ai 0.1.7

Library that help construct neural networks (AI) with Dart. Each type of network is one object and designed as close as possible to human brain.

Library for creating AI with Dart

Created under a MIT-style license.

Overview

This library represent an simple way to create neural network.

There are 2 types of neural network that can be created:

• MLP(multilayer and single-layer perceptron).
• AE(autoencoder).

API

MLP

Perseptron is designed due to Rosenblatt's perseptron.

The main class is the MLP which can contains Layers. Each layer consist of one or many Neurons. First layer always consist of InputNeurons where each of them take one input value and have weight equal to 1. All neurons of previous layer have contacts with each neurons of next layer.

In learning is used backpropagation algorithm.

AE

Architecture of AE is the same as MLP, except that first is used for encoding data.

Memory

Neural network have long-time and short-time memory. All information (knowledge - weights of synapces) of neural network during studying pass through short-time memory. When studying finished and knowledge is structured, then it pass to long-time memory. Knowledge is saved in JSON file knowledge.json in resources directory in the root of your library. For next time network take knowledge from file and initialize with proper weights.

Structure

Neural network can be created from predefined structure difined in structure.json file. You can place it anywhere you want, but default and preffered way is placing it in resources directory in the root of your library. Every structure.json must have type property that corresspond to neural network's names.

Structure of MLP:

{
  "type": "MLP",
  "input": 15, // count of `InputNeuron`s
  "hiddens": [3], // array length shows count of hidden `Layer`s and values are count of `Neuron`s of each layer
  "output": 3 // count of output `Neuron`s
}

Structure of AE:

{
  "type": "AE",
  "input": 15, // count of `InputNeuron`s
  "hiddens": [3], // array length shows count of hidden `Layer`s and values are count of `Neuron`s of each layer for encoded and decoded parts
  "encoded": 3 // count of `Neuron`s that encode data
}

Visualization

Visualization of process of training network is available. Implemented only Mean Squared Error (MSE). If visualise paramenter of train() method is true then MSE sends to console.

Sample

This testing network recognize number 5 from numbers in range from 0 to 9. Also she detects distorted numbers of 5.

import 'package:ai/ai.dart';

void main() {
  final l1 = Layer<InputNeuron>(<InputNeuron>[
    InputNeuron(),
    InputNeuron(),
    InputNeuron(),
    InputNeuron(),
    InputNeuron(),
    InputNeuron(),
    InputNeuron(),
    InputNeuron(),
    InputNeuron(),
    InputNeuron(),
    InputNeuron(),
    InputNeuron(),
    InputNeuron(),
    InputNeuron(),
    InputNeuron()
  ]);
  final l2 = Layer<Neuron>(<Neuron>[
    Neuron(15),
    Neuron(15),
    Neuron(15),
    Neuron(15),
    Neuron(15)
  ]);
  final l3 = Layer<Neuron>(<Neuron>[
    Neuron(5)
  ]);
  final n = MLP(<Layer<NeuronBase>>[
    l1,
    l2,
    l3
  ]);

  // Expected results according to learning data (10)
  final expected = <List<double>>[
    <double>[0.01],
    <double>[0.01],
    <double>[0.01],
    <double>[0.01],
    <double>[0.01],
    <double>[0.99], // 5
    <double>[0.01],
    <double>[0.01],
    <double>[0.01],
    <double>[0.01]
  ];

  // Learning data (10)
  final trainInput = <List<double>>[
    '111101101101111'.split('').map(double.parse).toList(),
    '001001001001001'.split('').map(double.parse).toList(),
    '111001111100111'.split('').map(double.parse).toList(),
    '111001111001111'.split('').map(double.parse).toList(),
    '101101111001001'.split('').map(double.parse).toList(),
    '111100111001111'.split('').map(double.parse).toList(), // 5
    '111100111101111'.split('').map(double.parse).toList(),
    '111001001001001'.split('').map(double.parse).toList(),
    '111101111101111'.split('').map(double.parse).toList(),
    '111101111001111'.split('').map(double.parse).toList()
  ];

  // Testing data
  final testInput = <List<double>>[
    '111100111000111'.split('').map(double.parse).toList(),
    '111100010001111'.split('').map(double.parse).toList(),
    '111100011001111'.split('').map(double.parse).toList(),
    '110100111001111'.split('').map(double.parse).toList(),
    '110100111001011'.split('').map(double.parse).toList(),
    '111100101001111'.split('').map(double.parse).toList()
  ];

  // Number which this network must recognize
  final num5 = '111100111001111'.split('').map(double.parse).toList();

  // This network trains
  n.train(input: trainInput, expected: expected, learningRate: 0.42, epoch: 5000, visualize: true);

  // This network predicts result
  print('Recognize 5? - ${n.predict(num5)}');
  for (var item in testInput) {
    print('Recognize distorted 5? - ${n.predict(item)[0]}');
  }
  print('Аnd 0? - ${n.predict(trainInput[0])}');
  print('Аnd 8? - ${n.predict(trainInput[8])}');
  print('Аnd 3? - ${n.predict(trainInput[3])}');
}

Features and bugs

Please file feature requests and bugs at the issue tracker.

With ❤️ to AI