Dynamic Memory Block Allocation for ML Sensor Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Ultra-low power sensor applications using machine learning networks, such as neural networks, are inefficient and consume excessive resources, leading to increased power consumption and resource utilization.
Innovation Solution
Implementing logical switches in machine learning networks to enable controllable memory block allocation based on performance levels, allowing for adaptive activation or deactivation of memory blocks to reduce data movement and power usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning networks use fixed memory block allocation to ensure full performance, then processing accuracy is maintained, but power consumption increases and resource efficiency decreases
Solution Approach 1:
The patent implements dynamic memory block allocation by introducing logical switches that can be activated or deactivated based on performance requirements. The system dynamically adjusts the number of active memory blocks from 1 to N, allowing the machine learning network to adapt its resource usage to match the actual processing needs, thereby reducing power consumption while maintaining necessary accuracy.
Solution Approach 2:
The system changes the operational parameters of the memory blocks by controlling their activation state. By varying the number of active memory blocks based on performance level requirements, the system optimizes the balance between processing accuracy and power consumption, enabling ultra-low power sensor applications to operate efficiently.
2Productivity
If machine learning networks activate all memory blocks to achieve full performance, then processing capability is maximized, but resource efficiency decreases
Solution Approach 1:
The patent applies partial action by activating only the necessary number of memory blocks required to meet the performance threshold, rather than activating all memory blocks. The logical switches enable selective activation of memory blocks based on the specific application's needs, reducing unnecessary resource consumption while maintaining adequate processing capability.
3Use of energy by moving object
If machine learning networks reduce memory block activation to save power, then power consumption decreases, but processing accuracy may deteriorate
Solution Approach 1:
The system incorporates feedback mechanisms that monitor processing performance and adjust the number of active memory blocks accordingly. The logical switches are controlled based on feedback regarding whether the current performance level meets the required threshold, ensuring that processing accuracy is maintained while minimizing power consumption.
Data Source
AI summary
Methods, systems, and devices for configuring a machine learning network are described. A device, which may be otherwise known as user equipment (UE), may support ultra-low power sensor applications. More specifically, the device may support memory block allocation of a machine learning network based on performance levels associated with the applications. For example, the device may identify a performance level associated with an application on the device. The device may determine that the performance level satisfies a condition, and subsequently determine a memory block allocation of a machine learning network of the device based on the performance level satisfying the condition. The memory block allocation may correspond to one or more connections of the machine learning network. Based on the memory block allocation, the device may adjust a quantity of memory blocks available for the machine learning network and process the application.