Spike Neural Network Circuit Pulse Detection Power Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional spike neural networks continuously consume power regardless of whether input spike signals have pulses, as they always compare voltage levels with a threshold voltage, leading to inefficient power usage.
Innovation Solution
A spike neural network circuit incorporating an input spike detecting circuit that generates an enable signal only when a pulse is detected, allowing the neuron circuit to compare the voltage level of an accumulated signal with a threshold voltage only when necessary, thereby reducing unnecessary power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the neuron circuit continuously compares voltage levels with threshold voltage, then the comparison operation is always performed, but power consumption increases continuously
Solution Approach 1:
The patent implements periodic action by using an enable signal that is generated only when a pulse is detected from input spike signals. This enable signal periodically activates the comparison operation in the neuron circuit, replacing continuous comparison with pulse-triggered comparison. The enable signal acts as a periodic control mechanism that turns the comparison operation on only when necessary, thereby reducing power consumption while maintaining reliable detection of spike events.
2Reliability
If the comparison operation is performed continuously, then no pulse detection is missed, but unnecessary power consumption occurs when no pulses are present
Solution Approach 1:
The patent applies preliminary action by introducing an input spike detecting circuit that performs preliminary detection of pulses in input spike signals before the main comparison operation. The enable signal generated by this preliminary detection circuit serves as a gatekeeper, activating the power-consuming comparison operation only when a pulse is actually present. This preliminary detection mechanism prevents energy waste by ensuring the comparison circuit operates only when necessary, while maintaining reliable pulse detection through the sequential detection-comparison process.
3Productivity
If the neuron circuit operates always, then all input signals are processed, but power consumption cannot be reduced
Solution Approach 1:
The patent implements local quality by differentiating the operational state of different circuit components. The input spike detecting circuit continuously monitors all input signals to maintain high productivity, while the neuron circuit's comparison operation is locally controlled by the enable signal to operate only when pulses are detected. This localized control strategy allows the system to maintain high signal processing throughput through continuous monitoring while reducing power consumption by activating the intensive comparison operation only when necessary.
Data Source
AI summary
Disclosed is a spike neural network circuit including a synaptic circuit including synapses arranged in rows and columns, an axon circuit that generates a first input spike signal to be provided to a first row among the rows, and a second input spike signal to be provided to a second row among the rows, an input spike detecting circuit that generates an enable signal when detecting a pulse from at least one of the first input spike signal and the second input spike signal, and a first neuron circuit that compares a voltage level of a first accumulated signal, which is output from a first column among the columns, with a threshold voltage level in response to the enable signal, and outputs a first output spike signal when the voltage level of the first accumulated signal exceeds the threshold voltage level.