neurons.c File Reference

Standalone neuron_s example with explicit connectivity. More...

Go to the source code of this file.

Detailed Description

Standalone neuron_s example with explicit connectivity.

Author

Oscar Sotomayor

This file provides a minimal, executable example that demonstrates how multiple neuron_s instances can be connected and evaluated without using the full NeuroTIC network container.

The example:

• constructs a small set of neurons manually,
• connects them using direct pointer references,
• evaluates them using an explicit external loop,
• and prints the resulting output.

The purpose of this file is illustrative. It shows how neuron_s behaves as a passive structure when driven by user-defined execution code.

This example does not represent the complete NeuroTIC execution model. No net_s, wiring_s, training, or activation dispatch is involved.

For the structural definition of neuron_s, see:

• ntcore.h
• neuron_s.dox

Example topology

The following diagram represents the concrete topology implemented in this example.

  /
#include "ntcore.h"
 
#define NEURON_0 NAND
#define NEURON_1 OR
#define NEURON_2 AND
 
#define AND             .w= (weight_t []){  0.1 ,  0.1 } , .b= (bias_t)(-0.15)
#define NAND            .w= (weight_t []){ -0.1 , -0.1 } , .b= (bias_t)( 0.15)
#define OR              .w= (weight_t []){  0.1 ,  0.1 } , .b= (bias_t)(-0.05)
#define NOR             .w= (weight_t []){ -0.1 , -0.1 } , .b= (bias_t)( 0.05)
#define CONJUNCION_A    .w= (weight_t []){ -0.1 ,  0.1 } , .b= (bias_t)(-0.05)
#define CONJUNCION_B    .w= (weight_t []){  0.1 , -0.1 } , .b= (bias_t)(-0.05)
#define IMPLICATION_A   .w= (weight_t []){ -0.1 ,  0.1 } , .b= (bias_t)( 0.05)
#define IMPLICATION_B   .w= (weight_t []){  0.1 , -0.1 } , .b= (bias_t)( 0.05)
 
#include <stdio.h>
 
int main( void ) {
    data_t inputs[2];
 
    neuron_s neuron[]={
        [0]= { .inputs= sizeof( inputs ) / sizeof( data_t ) , .bff_idx= 0 , .in= (data_t *[]){ &inputs[0] , &inputs[1] } , NEURON_0 , .fn= 0 , .out= (data_t)(0.0) },
        [1]= { .inputs= sizeof( inputs ) / sizeof( data_t ) , .bff_idx= 0 , .in= neuron[0].in , NEURON_1 , .fn= 0 , .out= (data_t)(0.0) },
        [2]= { .inputs= 2 , .bff_idx= 0 , .in= (data_t *[]){ &neuron[0].out , &neuron[1].out } , NEURON_2 , .fn= 0 , .out= (data_t)(0.0) }
    };
 
    for( uint8_t i= 0 ; i < 4 ; i++ ){
        for( size_t j= 0 , f= sizeof( inputs ) / sizeof( data_t ) ; j < f ; j++ ) inputs[j]= !!(data_t)( i & ( 1 << j ) );
        for( size_t j= 0 , f= sizeof( neuron ) / sizeof( neuron_s ) ; j < f ; j++ ){
            float sum= neuron[j].b;
            for( input_t k= 0 ; k < neuron[j].inputs ; k++ ) sum+= *neuron[j].in[k] * neuron[j].w[k];
            neuron[j].out= (data_t)( sum >= 0.0f );
        }
        printf( "| input_A: %0.0f | input_B: %0.0f | output: %0.0f |\n", inputs[0] , inputs[1] , neuron[2].out );
    }
 
    return 0;
}

Definition in file neurons.c.