Spiking Synaptic Elements Frequency Modulation Dynamic Range
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing hardware-implemented spiking neural networks face challenges in efficiently transmitting signals between neurons, as they typically convert spikes into continuous analog currents, limiting the dynamic range and efficiency of signal transmission.
Innovation Solution
The method involves modulating spikes received from pre-synaptic neurons in frequency based on synaptic weights to generate post-synaptic spikes, which are then transmitted to post-synaptic neurons, allowing for a higher dynamic range and efficient discrete event transmission, compatible with digital technology and memristive devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If spikes are converted to continuous analog currents at synapse output, then signal transmission is simplified, but the dynamic range is limited
Solution Approach 1:
The patent applies periodic action by converting continuous analog currents into discrete periodic spike trains. Each synapse outputs spikes at a frequency proportional to the analog current magnitude, transforming a continuous signal into a periodic discrete signal that preserves information while enabling higher dynamic range representation through spike counting and frequency modulation.
2Quantity of substance
If discrete spike events are used for signal transmission, then dynamic range is improved, but conversion complexity increases
Solution Approach 1:
The patent replaces complex analog-to-digital conversion mechanisms with a simpler frequency encoding scheme. Instead of converting spikes to analog currents and then back to digital values, the system directly encodes synaptic weight information into the frequency of output spikes, eliminating intermediate conversion steps and reducing overall system complexity.
3Measurement precision
If frequency modulation is used to encode synaptic weights, then signal precision is improved, but processing complexity increases
Solution Approach 1:
The patent implements self-service by having each synapse autonomously generate spike trains with frequencies determined by its own synaptic weight and the incoming spike rate. This distributed self-regulation eliminates the need for centralized control or complex processing units to manage signal transmission, as each synapse automatically encodes and transmits information based on local parameters.
Data Source
AI summary
Embodiment of the invention are directed to transmitting signals between neurons of a hardware-implemented, spiking neural network (or SNN). The network includes neuronal connections, each including a synaptic unit connecting a pre-synaptic neuron to a post-synaptic neuron. Spikes received from the pre-synaptic neuron of said each neuronal connection are first modulated, in frequency, based on a synaptic weight stored on said each synaptic unit, to generate post-synaptic spikes, such that a first number of spikes received from the pre-synaptic neuron are translated into a second number of post-synaptic spikes. At least some of the spikes received from the pre-synaptic neuron may, each, be translated into a train of two or more post-synaptic spikes. The post-synaptic spikes generated are subsequently transmitted to the post-synaptic neuron of said each neuronal connection. The novel approach makes it possible to obtain a higher dynamic range in the synapse output.