ML Model Status Notification for Wireless Communication Efficiency

The patent segments the machine learning system into distinct components: model generation unit, model selection unit, and inference unit. This segmentation allows each component to perform its specific function independently, managing complexity while maintaining overall system efficiency. The model generation unit creates ML models, the selection unit chooses appropriate models based on channel conditions, and the inference unit executes predictions, thereby resolving the contradiction between improved productivity and increased device complexity.

The patent employs copying by training machine learning models to replicate the complex relationship between channel inputs and state outputs. Instead of transmitting detailed channel state information, the system copies the essential patterns into pre-trained models that can be efficiently transmitted and executed. This approach maintains high feedback accuracy while significantly reducing the overhead required for channel state information transmission.

The patent applies preliminary action by pre-training machine learning models offline before deployment. The model generation unit creates trained models in advance based on historical channel data, storing them for later use. During actual communication, the inference unit executes these pre-trained models rather than performing training computations in real-time. This preliminary preparation significantly reduces power consumption during operation while maintaining high communication accuracy.