may be responsible for management of the ML model produced by the model training functional block 625. Such management functions may include deployment of a trained model, monitoring model performance, etc. In model deployment, the model management functional block 640 may allocate and schedule hardware and / or software resources for inference, based on received trained and tested models. As used herein, “inference” refers to the process of using trained ML model(s) to generate data analytics, actions, policies, etc. based on input inference data. In performance monitoring, based on wireless performance KPIs and model performance metrics, the model management functional block 640 may decide to terminate the running model, start model re-training, select another model, etc. In embodiments, the model management functional block 640 of the RAN 610 may be able to configure model management policies in the UE 605 as shown.
[0117] Another such element may be an inference data selection / filtering functional block 650. The inference data selection / filter functional block 650 may be responsible for generating datasets for model inference at the inference functional block 645, as described below. Specifically, inference data may be extracted from the data repository 615. The inference data selection / filter functional block 650 may select and / or filter the data based on the deployed ML model. Data may be transformed / augmented / pre-processed following the same transformation / augmentation / pre-processing as those in training data selection / filtering as described with respect to functional block 620. The produced inference dataset may be fed into the inference functional block 645.
[0118] Another such element may be the inference functional block 645. The inference functional block 645 may be responsible for executing inference as described above. Specifically, the inference functional block 645 may consume the inference dataset provided by the inference data selection / filtering functional block 650, and generate one or more outcomes. Such outcomes may be or include data analytics, actions, policies, etc. The outcome(s) may be provided to the performance measurement functional block 630. In some embodiments, one or both of functional blocks 645 and 650 may be, or may be part of, ML Inference Module 125.
[0119] The performance measurement functional block 630 may be configured to measure model performance metrics (e.g., accuracy, model bias, run-time latency, etc.) of deployed and executing models based on the inference outcome(s) for monitoring purpose. Model performance data may be stored in the data repository 615.
[0120] EXAMPLE PROCEDURES
[0121] In some embodiments, the electronic device(s), network(s), system(s), chip(s) or component(s), or portions or implementations thereof, of Figures 2-6, or some other figure herein, may be configured to perform one or more processes, techniques, or methods as described herein, or portions thereof. One such process is depicted in Figure 7. The process of Figure may include or relate to a method to be performed by a ML Management System. The process may include training or facilitating training, at 705, of a ML model by a ML training module, wherein training of the ML model generates a trained ML model, and wherein the ML model is related to a cellular network system; testing or facilitating testing, at 710, of the trained ML model by a ML testing module, wherein testing of the ML trained ML module includes evaluation of application of the trained ML module to testing data; emulating or facilitating emulating, at 715, of the trained ML model by a ML emulation module, wherein emulation of the ML model involves performing inference using the ML model in an emulation environment that is different than an inference environment of the cellular network; and identifying or facilitating identifying, at 720 based on the emulation of the trained ML model, whether to re-train the ML model by the ML training module or provide the trained ML model to a ML inference module of the cellular network for inference in the inference environment
[0122] For one or more embodiments, at least one of the components set forth in one or more of the preceding figures may be configured to perform one or more operations, techniques, processes, and / or methods as set forth in the example section below. For example, the baseband circuitry as described above in connection with one or more of the preceding figures may be configured to operate in accordance with one or more of the examples set forth below. For another example, circuitry associated with a UE, base station, network element, etc. as described above in connection with one or more of the preceding figures may be configured to operate in accordance with one or more of the examples set forth below in the example section.
[0123] EXAMPLES
[0124] Example 1 may include the method for supporting Al / ML for 5G system, comprising at least one of the following phases:
[0125] ML training;
[0126] ML emulation;
[0127] ML entity loading;
[0128] Al / ML inference.
[0129] Example 2 may include the method of example 1 or some other example herein, where ML training phase further includes:
[0130] ML entity training;
[0131] ML entity validation;
[0132] ML entity testing.
[0133] Example 3 may include the method according to example 2 or some other example herein, wherein ML entity training includes initial training of a new ML entity or retraining of an ML entity.
[0134] Example 4 may include the method according to examples 2 and 3 or some other example herein, wherein ML entity training is the training of an ML entity or joint training of a group of ML entities. Example 5 may include the method according to examples 2 or some other example herein, wherein ML entity testing is the testing of an ML entity or joint testing of a group of ML entities.
[0135] Example 6 may include the method according to example 1 or some other example herein, wherein ML emulation is to evaluate the inference performance in the emulation environment prior to applying it to the target network or system, by running an ML entity or Al / ML inference function for inference in an emulation environment is to that is the same as, or can simulate, the target network or system environment where the ML entity or Al / ML inference function will be used for inference.
[0136] Example 7 may include the method according to example 1 or some other example herein, wherein ML entity loading is to load a trained ML entity to the target Al / ML inference function which will use it for inference.
[0137] Example 8 may include the method according to example 1 or some other example herein, wherein Al / ML inference phase is to perform inference using the ML entity by the Al / ML inference function.
[0138] Example 9 may include the method according to example 1 or some other example herein, wherein each phase is managed by the respective management capabilities.
[0139] Example 10 may include the method according to example 1 or some other example herein, wherein the ML emulation phase may be skipped.
[0140] Example 11 may include the method according to example 1 or some other example herein, wherein the ML entity loading phase may be skipped.
[0141] Example 12 includes a method to be performed by an electronic device, wherein the method comprises: training a machine learning (ML) model that is related to a fifth generation (5G) network to generate a trained ML model; loading the trained ML model to a ML inference function; and performing inference of the trained ML model by the ML inference function.
[0142] Example 13 includes the method of example 2, and / or some other example herein, wherein the method further comprises emulating the trained ML model in a network environment that is similar to that of the 5G network in which the ML model will be implemented.
[0143] Example 14 includes a method to be performed by a machine learning (ML) management system of a cellular network, wherein the method comprises: training or facilitating training of a ML model by a ML training module, wherein training of the ML model generates a trained ML model, and wherein the ML model is related to a cellular network system; testing or facilitating testing of the trained ML model by a ML testing module, wherein testing of the ML trained ML module includes evaluation of application of the trained ML module to testing data; emulating or facilitating emulation of the trained ML model by a ML emulation module, wherein emulation of the ML model involves performing inference using the ML model in an emulation environment that is different than an inference environment of the cellular network; and identifying or facilitating identifying, based on the emulation of the trained ML model, whether to re-train the ML model by the ML training module or provide the trained ML model to a ML inference module of the cellular network for inference in the inference environment.
[0144] Example 15 includes the method of example 14, and / or some other example herein, wherein the inference environment is an environment in which the ML model is applied to a real-world (e.g., “live”) environment of the cellular network system.
[0145] Example 16 includes the method of any of examples 14-15, and / or some other example herein, wherein the cellular network system is a fifth generation core (5GC) system.
[0146] Example 17 includes the method of any of examples 14-15, and / or some other example herein, wherein the cellular network system is a Next Generation Radio Access Network (NG-RAN).
[0147] Example 18 includes the method of any of examples 14-15, and / or some other example herein, wherein the cellular network system is a management system of the 5G network.
[0148] Example 19 includes the method of any of examples 14-18, and / or some other example herein, wherein provision of the trained ML model to the ML inference module includes loading the trained ML model by a ML loading module.
[0149] Example 20 includes the method of any of examples 14-19, and / or some other example herein, wherein emulation of the trained ML model is performed based on a result of the evaluation of application of the trained ML module to the testing data.
[0150] Example 21 includes the method of any of examples 14-20, and / or some other example herein, wherein the emulation of the trained ML model is based on a request from a consumer of the ML model.
[0151] Example 22 includes the method of any of examples 14-21, and / or some other example herein, wherein the emulation of the trained ML model is based on a request from a producer of the ML model. Example 23 includes the method of any of examples 14-22, and / or some other example herein, wherein the ML management system is further configured to perform or facilitate performance of inference based on the trained ML model.
[0152] Example Z01 may include an apparatus comprising means to perform one or more elements of a method described in or related to any of examples 1-23, or any other method or process described herein.
[0153] Example Z02 may include one or more non-transitory computer-readable media comprising instructions to cause an electronic device, upon execution of the instructions by one or more processors of the electronic device, to perform one or more elements of a method described in or related to any of examples 1-23, or any other method or process described herein.
[0154] Example Z03 may include an apparatus comprising logic, modules, or circuitry to perform one or more elements of a method described in or related to any of examples 1- 23, or any other method or process described herein.
[0155] Example Z04 may include a method, technique, or process as described in or related to any of examples 1-23, or portions or parts thereof.
[0156] Example Z05 may include an apparatus comprising: one or more processors and one or more computer-readable media comprising instructions that, when executed by the one or more processors, cause the one or more processors to perform the method, techniques, or process as described in or related to any of examples 1-23, or portions thereof.
[0157] Example Z06 may include a signal as described in or related to any of examples 1- 23, or portions or parts thereof.
[0158] Example Z07 may include a datagram, packet, frame, segment, protocol data unit (PDU), or message as described in or related to any of examples 1-23, or portions or parts thereof, or otherwise described in the present disclosure.
[0159] Example Z08 may include a signal encoded with data as described in or related to any of examples 1-23, or portions or parts thereof, or otherwise described in the present disclosure.
[0160] Example Z09 may include a signal encoded with a datagram, packet, frame, segment, protocol data unit (PDU), or message as described in or related to any of examples 1-23, or portions or parts thereof, or otherwise described in the present disclosure. Example Z10 may include an electromagnetic signal carrying computer-readable instructions, wherein execution of the computer-readable instructions by one or more processors is to cause the one or more processors to perform the method, techniques, or process as described in or related to any of examples 1-23, or portions thereof.
[0161] Example Zll may include a computer program comprising instructions, wherein execution of the program by a processing element is to cause the processing element to carry out the method, techniques, or process as described in or related to any of examples 1-23, or portions thereof.
[0162] Example Z12 may include a signal in a wireless network as shown and described herein.
[0163] Example Z13 may include a method of communicating in a wireless network as shown and described herein.
[0164] Example Z14 may include a system for providing wireless communication as shown and described herein.
[0165] Example Z15 may include a device for providing wireless communication as shown and described herein.
[0166] Any of the above-described examples may be combined with any other example (or combination of examples), unless explicitly stated otherwise. The foregoing description of one or more implementations provides illustration and description, but is not intended to be exhaustive or to limit the scope of embodiments to the precise form disclosed. Modifications and variations are possible in light of the above teachings or may be acquired from practice of various embodiments.
[0167] Abbreviations
[0168] Unless used differently herein, terms, definitions, and abbreviations may be consistent with terms, definitions, and abbreviations defined in 3GPP TR 21.905 V16.0.0 (2019-06). For the purposes of the present document, the following abbreviations may apply to the examples and embodiments discussed herein. 3GPP Third 35 AN Access BCH Broadcast Generation Network Channel
[0169] Partnership AnLF Analytics 70 BER Bit Error Ratio Project Logical Function BFD Beam
[0170] 4G Fourth ANR Automatic Failure Detection Generation 40 Neighbour Relation BLER Block Error
[0171] 5G Fifth AOA Angle of Rate
[0172] Generation Arrival 75 BPSK Binary Phase
[0173] 5GC 5G Core AP Application Shift Keying network Protocol, Antenna BRAS Broadband AC 45 Port, Access Point Remote Access
[0174] Application API Application Server Client Programming 80 BSS Business
[0175] ACR Application Interface Support System Context Relocation APN Access Point BS Base Station ACK 50 Name BSR Buffer Status
[0176] Acknowledgem ARP Allocation and Report ent Retention Priority 85 BW Bandwidth ACID ARQ Automatic BWP Bandwidth Part
[0177] Application Repeat Request C-RNT1 Cell Client Identification 55 AS Access Stratum Radio Network ADRF Analytics Data ASP Temporary
[0178] Repository Application Service 90 Identity
[0179] Function Provider CA Carrier
[0180] AF Application Aggregation, Function 60 ASN.l Abstract Syntax Certification
[0181] AM Acknowledged Notation One Authority Mode AUSF Authentication 95 CAPEX CAPital
[0182] AMBR Aggregate Server Function Expenditure Maximum Bit Rate AWGN Additive White CBD Candidate AMF Access and 65 Gaussian Noise Beam Detection
[0183] Mobility BAP Backhaul CBRA Contention
[0184] Management Adaptation Protocol 100 Based Random Function Access CC Component 35 CK Cipher Key C / R Carrier, Country CM Connection 70 Command / Res Code, Cryptographic Management, ponse field bit Checksum Conditional CRAN Cloud Radio CCA Clear Channel Mandatory Access Assessment 40 CMAS Commercial Network, Cloud CCE Control Channel Mobile Alert Service 75 RAN Element CMD Command CRB Common CCCH Common CMS Cloud Resource Block Control Channel Management System CRC Cyclic CE Coverage 45 CO Conditional Redundancy Check Enhancement Optional 80 CR1 Channel-State CDM Content CoMP Coordinated Information Delivery Network Multi-Point Resource CDMA Code-Division CORESET Control Indicator, CS1-RS Multiple Access 50 Resource Set Resource CDR Charging Data COTS Commercial 85 Indicator Request Off-The-Shelf C-RNT1 Cell
[0185] CDR Charging Data CP Control Plane, RNT1 Response Cyclic Prefix, CS Circuit
[0186] CFRA Contention Free 55 Connection Switched Random Access Point 90 CSCF call CG Cell Group CPD Connection session control CGF Charging Point Descriptor function
[0187] Gateway Function CPE Customer CSAR Cloud Service CHF Charging 60 Premise Archive
[0188] Function Equipment 95 CS1 Channel-State
[0189] Cl Cell Identity CP1CH Common Pilot Information
[0190] CID Cell-ID (e.g., Channel CS1-1M CS1 Interference positioning method) CQ1 Channel Quality Measurement C1M Common 65 Indicator CS1-RS CS1 Reference Information Model CPU CS1 processing 100 Signal C1R Carrier to unit, Central Interference Ratio Processing Unit CSI-RSRP CSI DF Deployment EAS Edge reference signal 35 Flavour Application Server received power DL Downlink 70 ECCA extended clear
[0191] CSI-RSRQ CSI DMTF Distributed channel reference signal Management Task assessment, received quality Force extended CCA
[0192] CS1-S1NR CSI 40 DPDK Data Plane ECCE Enhanced signal-to-noise and Development Kit 75 Control Channel interference DM-RS, DMRS Element, ratio Demodulation Enhanced CCE CSMA Carrier Sense Reference Signal ED Energy Multiple Access 45 DN Data network Detection
[0193] CSMA / CA CSMA DNN Data Network 80 EDGE Enhanced with collision Name Datarates for GSM avoidance DNA1 Data Network Evolution (GSM CSS Common Access Identifier Evolution) Search Space, Cell50 EAS Edge specific Search DRB Data Radio 85 Application Server
[0194] Space Bearer EAS1D Edge
[0195] CTF Charging DRS Discovery Application Server Trigger Function Reference Signal Identification CTS Clear-to-Send 55 DRX Discontinuous ECS Edge CW Codeword Reception 90 Configuration Server CWS Contention DSL Domain Specific ECSP Edge Window Size Language. Digital Computing Service D2D Device-to- Subscriber Line Provider Device 60 DSLAM DSL EDN Edge DC Dual Access Multiplexer 95 Data Network Connectivity, Direct DwPTS EEC Edge Current Downlink Pilot Enabler Client
[0196] DC1 Downlink Time Slot EEC1D Edge Control 65 E-LAN Ethernet Local Enabler Client
[0197] Information Area Network 100 Identification
[0198] E2E End-to-End EES Edge 35 Physical FACCHFast Associated
[0199] Enabler Server Downlink Control Control CHannel EESID Edge Cannel 70 FACCH / F Fast
[0200] Enabler Server EPRE Energy per Associated Control Identification resource element Channel / Full
[0201] EHE Edge 40 EPS Evolved Packet rate
[0202] Hosting Environment System FACCH / H Fast EGMF Exposure EREG enhanced REG, 75 Associated Control Governance enhanced resource Channel / Half Management element groups rate
[0203] Function 45 ETS1 European FACH Forward Access
[0204] EGPRS Enhanced GPRS Telecommunica Channel
[0205] E1R Equipment tions Standards 80 FAUSCH Fast Identity Register Institute Uplink Signalling eLAA enhanced ETWS Earthquake and Channel Licensed Assisted 50 Tsunami Warning FB Functional Access, System Block enhanced LAA eUlCC embedded 85 FBI Feedback EM Element U1CC, embedded Information Manager Universal FCC Federal eMBB Enhanced 55 Integrated Circuit Communications Mobile Card Commission
[0206] Broadband E-UTRA Evolved 90 FCCH Frequency
[0207] EMS Element UTRA Correction CHannel Management System E-UTRAN Evolved FDD Frequency eNB evolved NodeB, 60 UTRAN Division Duplex E-UTRAN Node B EV2X Enhanced V2X FDM Frequency EN-DC E-UTRA-NR F1AP Fl Application 95 Division Dual Connectivity Protocol Multiplex EPC Evolved Packet Fl-C Fl Control FDMA Frequency Core 65 plane interface Division Multiple EPDCCH Fl-U Fl User plane Access enhanced interface 100 FE Front End PDCCH, enhanced FEC Forward Error 35 Global Navigation GTS Go To Sleep Correction Satellite Signal (related to
[0208] FFS For Further System) 70 WUS)
[0209] Study gNB Next GUMME1 Globally
[0210] FFT Fast Fourier Generation NodeB Unique MME
[0211] Transformation 40 gNB-CU gNB- Identifier feLAA further centralized unit, Next GUT1 Globally Unique enhanced Licensed Generation 75 Temporary UE
[0212] Assisted Access, NodeB Identity further centralized unit HARQ Hybrid ARQ, enhanced LAA 45 gNB-DU gNB- Hybrid
[0213] FN Frame Number distributed unit, Next Automatic
[0214] FPGA Field- Generation 80 Repeat Request
[0215] Programmable Gate NodeB HANDO Handover
[0216] Array distributed unit HFN HyperFrame
[0217] FR Frequency 50 GNSS Global Number
[0218] Range Navigation Satellite HHO Hard Handover
[0219] FQDN Fully Qualified System 85 HLR Home Location Domain Name GPRS General Packet Register
[0220] G-RNT1 GERAN Radio Service HN Home Network
[0221] Radio Network 55 GPS1 Generic HO Handover
[0222] Temporary Public Subscription HPLMN Home Identity Identifier 90 Public Land Mobile GERAN GSM Global System Network
[0223] GSM EDGE RAN, for Mobile HSDPA High GSM EDGE Radio 60 Communication Speed Downlink Access Network s, Groupe Special Packet Access GGSN Gateway GPRS Mobile 95 HSN Hopping Support Node GTP GPRS Tunneling Sequence Number GLONASS Protocol HSPA High speed
[0224] GLObal'naya 65 GTP-U GPRS Packet Access
[0225] NAvigatsionnay Tunnelling Protocol HSS Home a Sputnikovaya for User Plane 100 Subscriber Server Sistema [EngL: HSUPA High 35 1E1 Information loT Internet of
[0226] Speed Uplink Packet Element 70 Things Access Identifier IP Internet
[0227] HTTP Hyper Text 1E1DL Information Protocol Transfer Protocol Element Ipsec IP Security,
[0228] HTTPSHyper Text 40 Identifier Data Internet Protocol Transfer Protocol Length 75 Security
[0229] Secure (https is IETF Internet IP-CAN 1P- http / 1.1 over Engineering Task Connectivity Access
[0230] SSL, i.e. port 443) Force Network 1-Block 45 IF Infrastructure 1P-M IP Multicast
[0231] Information HOT Industrial 80 IPv4 Internet
[0232] Block Internet of Things Protocol Version 4
[0233] 1CC1D Integrated IM Interference IPv6 Internet Circuit Card Measurement, Protocol Version 6
[0234] Identification 50 Intermodulatio 1R Infrared 1AB Integrated n, IP Multimedia 85 IS In Sync Access and Backhaul 1MC IMS Credentials 1RP Integration 1C1C Inter-Cell 1ME1 International Reference Point Interference Mobile ISDN Integrated
[0235] Coordination 55 Equipment Services Digital
[0236] ID Identity, Identity 90 Network identifier 1MG1 International 1S1M IM Services 1DFT Inverse mobile group identity Identity Module Discrete Fourier 1MP1 IP Multimedia ISO International
[0237] Transform 60 Private Identity Organisation for
[0238] IE Information 1MPU IP Multimedia 95 Standardisation element PUblic identity ISP Internet Service 1BE In-Band IMS IP Multimedia Provider Emission Subsystem 1WF Interworking-
[0239] IEEE Institute of 65 1MS1 International Function
[0240] Electrical and Mobile 100 1-WLAN
[0241] Electronics Subscriber Interworking Engineers Identity WLAN Constraint 35 LAA Licensed LW1P LTE / WLAN length of the Assisted Access Radio Level convolutional LAN Local Area 70 Integration code, US1M Network with IPsec Tunnel
[0242] Individual key LADN Local LTE Long Term
[0243] 40 Area Data Network Evolution kB Kilobyte (1000 LBT Listen Before M2M Machine-to- bytes) Talk 75 Machine kbps kilo-bits per LCM LifeCycle MAC Medium Access second Management Control
[0244] Kc Ciphering key 45 LCR Low Chip Rate (protocol
[0245] Ki Individual LCS Location layering context) subscriber Services 80 MAC Message authentication LC1D Logical authentication key Channel ID code
[0246] KPI Key 50 LI Layer Indicator (security / encryption
[0247] Performance Indicator LLC Logical Link context)
[0248] KQ1 Key Quality Control, Low Layer 85 MAC-AMAC used for Indicator Compatibility authentication
[0249] KS1 Key Set LMF Location and key Identifier 55 Management Function agreement (TSG ksps kilo-symbols LOS Line of T WG3 context) per second Sight 90 MAC-1 MAC used for
[0250] KVM Kernel Virtual LPLMN Local data integrity of
[0251] Machine PLMN signalling messages
[0252] LI Layer 1 60 LPP LTE Positioning (TSG T WG3 context) (physical layer) Protocol MANO Management Ll-RSRP Layer 1 LSB Least 95 and Orchestration reference signal Significant Bit MBMS Multimedia received power LTE Long Term Broadcast and
[0253] L2 Layer 2 (data 65 Evolution Multicast link layer) LWA LTE-WLAN Service
[0254] L3 Layer 3 aggregation 100 MBSFN
[0255] (network layer) Multimedia Broadcast M1B Master MPLS MultiProtocol multicast 35 Information Block, Label Switching service Single Management 70 MS Mobile Station
[0256] Frequency Information Base MSB Most Significant
[0257] Network M1M0 Multiple Input Bit
[0258] MCC Mobile Country Multiple Output MSC Mobile
[0259] Code 40 MLC Mobile Location Switching Centre
[0260] MCG Master Cell Centre 75 MSI Minimum
[0261] Group MM Mobility System
[0262] MCOT Maximum Management Information,
[0263] Channel MME Mobility MCH Scheduling
[0264] Occupancy 45 Management Entity Information
[0265] Time MN Master Node 80 MS1D Mobile Station
[0266] MCS Modulation and MNO Mobile Identifier coding scheme Network Operator MS1N Mobile Station
[0267] MDAF Management MO Measurement Identification
[0268] Data Analytics 50 Object, Mobile Number
[0269] Function Originated 85 MS1SDN Mobile
[0270] MDAS Management MPBCH MTC Subscriber ISDN
[0271] Data Analytics Physical Broadcast Number
[0272] Service CHannel MT Mobile
[0273] MDT Minimization of 55 MPDCCH MTC Terminated, Mobile
[0274] Drive Tests Physical Downlink 90 Termination
[0275] ME Mobile Control MTC Machine-Type
[0276] Equipment CHannel Communication
[0277] MeNB master eNB MPDSCH MTC s
[0278] MER Message Error 60 Physical Downlink MTLF Model Training
[0279] Ratio Shared CHannel 95 Logical
[0280] MGL Measurement MPRACH MTC Functions
[0281] Gap Length Physical Random mMTC massive MTC,
[0282] MGRP Measurement Access CHannel massive
[0283] Gap Repetition 65 MPUSCH MTC Machine-Type
[0284] Period Physical Uplink Shared 100 Communication
[0285] Channel s MU-MIMO Multi 35 NFV1 NFV Physical Random User MIMO Infrastructure 70 Access CHannel MWUSMTC wake-up NFVO NFV NPUSCH signal, MTC WUS Orchestrator Narrowband NACK Negative NG Next Physical Uplink Acknowledgement 40 Generation, Next Gen Shared CHannel NA1 Network Access NGEN-DC NG-RAN 75 NPSS Narrowband Identifier E-UTRA-NR Dual Primary
[0286] NAS Non-Access Connectivity Synchronizatio
[0287] Stratum, Non-Access NM Network n Signal Stratum layer 45 Manager NSSS Narrowband NCT Network NMS Network 80 Secondary Connectivity Management System Synchronizatio
[0288] Topology N-PoP Network Point n Signal NC-JT Nonof Presence NR New Radio, coherent Joint 50 NM1B, N-M1B Neighbour Relation
[0289] Transmission Narrowband M1B 85 NRF NF Repository
[0290] NEC Network NPBCH Function Capability Narrowband NRS Narrowband
[0291] Exposure Physical Reference Signal NE-DC NR-E-UTRA 55 Broadcast NS Network Dual Connectivity CHannel 90 Service NEF Network NPDCCH NSA Non-Standalone Exposure Function Narrowband operation mode NF Network Physical NSD Network Function 60 Downlink Service Descriptor
[0292] NFP Network Control CHannel 95 NSR Network Forwarding Path NPDSCH Service Record NFPD Network Narrowband NSSA1 Network Slice Forwarding Path Physical Selection
[0293] Descriptor 65 Downlink Assistance NFV Network Shared CHannel 100 Information Functions NPRACH S-NNSA1 Single-
[0294] Virtualization Narrowband NSSA1 NSSF N etwork Slice 35 PAR Peak to Average PDCP Packet Data Selection Function Ratio Convergence Protocol
[0295] NW Network PBCH Physical 70 PDN Packet Data NWDAF Network Broadcast Channel Network, Public
[0296] Data Analytics PC Power Control, Data Network Function 40 Personal PDSCHPhysical NWUS Narrowband Computer Downlink Shared wake-up signal, PCC Primary 75 Channel Narrowband WUS Component Carrier, PDU Protocol Data NZP Non-Zero Primary CC Unit Power 45 P-CSCF Proxy PEI Permanent O&M Operation and CSCF Equipment Maintenance PCell Primary Cell 80 Identifiers ODU2 Optical channel PCI Physical Cell ID, PFD Packet Flow Data Unit - type 2 Physical Cell Description OFDM Orthogonal 50 Identity P-GW PDN Gateway Frequency Division PCEF Policy and PH1CH Physical hybrid- Multiplexing Charging 85 ARQ indicator OFDMA Enforcement channel
[0297] Orthogonal Function PHY Physical layer Frequency Division 55 PCF Policy Control PLMN Public Land Multiple Access Function Mobile Network OOB Out-of-band PCRF Policy Control 90 PIN Personal OOS Out of Sync and Charging Rules Identification Number OPEX Operating Function PM Performance EXpense 60 PDCP Packet Data Measurement OS1 Other System Convergence Protocol, PM1 Precoding Information Packet Data 95 Matrix Indicator OSS Operations Convergence PNF Physical Support System Protocol layer Network Function OTA over-the-air 65 PDCCH Physical PNFD Physical PAPR Peak-to- Downlink Control Network Function Average Power Channel 100 Descriptor Ratio PNFR Physical 35 PSSCH Physical QPSK Quadrature
[0298] Network Function Sidelink Shared (Quaternary) Phase
[0299] Record Channel Shift Keying
[0300] POC PTT over PSFCH physical 70 4SS Quasi-Zenith
[0301] Cellular sidelink feedback Satellite System
[0302] PP, PTP Point-to- 40 channel RA-RNT1 Random
[0303] Point PSCell Primary SCell Access RNT1
[0304] PPP Point-to-Point PSS Primary RAB Radio Access
[0305] Protocol Synchronization 75 Bearer, Random
[0306] PRACH Physical Signal Access Burst
[0307] RACH 45 PSTN Public Switched RACH Random Access
[0308] PRE Physical Telephone Network Channel resource block PT-RS Phase-tracking RADIUS Remote
[0309] PRG Physical reference signal 80 Authentication Dial In resource block PTT Push-to-Talk User Service group 50 PUCCH Physical RAN Radio Access
[0310] ProSe Proximity Uplink Control Network
[0311] Services, Channel RAND RANDom
[0312] Proximity- PUSCHPhysical Uplink 85 number (used for
[0313] Based Service Shared Channel authentication)
[0314] PRS Positioning 55 QAM Quadrature RAR Random Access
[0315] Reference Signal Amplitude Response
[0316] PRR Packet Modulation RAT Radio Access
[0317] Reception Radio QC1 QoS class of 90 Technology
[0318] PS Packet Services identifier RAU Routing Area
[0319] PSBCH Physical 60 QCL Quasi coUpdate
[0320] Sidelink Broadcast location RB Resource block,
[0321] Channel QF1 QoS Flow ID, Radio Bearer
[0322] PSDCHPhysical QoS Flow 95 RBG Resource block
[0323] Sidelink Downlink Identifier group
[0324] Channel 65 QoS Quality of REG Resource
[0325] PSCCH Physical Service Element Group
[0326] Sidelink Control Rel Release
[0327] Channel 100 REQ REQuest RF Radio RNC Radio Network Rx Reception, Frequency 35 Controller Receiving, Receiver RI Rank Indicator RNL Radio Network 70 S1AP SI Application
[0328] RIV Resource Layer Protocol indicator value RNTI Radio Network Sl-MME SI for RL Radio Link Temporary the control plane
[0329] RLC Radio Link 40 Identifier Sl-U SI for the user
[0330] Control, Radio ROHC RObust Header 75 plane
[0331] Link Control Compression S-CSCF serving layer RRC Radio Resource CSCF
[0332] RLC AM RLC Control, Radio S-GW Serving Acknowledged Mode 45 Resource Control Gateway
[0333] RLC UM RLC layer 80 S-RNT1 SRNC
[0334] Unacknowledged RRM Radio Resource Radio Network
[0335] Mode Management Temporary
[0336] RLF Radio Link RS Reference Identity
[0337] Failure 50 Signal S-TMS1 SAE
[0338] RLM Radio Link RSRP Reference 85 Temporary Mobile
[0339] Monitoring Signal Received Station
[0340] RLM-RS Power Identifier
[0341] Reference RSRQ Reference SA Standalone
[0342] Signal for RLM 55 Signal Received operation mode
[0343] RM Registration Quality 90 SAE System Management RSSI Received Signal Architecture
[0344] RMC Reference Strength Indicator Evolution
[0345] Measurement RSU Road Side Unit SAP Service Access
[0346] Channel 60 RSTD Reference Point
[0347] RMSI Remaining MSI, Signal Time 95 SAPD Service Access Remaining difference Point Descriptor
[0348] Minimum RTP Real Time SAPI Service Access
[0349] System Protocol Point Identifier
[0350] Information 65 RTS Ready-To-Send SCC Secondary
[0351] RN Relay Node RTT Round Trip 100 Component Carrier, Time Secondary CC SCell Secondary Cell 35 SDU Service Data SLA Service Level SCEF Service Unit 70 Agreement Capability Exposure SEAF Security Anchor SM Session Function Function Management
[0352] SC-FDMA Single SeNB secondary eNB SMF Session Carrier Frequency 40 SEPP Security Edge Management Function
[0353] Division Protection Proxy 75 SMS Short Message Multiple Access SF1 Slot format Service SCG Secondary Cell indication SMSF SMS Function Group SFTD Space- SMTC SSB-based SCM Security 45 Frequency Time Measurement Context Diversity, SFN 80 Timing
[0354] Management and frame timing Configuration SCS Subcarrier difference SN Secondary Spacing SFN System Frame Node, Sequence SCTP Stream Control 50 Number Number
[0355] Transmission SgNB Secondary gNB 85 SoC System on Chip
[0356] Protocol SGSN Serving GPRS SON Self-Organizing
[0357] SDAP Service Data Support Node Network Adaptation Protocol, S-GW Serving SpCell Special Cell Service Data 55 Gateway SP-CSl-RNTlSemi-
[0358] Adaptation SI System 90 Persistent CS1 RNT1 Protocol layer Information SPS Semi-Persistent SDL Supplementary S1-RNT1 System Scheduling Downlink Information RNT1 SQN Sequence
[0359] SDNF Structured Data 60 SIB System number Storage Network Information Block 95 SR Scheduling Function SIM Subscriber Request
[0360] SDP Session Identity Module SRB Signalling Radio
[0361] Description Protocol SIP Session Bearer SDSF Structured Data 65 Initiated Protocol SRS Sounding Storage Function SiP System in 100 Reference Signal SDT Small Data Package
[0362] Transmission SL Sidelink ss 35 and Interference TCP Transmission
[0363] Synchronizatio Ratio Communication n Signal SSS Secondary 70 Protocol SSB Synchronization TDD Time Division
[0364] Synchronizatio Signal Duplex n Signal Block 40 SSSG Search Space TDM Time Division SSID Service Set Set Group Multiplexing Identifier SSS1F Search Space 75 TDMA Time Division
[0365] SS / PBCH Block Set Indicator Multiple Access SSBR1 SS / PBCH Block SST Slice / Service TE Terminal Resource 45 Types Equipment
[0366] Indicator, SU-M1M0 Single TE1D Tunnel End Synchronization User MIMO 80 Point Identifier
[0367] Signal Block SUL Supplementary TFT Traffic Flow Resource Indicator Uplink Template SSC Session and 50 TA Timing TMS1 Temporary Service Advance, Tracking Mobile
[0368] Continuity Area 85 Subscriber SS-RSRP TAC Tracking Area Identity
[0369] Synchronizatio Code TNL Transport n Signal based 55 TAG Timing Advance Network Layer Reference Group TPC Transmit Power Signal Received TAI Tracking 90 Control Power Area Identity TPM1 Transmitted
[0370] SS-RSRQ TAU Tracking Area Precoding Matrix
[0371] Synchronizatio 60 Update Indicator n Signal based TB Transport Block TR Technical Reference TBS Transport Block 95 Report Signal Received Size TRP, TRxP Quality TBD To Be Defined Transmission
[0372] SS-S1NR 65 TCI Transmission Reception Point
[0373] Synchronizatio Configuration TRS Tracking n Signal based Indicator 100 Reference Signal Signal to Noise TRx Transceiver TS Technical 35 UML Unified V2P Vehicle-to-
[0374] Specifications, Modelling Language Pedestrian
[0375] Technical UMTS Universal 70 V2V Vehicle-to-
[0376] Standard Mobile Vehicle
[0377] TT1 Transmission Telecommunica V2X Vehicle-to-
[0378] Time Interval 40 tions System everything
[0379] Tx Transmission, UP User Plane VIM Virtualized
[0380] Transmitting, UPF User Plane 75 Infrastructure Transmitter Function Manager
[0381] U-RNT1 UTRAN UR1 Uniform VL Virtual Link,
[0382] Radio Network 45 Resource Identifier VLAN Virtual LAN,
[0383] Temporary URL Uniform Virtual Local Area Identity Resource Locator 80 Network
[0384] UART Universal URLLC Ultra-Reliable VM Virtual Machine
[0385] Asynchronous and Low Latency VNF Virtualized
[0386] Receiver and 50 USB Universal Serial Network Function
[0387] Transmitter Bus VNFFG VNF
[0388] UC1 Uplink Control US1M Universal 85 Forwarding Graph Information Subscriber Identity VNFFGD VNF
[0389] UE User Equipment Module Forwarding Graph UDM Unified Data 55 USS UE-specific Descriptor
[0390] Management search space VNFM VNF Manager
[0391] UDP User Datagram UTRA UMTS 90 VoIP Voice-over-lP, Protocol Terrestrial Radio Voice-over- Internet
[0392] UDSF Unstructured Access Protocol
[0393] Data Storage Network 60 UTRAN VPLMN Visited Function Universal Public Land Mobile
[0394] U1CC Universal Terrestrial Radio 95 Network
[0395] Integrated Circuit Access Network VPN Virtual Private
[0396] Card UwPTS Uplink Network
[0397] UL Uplink 65 Pilot Time Slot VRB Virtual
[0398] UM V21 Vehicle-to- Resource Block
[0399] Unacknowledge Infrastruction 100 WiMAX d Mode Worldwide Interoperability for Microwave
[0400] Access
[0401] WLAN Wireless Local
[0402] Area Network
[0403] WMAN Wireless
[0404] Metropolitan Area
[0405] Network
[0406] WPAN Wireless
[0407] Personal Area
[0408] Network
[0409] X2-C X2-Control plane
[0410] X2-U X2-User plane XML extensible Markup
[0411] Language
[0412] XRES EXpected user RESponse
[0413] XOR exclusive OR ZC Zadoff-Chu ZP Zero Power
[0414] Terminology
[0415] For the purposes of the present document, the following terms and definitions are applicable to the examples and embodiments discussed herein.
[0416] The term “application” may refer to a complete and deployable package, environment to achieve a certain function in an operational environment. The term “Al / ML application” or the like may be an application that contains some Al / ML models and application-level descriptions.
[0417] The term “circuitry” as used herein refers to, is part of, or includes hardware components such as an electronic circuit, a logic circuit, a processor (shared, dedicated, or group) and / or memory (shared, dedicated, or group), an Application Specific Integrated Circuit (ASIC), a field-programmable device (FPD) (e.g., a field-programmable gate array (FPGA), a programmable logic device (PLD), a complex PLD (CPLD), a high-capacity PLD (HCPLD), a structured ASIC, or a programmable SoC), digital signal processors (DSPs), etc., that are configured to provide the described functionality. In some embodiments, the circuitry may execute one or more software or firmware programs to provide at least some of the described functionality. The term “circuitry” may also refer to a combination of one or more hardware elements (or a combination of circuits used in an electrical or electronic system) with the program code used to carry out the functionality of that program code. In these embodiments, the combination of hardware elements and program code may be referred to as a particular type of circuitry.
[0418] The term “processor circuitry” as used herein refers to, is part of, or includes circuitry capable of sequentially and automatically carrying out a sequence of arithmetic or logical operations, or recording, storing, and / or transferring digital data. Processing circuitry may include one or more processing cores to execute instructions and one or more memory structures to store program and data information. The term “processor circuitry” may refer to one or more application processors, one or more baseband processors, a physical central processing unit (CPU), a single-core processor, a dual-core processor, a triple-core processor, a quad-core processor, and / or any other device capable of executing or otherwise operating computer-executable instructions, such as program code, software modules, and / or functional processes. Processing circuitry may include more hardware accelerators, which may be microprocessors, programmable processing devices, or the like. The one or more hardware accelerators may include, for example, computer vision (CV) and / or deep learning (DL) accelerators. The terms “application circuitry” and / or “baseband circuitry” may be considered synonymous to, and may be referred to as, “processor circuitry.”
[0419] The term “interface circuitry” as used herein refers to, is part of, or includes circuitry that enables the exchange of information between two or more components or devices. The term “interface circuitry” may refer to one or more hardware interfaces, for example, buses, I / O interfaces, peripheral component interfaces, network interface cards, and / or the like.
[0420] The term “user equipment” or “UE” as used herein refers to a device with radio communication capabilities and may describe a remote user of network resources in a communications network. The term “user equipment” or “UE” may be considered synonymous to, and may be referred to as, client, mobile, mobile device, mobile terminal, user terminal, mobile unit, mobile station, mobile user, subscriber, user, remote station, access agent, user agent, receiver, radio equipment, reconfigurable radio equipment, reconfigurable mobile device, etc. Furthermore, the term “user equipment” or “UE” may include any type of wireless / wired device or any computing device including a wireless communications interface.
[0421] The term “network element” as used herein refers to physical or virtualized equipment and / or infrastructure used to provide wired or wireless communication network services. The term “network element” may be considered synonymous to and / or referred to as a networked computer, networking hardware, network equipment, network node, router, switch, hub, bridge, radio network controller, RAN device, RAN node, gateway, server, virtualized VNF, NFV1, and / or the like.
[0422] The term “computer system” as used herein refers to any type interconnected electronic devices, computer devices, or components thereof. Additionally, the term “computer system” and / or “system” may refer to various components of a computer that are communicatively coupled with one another. Furthermore, the term “computer system” and / or “system” may refer to multiple computer devices and / or multiple computing systems that are communicatively coupled with one another and configured to share computing and / or networking resources.
[0423] The term “appliance,” “computer appliance,” or the like, as used herein refers to a computer device or computer system with program code (e.g., software or firmware) that is specifically designed to provide a specific computing resource. A “virtual appliance” is a virtual machine image to be implemented by a hypervisor-equipped device that virtualizes or emulates a computer appliance or otherwise is dedicated to provide a specific computing resource.
[0424] The term “resource” as used herein refers to a physical or virtual device, a physical or virtual component within a computing environment, and / or a physical or virtual component within a particular device, such as computer devices, mechanical devices, memory space, processor / CPU time, processor / CPU usage, processor and accelerator loads, hardware time or usage, electrical power, input / output operations, ports or network sockets, channel / link allocation, throughput, memory usage, storage, network, database and applications, workload units, and / or the like. A “hardware resource” may refer to compute, storage, and / or network resources provided by physical hardware element(s). A “virtualized resource” may refer to compute, storage, and / or network resources provided by virtualization infrastructure to an application, device, system, etc. The term “network resource” or “communication resource” may refer to resources that are accessible by computer devices / systems via a communications network. The term “system resources” may refer to any kind of shared entities to provide services, and may include computing and / or network resources. System resources may be considered as a set of coherent functions, network data objects or services, accessible through a server where such system resources reside on a single host or multiple hosts and are clearly identifiable.
[0425] The term “channel” as used herein refers to any transmission medium, either tangible or intangible, which is used to communicate data or a data stream. The term “channel” may be synonymous with and / or equivalent to “communications channel,” “data communications channel,” “transmission channel,” “data transmission channel,” “access channel,” “data access channel,” “link,” “data link,” “carrier,” “radiofrequency carrier,” and / or any other like term denoting a pathway or medium through which data is communicated. Additionally, the term “link” as used herein refers to a connection between two devices through a RAT for the purpose of transmitting and receiving information.
[0426] The terms “instantiate,” “instantiation,” and the like as used herein refers to the creation of an instance. An “instance” also refers to a concrete occurrence of an object, which may occur, for example, during execution of program code.
[0427] The terms “coupled,” “communicatively coupled,” along with derivatives thereof are used herein. The term “coupled” may mean two or more elements are in direct physical or electrical contact with one another, may mean that two or more elements indirectly contact each other but still cooperate or interact with each other, and / or may mean that one or more other elements are coupled or connected between the elements that are said to be coupled with each other. The term “directly coupled” may mean that two or more elements are in direct contact with one another. The term “communicatively coupled” may mean that two or more elements may be in contact with one another by a means of communication including through a wire or other interconnect connection, through a wireless communication channel or link, and / or the like.
[0428] The term “information element” refers to a structural element containing one or more fields. The term “field” refers to individual contents of an information element, or a data element that contains content.
[0429] The term “SMTC” refers to an SSB-based measurement timing configuration configured by SSB-MeasurementTimingConfiguration.
[0430] The term “SSB” refers to an SS / PBCH block.
[0431] The term “a “Primary Cell” refers to the MCG cell, operating on the primary frequency, in which the UE either performs the initial connection establishment procedure or initiates the connection re-establishment procedure.
[0432] The term “Primary SCG Cell” refers to the SCG cell in which the UE performs random access when performing the Reconfiguration with Sync procedure for DC operation.
[0433] The term “Secondary Cell” refers to a cell providing additional radio resources on top of a Special Cell for a UE configured with CA.
[0434] The term “Secondary Cell Group” refers to the subset of serving cells comprising the PSCell and zero or more secondary cells for a UE configured with DC.
[0435] The term “Serving Cell” refers to the primary cell for a UE in RRC_CONNECTED not configured with CA / DC there is only one serving cell comprising of the primary cell.
[0436] The term “serving cell” or “serving cells” refers to the set of cells comprising the Special Cell(s) and all secondary cells for a UE in RRC_CONNECTED configured with CA / .
[0437] The term “Special Cell” refers to the PCell of the MCG or the PSCell of the SCG for DC operation; otherwise, the term “Special Cell” refers to the Pcell.
[0438] The term “machine learning” or “ML” refers to the use of computer systems implementing algorithms and / or statistical models to perform specific task(s) without using explicit instructions, but instead relying on patterns and inferences. ML algorithms build or estimate mathematical model(s) (referred to as “ML models” or the like ] based on sample data (referred to as “training data,” “model training information,” or the like ] in order to make predictions or decisions without being explicitly programmed to perform such tasks. Generally, an ML algorithm is a computer program that learns from experience with respect to some task and some performance measure, and an ML model may be any object or data structure created after an ML algorithm is trained with one or more training datasets. After training, an ML model may be used to make predictions on new datasets. Although the term “ML algorithm” refers to different concepts than the term “ML model,” these terms as discussed herein may be used interchangeably for the purposes of the present disclosure.
[0439] The term “machine learning model,” “ML model,” or the like may also refer to ML methods and concepts used by an ML-assisted solution. An “ML-assisted solution” is a solution that addresses a specific use case using ML algorithms during operation. ML models include supervised learning (e.g., linear regression, k-nearest neighbor (KNN), descision tree algorithms, support machine vectors, Bayesian algorithm, ensemble algorithms, etc.) unsupervised learning (e.g., K-means clustering, principle component analysis (PCA), etc.), reinforcement learning (e.g., Q-learning, multi-armed bandit learning, deep RL, etc.), neural networks, and the like. Depending on the implementation a specific ML model could have many sub-models as components and the ML model may train all sub-models together. Separately trained ML models can also be chained together in an ML pipeline during inference. An “ML pipeline” is a set of functionalities, functions, or functional entities specific for an ML-assisted solution; an ML pipeline may include one or several data sources in a data pipeline, a model training pipeline, a model evaluation pipeline, and an actor. The “actor” is an entity that hosts an ML assisted solution using the output of the ML model inference). The term “ML training host” refers to an entity, such as a network function, that hosts the training of the model. The term “ML inference host” refers to an entity, such as a network function, that hosts model during inference mode (which includes both the model execution as well as any online learning if applicable). The ML-host informs the actor about the output of the ML algorithm, and the actor takes a decision for an action (an “action” is performed by an actor as a result of the output of an ML assisted solution). The term “model inference information” refers to information used as an input to the ML model for determining inference(s); the data used to train an ML model and the data used to determine inferences may overlap, however, “training data” and “inference data” refer to different concepts.
Claims
CLAIMS1. An electronic system comprising: one or more processors configured to implement a machine learning (ML) management system of a cellular network; and one or more computer-readable media comprising instructions that, upon execution of the instructions, are to cause the ML management system to: train or facilitate training of a ML model by a ML training module, wherein training of the ML model generates a trained ML model, and wherein the ML model is related to a cellular network system; test or facilitate testing of the trained ML model by a ML testing module, wherein testing of the ML trained ML module includes evaluation of application of the trained ML module to testing data; emulate or facilitate emulation of the trained ML model by a ML emulation module, wherein emulation of the ML model involves performing inference using the ML model in an emulation environment that is different than an inference environment of the cellular network; and identify or facilitate identification regarding, based on the emulation of the trained ML model, whether to re-train the ML model by the ML training module or provide the trained ML model to a ML inference module of the cellular network for inference in the inference environment.
2. The electronic system of claim 1, wherein the inference environment is an environment in which the ML model is applied to real-world live environment of the cellular network system.
3. The electronic system of claim 1, wherein the cellular network system is a fifth generation core (5GC) system.
4. The electronic system of claim 1, wherein the cellular network system is a Next Generation Radio Access Network (NG-RAN).
5. The electronic system of claim 1, wherein the cellular network system is a management system of the 5G network6. The electronic system of any of claims 1-5, wherein provision of the trained ML model to the ML inference module includes loading the trained ML model by a ML loading module.
7. The electronic system of any of claims 1-5, wherein emulation of the trained ML model is performed based on a result of the evaluation of application of the trained ML module to the testing data.
8. The electronic system of any of claims 1-5, wherein the emulation of the trained ML model is based on a request from a consumer of the ML model.
9. The electronic system of any of claims 1-5, wherein the emulation of the trained ML model is based on a request from a producer of the ML model.
10. One or more computer-readable media comprising instructions that, upon execution of the instructions by one or more processors of one or more electronic devices, are to cause a machine learning (ML) management system of a cellular network to: train or facilitate training of a ML model by a ML training module, wherein training of the ML model generates a trained ML model, and wherein the ML model is related to a cellular network system; test or facilitate testing of the trained ML model by a ML testing module, wherein testing of the ML trained ML module includes evaluation of application of the trained ML module to testing data; emulate or facilitate emulation of the trained ML model by a ML emulation module, wherein emulation of the ML model involves performing inference using the ML model in an emulation environment that is different than an inference environment of the cellular network; and identify or facilitate identification regarding, based on the emulation of the trained ML model, whether to re-train the ML model by the ML training module or provide the trained ML model to a ML inference module of the cellular network for inference in the inference environment.
11. The one or more computer-readable media of claim 10, wherein the inference environment is an environment in which the ML model is applied to a live environment of the cellular network system.
12. The one or more computer-readable media of any of claims 10-11, wherein provision of the trained ML model to the ML inference module includes loading the trained ML model by a ML loading module.
13. The one or more computer-readable media of any of claims 10-11, wherein emulation of the trained ML model is performed based on a result of the evaluation of application of the trained ML module to the testing data.
14. The one or more computer-readable media of any of claims 10-11, wherein the emulation of the trained ML model is based on a request from a consumer of the ML model.
15. The one or more computer-readable media of any of claims 10-11, wherein the emulation of the trained ML model is based on a request from a producer of the ML model.
16. A machine learning (ML) management system of a cellular network comprising: a management services (MnS) producer configured to: train a ML model by a ML training module, wherein training of the ML model generates a trained ML model, and wherein the ML model is related to a cellular network system; test the trained ML model by a ML testing module, wherein testing of the ML trained ML module includes evaluation of application of the trained ML module to testing data; emulate the trained ML model by a ML emulation module, wherein emulation of the ML model involves performing inference using the ML model in an emulation environment that is different than an inference environment of the cellular network; and identify, based on the emulation of the trained ML model, whether to re-train the ML model by the ML training module or provide the trained ML model to a ML inference module of the cellular network for inference in the inference environment; anda MnS consumer configured to perform inference in the inference environment based on the trained ML model.
17. The ML management system of claim 10, wherein the inference environment is an environment in which the ML model is applied to a live environment of the cellular network system.
18. The ML management system of any of claims 16-17, wherein emulation of the trained ML model is performed based on a result of the evaluation of application of the trained ML module to the testing data.
19. The ML management system of any of claims 16-17, wherein the emulation of the trained ML model is based on a request from a consumer of the ML model.
20. The ML management system of any of claims 16-17, wherein the emulation of the trained ML model is based on a request from a producer of the ML model.
Citation Information
Patent Citations
Apparatus and method for man-machine interaction of AI / ML training hosts
CN114443556A
System and method for providing robust artificial intelligence inference in edge computing devices
US20230108920A1
Managing a wireless device that is operable to connect to a communication network
WO2022013090A1
