Features:

• Image Classifier
• Multi-Layer Perceptron
• Hopfield Network

Planned:

• Discreet Sequence Recall
• Genetic Algorithm
• KNN
• etc

Image Classifier

No training required but does require a manual tap interaction step to use.

Example Use

Classification

Classification is done by passing a dictionary containing a base64 encoded PNG Image.

{
  "Architecture": "Image Classifier",
  "Image": "iVBORw0K..."
}

Multi-Layer Perceptron

Training

Training the Perceptron is done by passing a dictionary describing the inputs, hidden layer, and outputs.

Training data is a \n separated list of the form [inputs,...]=>[output]

{
  "Name": "XOR",
  "Architecture": "Perceptron",
  "Inputs": 2,
  "Hidden": [2],
  "Outputs": 1,
  "Training Data": "[0,1]=>[1]\n[1,0]=>[1]\n[1,1]=>[0]\n[0,0]=>[0]",
  "Error": 0.0001,
  "Iterations": 50000
}

Inference

Inference is done by passing a dictionary with Inputs provided as [inputs,...]

{
  "Name":"XOR",
  "Architecture":"Perceptron",
  "Inputs":"[0,0]"
}

Hopfield Network

Training

Training the Hopfield network is done by passing a dictionary describing the inputs binary sequences and bit-length of the network.

{
  "Name":"10-Bit Example",
  "Architecture":"Hopfield",
  "Bit Length":10,
  "Training Data":[
    "[0, 1, 0, 1, 0, 1, 0, 1, 0, 1]",
    "[1, 1, 1, 1, 1, 0, 0, 0, 0, 0]"
]}

Feeding

Feeding the hopfield is done by passing a dictionary with Inputs provided as [input,...]

{
  "Name":"10-Bit Example",
  "Architecture":"Hopfield",
  "Inputs":"[0,1,0,1,0,1,0,1,1,1]"
}

Dependency Badge

[![This shortcut depends on ShortcutML](http://magaimg.net/img/8kdh.png)](https://routinehub.co/shortcut/3246)

Credits

Perceptron & Hopfield Based on Synaptic

Image Classifier Based on ml5.js

Supports