NanoBrain-unitypackage/Neuron.cs

using System;
using System.Collections.Generic;
using System.Linq;
using UnityEngine;
using UnityEditor;
using Unity.Mathematics;
using static Unity.Mathematics.math;

[Serializable]
public class Neuron : Nucleus {

    public Neuron(Cluster parent, string name) {
        this.parent = parent;
        this.name = name;
        this.parent?.nuclei.Add(this);
    }
    public Neuron(ClusterPrefab parent, string name) {
        this.cluster = parent;
        this.name = name;
        if (this.cluster != null) {
            this.cluster.nuclei.Add(this);
        }
        // else
        //     Debug.LogError("No neuroid network");
    }

    #region Serialization

    public enum CurvePresets {
        Linear,
        Power,
        Sqrt,
        Reciprocal,
        Custom
    }
    [SerializeField]
    private CurvePresets _curvePreset;
    public CurvePresets curvePreset {
        get { return _curvePreset; }
        set {
            _curvePreset = value;
            this.curve = GenerateCurve();
        }
    }
    public AnimationCurve curve;
    public float curveMax = 1.0f;

    #region Parameters

    public bool average = false;

    #endregion Parameters

    public AnimationCurve GenerateCurve() {
        switch (this.curvePreset) {
            case CurvePresets.Linear:
                this.curveMax = 1;
                return Presets.Linear(1);
            case CurvePresets.Power:
                this.curveMax = 1;
                return Presets.Power(2.0f, 1);
            case CurvePresets.Sqrt:
                this.curveMax = 1;
                return Presets.Power(0.5f, 1);
            case CurvePresets.Reciprocal:
                this.curveMax = 1 / 0.01f * 1;
                return Presets.Reciprocal(1);
            default:
                this.curveMax = 1;
                return this.curve;
        }
    }

    public virtual void Deserialize(Neuron nucleus) { }

    #endregion Serialization

    #region Runtime state (not serialized)

    #region Activation

    public static class Presets {
        private const int samples = 32;
        public static AnimationCurve Linear(float weight) {
            return AnimationCurve.Linear(0f, 0f, 1000f, weight * 1000);
        }
        public static AnimationCurve Power(float exponent, float weight) {
            // build keyframes
            Keyframe[] keys = new Keyframe[samples];
            for (int i = 0; i < samples; i++) {
                float t = i / (float)(samples - 1);
                float v = Mathf.Pow(t, exponent) * weight;
                keys[i] = new Keyframe(t, v);
            }

            AnimationCurve curve = new(keys);

            // set tangent modes for each key to Auto (smooth). Use Linear if you prefer straight segments.
            for (int i = 0; i < curve.length; i++) {
                AnimationUtility.SetKeyLeftTangentMode(curve, i, AnimationUtility.TangentMode.Auto);
                AnimationUtility.SetKeyRightTangentMode(curve, i, AnimationUtility.TangentMode.Auto);
            }

            return curve;
        }
        public static AnimationCurve Reciprocal(float weight) {
            int samples = 128;
            float xMin = 0.001f;
            float xMax = 1;
            var keys = new Keyframe[samples];
            for (int i = 0; i < samples; i++) {
                float t = i / (float)(samples - 1);
                float x = Mathf.Lerp(xMin, xMax, t);
                float y = 1f / x * weight;
                keys[i] = new Keyframe(x, y);
            }
            var curve = new AnimationCurve(keys);
            for (int i = 0; i < curve.length; i++) {
                AnimationUtility.SetKeyLeftTangentMode(curve, i, AnimationUtility.TangentMode.Linear);
                AnimationUtility.SetKeyRightTangentMode(curve, i, AnimationUtility.TangentMode.Linear);
            }
            return curve;
        }
    }

    #endregion Activation

    #endregion Runtime state

    // this clone the nucleus without the synapses and receivers
    public override Nucleus ShallowCloneTo(Cluster newParent) {
        Neuron clone = new(newParent, this.name);
        CloneFields(clone);
        return clone;
    }

    public override Nucleus Clone(ClusterPrefab prefab) {
        Neuron clone = new(prefab, this.name);
        CloneFields(clone);
        foreach (Synapse synapse in this.synapses) {
            Synapse clonedSynapse = clone.AddSynapse(synapse.nucleus);
            clonedSynapse.weight = synapse.weight;
        }
        foreach (Nucleus receiver in this.receivers) {
            clone.AddReceiver(receiver);
        }
        return clone;
    }

    protected virtual void CloneFields(Neuron clone) {
        clone.array = null;
        clone.bias = this.bias;
        clone.curve = this.curve;
        clone.curvePreset = this.curvePreset;
        clone.curveMax = this.curveMax;
        clone.average = this.average;
    }

    public static void Delete(Nucleus nucleus) {
        foreach (Synapse synapse in nucleus.synapses) {
            if (synapse.nucleus is Neuron synapse_nucleus) {
                if (synapse_nucleus.receivers.Count > 1) {
                    // there is another nucleus feeding into this input nucleus
                    synapse_nucleus.receivers.RemoveAll(r => r == nucleus);
                }
                else {
                    // No other links, delete it.
                    Neuron.Delete(synapse_nucleus);
                }
            }
        }
        foreach (Nucleus receiver in nucleus.receivers) {
            if (receiver != null && receiver.synapses != null)
                receiver.synapses.RemoveAll(s => s.nucleus == nucleus);
        }

        if (nucleus.cluster != null) {
            nucleus.cluster.nuclei.RemoveAll(n => n == nucleus);
            nucleus.cluster.GarbageCollection();
        }
    }

    public override void UpdateStateIsolated() {
        Vector3 sum = this.bias;
        int n = 0;

        //Applying the weight factgors
        foreach (Synapse synapse in this.synapses) {
            sum += synapse.weight * synapse.nucleus.outputValue;

            // Perhaps synapses should be removed when the output value goes to 0....
            if (lengthsq(synapse.nucleus.outputValue) != 0) {
                n++;
                this.stale = 0;
            }
        }
        if (this.average && n > 0)
            sum /= n;

        // Activation function
        float3 result = Activation(sum);
        if (this.stale > staleValueForSleep)
            this.outputValue = new float3(0, 0, 0);
        else
            this.outputValue = result;
    }

    protected float3 Activation(float3 input) {
        float3 result = Vector3.zero;
        switch (this.curvePreset) {
            case CurvePresets.Linear:
                result = input;
                break;
            case CurvePresets.Sqrt:
                result = normalize(input) * System.MathF.Sqrt(length(input));
                break;
            case CurvePresets.Power:
                result = normalize(input) * System.MathF.Pow(length(input), 2);
                break;
            case CurvePresets.Reciprocal: {
                    float magnitude = length(input);
                    if (magnitude > 0)
                        result = normalize(input) * (1 / magnitude);
                    break;
                }
            default:
                float activatedValue = this.curve.Evaluate(length(input));
                result = normalize(input) * activatedValue;
                break;
        }
        return result;
    }

    public virtual void ProcessStimulus(Vector3 inputValue, string thingName = null) {
        this.stale = 0;
        this.bias = inputValue;
        this.parent.UpdateFromNucleus(this);
    }
}