AI/ML Capability Signaling for Scenario-Based Model Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless communication systems face challenges in coordinating AI/ML model selection and management between network nodes, leading to inconsistent performance due to lack of awareness of applicable models for current scenarios and configurations.
Innovation Solution
Implement signaling mechanisms for exchanging AI/ML capability and applicability information between network nodes, allowing nodes to select appropriate models based on current conditions and configurations, ensuring consistent performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If signaling mechanisms for AI/ML capability exchange are implemented, then performance consistency is improved, but device complexity increases
Solution Approach 1:
The patent introduces a signaling mechanism that acts as an intermediary between network nodes and UE devices. This mechanism includes capability information elements that carry AI/ML model identifiers, applicability indicators, and configuration parameters. The signaling framework mediates the exchange of information between the network side (which maintains the AI/ML model database) and the UE side (which selects and executes models), enabling coordinated model selection without requiring complex direct communication between all network entities.
Solution Approach 2:
The patent segments the AI/ML management functionality into distinct components: (1) a database component for storing AI/ML models and metadata, (2) a signaling component for exchanging capability information, and (3) a selection component for choosing appropriate models. This segmentation allows each component to be optimized independently and simplifies the overall system architecture by separating concerns between model storage, information exchange, and model selection.
2Adaptability or versatility
If nodes select models based on current conditions and configurations, then adaptability is improved, but information processing requirements increase
Solution Approach 1:
The patent implements preliminary action by pre-configuring AI/ML models with metadata including applicability indicators, supported scenarios, and configuration requirements. The capability information elements are prepared in advance with structured data about model compatibility with different network conditions and UE configurations. This preliminary organization of information allows nodes to quickly determine model applicability without performing complex real-time analysis, reducing the information processing load during actual model selection.
Data Source
AI summary
Various aspects of the present disclosure relate to an apparatus and method for signaling artificial intelligence (AI)/machine learning (ML) functionality. A first request message for capability information associated with AI can be received. A first response message comprising the capability information can be transmitted in response to the received first request message, where the capability information indicates one or more AI functionalities supported by a UE. A second request message can be received based at least in part on the transmitted first response message, where the second request message includes at least one configuration for AI. A second response message can be transmitted in response to the received second request message, where the second response message includes feedback that indicates whether the at least one AI functionality supported by the UE is applicable for the at least one configuration.