Multi-encoding Spike Neural Network Apparatus
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current spike neural network-based neuromorphic systems, primarily using leaky-integrate-and-fire neuron models, do not fully utilize the characteristics of various neuronal models studied in the human brain, limiting their information processing capabilities.
Innovation Solution
A spike neural network apparatus that preprocesses input signals through a combination of encoding methods such as rate coding, temporal coding, phase coding, and synchronous coding, and performs these operations using a neuromorphic chip with a network-on-chip architecture, allowing for more efficient signal processing and classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single encoding method is used in spike neural networks, then the system is simpler to implement, but the information processing capability is limited
Solution Approach 1:
The patent combines multiple encoding methods (rate coding, temporal coding, phase coding, synchronous coding) into a unified encoding system that processes input signals through multiple encoding paths simultaneously. This merging of different encoding approaches enables the spike neural network to utilize diverse neuronal model characteristics, thereby improving information processing capability while managing system complexity through integrated architecture design.
2Productivity
If multiple encoding methods are implemented, then information processing capability improves, but hardware requirements increase
Solution Approach 1:
The patent segments the encoding system into distinct functional modules, each responsible for a specific encoding method (rate coding module, temporal coding module, phase coding module, synchronous coding module). This segmentation allows independent implementation and optimization of each encoding function, improving signal processing efficiency while managing hardware complexity through modular architecture that can be selectively configured.
3Measurement precision
If multiple encoding methods are used, then classification accuracy improves, but operation complexity increases
Solution Approach 1:
The patent applies multiple encoding methods as preliminary processing steps before the main spike neural network classification operation. By pre-encoding input signals through rate coding, temporal coding, phase coding, and synchronous coding, the system prepares the data in multiple representations that enhance classification accuracy. This preliminary encoding action simplifies the subsequent classification operation while maintaining high precision through multi-encoded input.
Data Source
AI summary
Disclosed are a spike neural network apparatus based on a multi-encoding and an operating method thereof. The method of operating a spike neural network (SNN) apparatus that performs a multi-encoding, includes receiving an input signal by an encoding module, performing a rate coding and a temporal coding on the received input signal by the encoding module, generating an SNN input signal based on the performance result of the rate coding and the temporal coding, and transmitting the generated SNN input signal to a neuromorphic chip that performs a spike neural network (SNN) operation.