Terminal ML Resource Scheduling for MIMO Heat Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increased computational demands of multiple ML modes in wireless communication networks lead to high device temperatures, which can compromise communication performance and require excessive computing resources.
Innovation Solution
A method for scheduling ML processing resources and adjusting MIMO specifications based on factors like real-time requirements, latency, collaboration level, and device temperature, allowing for efficient allocation of resources and temperature control without compromising performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple ML models are operated at the terminal device to improve communication performance, then communication performance is improved, but device temperature increases and computing resources are exhausted
Solution Approach 1:
The patent implements dynamic scheduling of ML processing resources based on real-time device temperature and communication requirements. The terminal device adjusts the number of ML processing resources allocated to different ML models dynamically, increasing resources when temperature is low and communication performance needs improvement, and reducing resources when temperature exceeds thresholds, thereby resolving the contradiction between maintaining communication performance and controlling device temperature.
2Reliability
If multiple ML models are operated at the terminal device to improve communication performance, then communication performance is improved, but computing resources are exhausted
Solution Approach 1:
The patent changes the parameter of ML processing resource allocation based on device state and communication requirements. By adjusting the number of processing resources assigned to each ML model according to temperature thresholds and communication performance needs, the system optimizes computing resource utilization while maintaining necessary communication functions, resolving the contradiction between improved communication performance and reduced computing resource consumption.
3Reliability
If ML processing resources are increased to handle multiple ML models, then communication performance is improved, but device complexity increases
Solution Approach 1:
The patent segments the ML processing resources into distinct allocable units that can be independently assigned to different ML models based on priority and device state. This segmentation allows the system to manage complexity by treating resource allocation as a modular scheduling problem rather than a monolithic system, enabling fine-grained control over resource distribution while maintaining communication performance.
Data Source
AI summary
Example embodiments of the present disclosure relate to an effective mechanism for communication. In this solution, the device determines, at a terminal device supporting a plurality of machine learning (ML) models, the respective number of ML processing resources or a respective priority for each ML model of at least one ML model being operated at the terminal device and schedules, based on the respective numbers of ML processing resources or the respective priorities, ML processing resources for the at least one ML model. In this way, the limited ML processing resource/capacity may be allocated to the most importance ML model(s).