Fixed Point Neural Network Quantization Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Neural networks require significant computational resources for processing complex input data, which is a challenge for devices with limited resources like smartphones, necessitating a technique to reduce operations and maximize performance.
Innovation Solution
A method and apparatus for generating a fixed point neural network by selecting layers, forming candidate parameter sets, determining update parameter values based on network performance, and updating quantization parameters to optimize network performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If floating point neural network is used, then network performance and accuracy are improved, but computational complexity and resource consumption increase
Solution Approach 1:
The patent changes the numerical representation parameter from floating point to fixed point format. This parameter change reduces computational complexity and resource consumption while maintaining acceptable network performance through optimized quantization schemes and layer-wise fixed point precision adjustment
Solution Approach 2:
The patent applies different fixed point precision levels to different layers of the neural network based on their specific requirements. Critical layers receive higher precision while less critical layers use lower precision, optimizing the balance between performance and computational complexity
2Use of energy by moving object
If quantization parameter is reduced for fixed point conversion, then resource consumption is reduced, but quantization error increases
Solution Approach 1:
The patent optimizes the quantization parameter (bit width) for each layer's fixed point representation. By carefully selecting parameter values that balance precision and resource usage, the system reduces quantization error while maintaining low resource consumption appropriate for fixed point arithmetic
Solution Approach 2:
The patent employs dynamic quantization parameter adjustment where the precision level adapts based on the specific layer's needs and input characteristics. This allows the system to minimize quantization error in critical paths while maintaining resource efficiency overall
Data Source
AI summary
A method and apparatus for generating a fixed point neural network are provided. The method includes selecting at least one layer of a neural network as an object layer, wherein the neural network includes a plurality of layers, each of the plurality of layers corresponding to a respective one of plurality of quantization parameters; forming a candidate parameter set including candidate parameter values with respect to a quantization parameter of the plurality of quantization parameters corresponding to the object layer; determining an update parameter value from among the candidate parameter values based on levels of network performance of the neural network, wherein each of the levels of network performance correspond to a respective one of the candidate parameter values; and updating the quantization parameter with respect to the object layer based on the update parameter value.