Conditional-Bias Comparator in Spiking Neural Neuron Circuits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing number of neurons in spike neural networks leads to higher power consumption, which is not efficiently managed by existing technologies.
Innovation Solution
A spike neural network circuit that includes a comparator operated by a conditional bias current, where the bias circuit supplies a bias current only when necessary, based on the membrane signal, reducing power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the number of neurons is increased to enhance processing capability, then the computational power and data processing ability are improved, but the power consumption increases significantly
Solution Approach 1:
The bias circuit operates periodically by monitoring the membrane signal voltage and enabling bias current only when it exceeds the threshold voltage, creating a pulse-like operation pattern that reduces average power consumption while maintaining computational functionality during active periods
Solution Approach 2:
The bias current is dynamically adjusted based on the membrane signal voltage level, transitioning from a static always-on bias to a dynamic conditional bias that adapts to the operational state of the neuron, enabling power savings during inactive states while maintaining performance during active computation
2Reliability
If continuous bias current is supplied to the comparator to ensure stable operation, then the reliability and operational stability are improved, but the power consumption increases
Solution Approach 1:
The bias circuit incorporates feedback by continuously monitoring the membrane signal voltage and using this information to control the bias current supply, ensuring that bias current is provided only when the neuron is actively processing information, thus maintaining reliability during computation while reducing unnecessary power consumption during idle states
Data Source
AI summary
Provided is a spike neural network circuit including a synapse configured to generate an operation signal based on an input spike signal and a weight, and a neuron configured to generate an output spike signal using a comparator configured to compare a voltage of a membrane signal generated based on the operation signal with a voltage of a threshold signal, wherein the comparator includes a bias circuit configured to conditionally supply a bias current of the comparator depending on the membrane signal.