Wireless Terminal AI/ML Model Inference and Feedback Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wireless communication systems face challenges in efficiently sharing and managing artificial intelligence (AI)/machine learning (ML) models, particularly in reporting model performance feedback (MPF) to base stations.
Innovation Solution
A method and device for operating terminals and base stations in wireless communication systems, which involve receiving AI/ML model information and MPF-related information, performing model inference, and determining whether to transmit MPF to the base station based on performance evaluation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If AI/ML model information is shared between terminal and base station, then model performance evaluation is improved, but communication overhead and processing complexity increase
Solution Approach 1:
The patent divides the AI/ML model processing into two segments: model inference execution at the terminal and model performance feedback (MPF) generation at the base station. The terminal performs inference using received model information and generates MPF locally, while the base station receives and processes this feedback. This segmentation reduces the processing burden on individual devices while maintaining comprehensive performance evaluation capabilities.
Solution Approach 2:
The patent introduces MPF (model performance feedback) as an intermediary mechanism that facilitates efficient communication between terminal and base station. The MPF serves as a structured intermediate representation that conveys performance evaluation results without requiring transmission of complete model data or complex processing at both ends, thereby reducing overall system complexity while preserving evaluation accuracy.
2Reliability
If MPF is transmitted to base station for model updates, then model accuracy is improved, but communication overhead increases
Solution Approach 1:
The patent extracts only the essential performance evaluation results into the MPF structure, separating critical performance metrics from redundant model data. By taking out only the necessary feedback information (such as performance measurements and comparisons against reference values) rather than transmitting complete model datasets, the system maintains model accuracy improvements while significantly reducing communication overhead and energy consumption.
Solution Approach 2:
The patent implements partial action by transmitting only the necessary MPF information required for model updates rather than all available data. The MPF structure includes selective performance metrics and comparison results that are sufficient for base station model optimization, avoiding the excessive transmission of complete model parameters or unnecessary performance data, thus balancing reliability with communication efficiency.
3Adaptability or versatility
If model inference is performed at terminal, then communication capability is improved, but processing power consumption increases
Solution Approach 1:
The patent applies preliminary action by receiving and storing AI/ML model information at the terminal before performing inference. The terminal pre-loads necessary model parameters and configurations from the base station, enabling local inference execution without requiring continuous high-power communication or cloud processing. This preliminary preparation allows the terminal to perform communication capability functions using its existing processing resources, reducing real-time power consumption while maintaining adaptability.
Data Source
AI summary
In the present disclosure, a method for operating a terminal in a wireless communication may include receiving, by the terminal, a synchronization signal from a base station, transmitting a random access preamble to the base station based on the synchronization signal, receiving a random access response based on the random access preamble, performing connection with the base station after receiving the random access response, receiving at least any one of AI/ML model information and model performance feedback-related information for an AI/ML model from the base station, and performing model inference in the AI/ML model based on the AI/ML model information and determining, through model performance evaluation, whether or not to transmit MPF of the AI/ML model to the base station.