Struct AdvancedMemristiveLearning 

pub struct AdvancedMemristiveLearning { /* private fields */ }

Advanced memristive learning system with synaptic plasticity and homeostasis

This system implements a comprehensive memristive computing framework with multiple types of plasticity, homeostatic regulation, and neuromodulation. It supports various memristive device types and advanced learning dynamics.

Features

• Multiple memristive device types (TiO2, HfO2, Phase Change, etc.)
• Advanced plasticity mechanisms (STDP, homeostatic scaling, etc.)
• Homeostatic regulation for stable learning
• Metaplasticity for learning-to-learn capabilities
• Neuromodulation systems (dopamine, serotonin, etc.)
• Memory consolidation and forgetting protection
• Comprehensive learning history tracking

Example

use scirs2_core::ndarray::{Array1, Array2};
use scirs2_spatial::neuromorphic::algorithms::AdvancedMemristiveLearning;
use scirs2_spatial::neuromorphic::MemristiveDeviceType;

let mut learning_system = AdvancedMemristiveLearning::new(4, 2, MemristiveDeviceType::TitaniumDioxide)
    .with_forgetting_protection(true);

// Training data
let spatial_data = Array2::from_shape_vec((4, 4), vec![
    0.0, 0.0, 1.0, 1.0,
    1.0, 0.0, 0.0, 1.0,
    0.0, 1.0, 1.0, 0.0,
    1.0, 1.0, 0.0, 0.0
]).unwrap();
let targets = Array1::from_vec(vec![0.0, 1.0, 1.0, 0.0]);

// Train the system
let result = learning_system.train_spatial_data(&spatial_data.view(), &targets.view(), 50).await.unwrap();
println!("Final accuracy: {:.2}", result.training_metrics.last().unwrap().accuracy);

Implementations

impl AdvancedMemristiveLearning

pub fn new(rows: usize, cols: usize, device_type: MemristiveDeviceType) -> Self

Create new advanced memristive learning system

Arguments

• rows - Number of rows in crossbar array
• cols - Number of columns in crossbar array
• device_type - Type of memristive device to simulate

Returns

A new AdvancedMemristiveLearning system with default parameters

pub fn enable_plasticity(self, plasticity_type: PlasticityType) -> Self

Enable specific plasticity mechanism

Arguments

• plasticity_type - Type of plasticity to enable

pub fn with_homeostatic_regulation(self, target_rates: Array1<f64>) -> Self

Configure homeostatic regulation

Arguments

• target_rates - Target firing rates for each neuron

pub fn with_forgetting_protection(self, enabled: bool) -> Self

Enable catastrophic forgetting protection

Arguments

• enabled - Whether to enable forgetting protection mechanisms

pub async fn train_spatial_data( &mut self, spatial_data: &ArrayView2<'_, f64>, target_outputs: &ArrayView1<'_, f64>, epochs: usize, ) -> SpatialResult<TrainingResult>

Train on spatial data with advanced plasticity

Performs training using all enabled plasticity mechanisms, homeostatic regulation, and neuromodulation.

Arguments

• spatial_data - Input spatial data (n_samples × n_features)
• target_outputs - Target outputs for each sample
• epochs - Number of training epochs

Returns

Training results including final weights and learning history

pub fn crossbar_dimensions(&self) -> (usize, usize)

Get crossbar dimensions

pub fn device_type(&self) -> &MemristiveDeviceType

Get device type

pub fn conductances(&self) -> &Array2<f64>

Get current conductances

pub fn learning_history(&self) -> &LearningHistory

Get learning history

Trait Implementations

impl Clone for AdvancedMemristiveLearning

fn clone(&self) -> AdvancedMemristiveLearning

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for AdvancedMemristiveLearning

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.