Automated Neuron Model Parameter Tuning for Spiking Networks

Solution Overview

Problem

Researchers face challenges in manually tuning parameters of artificial neural networks to match prototypical neuron dynamics, which is time-consuming and requires extensive mathematical expertise, especially when trying to replicate the behavior of biological neurons in spiking neural networks.

Innovation Solution

An automated method using a piecewise linear neuron model to determine parameters that match prototypical neuron dynamics, employing optimization metrics to quantify differences in membrane voltages and iteratively adjust parameters for optimal alignment.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual tuning of neuron model parameters is performed to match prototypical neuron dynamics, then the accuracy of replicating biological neuron behavior is improved, but the time required and expertise needed increase significantly

Engineering Contradiction:

Improveaccuracy of neuron dynamics matchingVSAvoidtime required for parameter tuning

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs self-tuning by automatically adjusting neuron model parameters to match prototypical neuron dynamics. The optimization algorithm independently modifies parameters without requiring manual intervention, enabling the system to self-calibrate and achieve accurate replication of biological neuron behavior autonomously

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The invention systematically varies and optimizes neuron model parameters (such as membrane conductance, threshold voltage, and time constants) to match the dynamics of prototypical neurons. By automatically adjusting these parameters through optimization algorithms, the system achieves precise matching of neural dynamics without manual tuning

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If manual tuning of neuron model parameters is performed to match prototypical neuron dynamics, then the accuracy of replicating biological neuron behavior is improved, but the complexity of the process increases

Engineering Contradiction:

Improveaccuracy of neuron dynamics matchingVSAvoidcomplexity of parameter tuning process

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs self-tuning by automatically adjusting neuron model parameters to match prototypical neuron dynamics. The optimization algorithm independently modifies parameters without requiring manual intervention, enabling the system to self-calibrate and achieve accurate replication of biological neuron behavior autonomously

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The invention replaces the manual mechanical process of parameter tuning with an automated computational optimization system. Instead of researchers manually adjusting parameters based on mathematical models, the system uses algorithms to automatically compute and adjust parameters, substituting human cognitive processes with computational processes

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If automated parameter optimization is implemented, then the time required for tuning is reduced, but the complexity of the optimization system increases

Engineering Contradiction:

Improvespeed of parameter tuningVSAvoidcomplexity of optimization system

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The invention replaces the manual mechanical process of parameter tuning with an automated computational optimization system. Instead of researchers manually adjusting parameters based on mathematical models, the system uses algorithms to automatically compute and adjust parameters, substituting human cognitive processes with computational processes

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The invention systematically varies and optimizes neuron model parameters (such as membrane conductance, threshold voltage, and time constants) to match the dynamics of prototypical neurons. By automatically adjusting these parameters through optimization algorithms, the system achieves precise matching of neural dynamics without manual tuning

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS9305256B2Automated method for modifying neural dynamics

Publication Date: 2016.04.05 QUALCOMM INC

AI summary

A method for improving neural dynamics includes obtaining prototypical neuron dynamics. The method also includes modifying parameters of a neuron model so that the neuron model matches the prototypical neuron dynamics. The neuron dynamics comprise membrane voltages and/or spike timing.