Awesome

gleam_synapses

A plug-and-play library for neural networks written in Gleam!

Basic usage

Install synapses

Run gleam add gleam_synapses in the directory of your project.

Import the `Net` module

import gleam_synapses/net.{type Net}

Create a random neural network by providing its layer sizes

let rand_net = net.new([2, 3, 1])

Input layer: the first layer of the network has 2 nodes.
Hidden layer: the second layer has 3 neurons.
Output layer: the third layer has 1 neuron.

Get the json of the random neural network

net.to_json(rand_net)
// "[
//   [{\"activationF\" : \"sigmoid\", \"weights\" : [-0.5,0.1,0.8]},
//    {\"activationF\" : \"sigmoid\", \"weights\" : [0.7,0.6,-0.1]},
//    {\"activationF\" : \"sigmoid\", \"weights\" : [-0.8,-0.1,-0.7]}],
//   [{\"activationF\" : \"sigmoid\", \"weights\" : [0.5,-0.3,-0.4,-0.5]}]
// ]"

Create a neural network by providing its json

let network = net.from_json("[
   [{\"activationF\" : \"sigmoid\", \"weights\" : [-0.5,0.1,0.8]},
    {\"activationF\" : \"sigmoid\", \"weights\" : [0.7,0.6,-0.1]},
    {\"activationF\" : \"sigmoid\", \"weights\" : [-0.8,-0.1,-0.7]}],
   [{\"activationF\" : \"sigmoid\", \"weights\" : [0.5,-0.3,-0.4,-0.5]}]
 ]")

Make a prediction

net.predict(network, [0.2, 0.6])
// [0.49131100324012494]

Train a neural network

net.fit(network, 0.1, [0.2, 0.6], [0.9])

The fit method returns the neural network with its weights adjusted to a single observation.

Advanced usage

Fully train a neural network

In practice, for a neural network to be fully trained, it should be fitted with multiple observations, usually by folding over an iterator.

[#([0.2, 0.6], [0.9]),
 #([0.1, 0.8], [0.2]),
 #([0.5, 0.4], [0.6])]
|> iterator.from_list
|> iterator.fold(network, fn(acc, t) {
  let #(xs, ys) = t
  net.fit(acc, 0.1, xs, ys)
})

Boost the performance

Every function is efficient because its implementation is based on lazy list and all information is obtained at a single pass.

For a neural network that has huge layers, the performance can be further improved by using the parallel counterparts of predict and fit (par_predict and par_fit).

Create a neural network for testing

net.new_with_seed([2, 3, 1], 1000)

We can provide a seed to create a non-random neural network. This way, we can use it for testing.

Define the activation functions and the weights

import gleam_synapses/fun.{type Fun}
import gleam/float

let activation_f = fn(layer_index: Int) -> Fun {
  case layer_index {
    0 -> fun.sigmoid()
    1 -> fun.identity()
    2 -> fun.leaky_re_lu()
    3 -> fun.tanh()
  }
}

let weight_init_f = fn(_: Int) -> Float {
  float.random(0.0, 1.0)
}

let custom_net = net.new_custom([4, 6, 8, 5, 3], activation_f, weight_init_f)

The activation_f function accepts the index of a layer and returns an activation function for its neurons.
The weight_init_f function accepts the index of a layer and returns a weight for the synapses of its neurons.

If we don't provide these functions, the activation function of all neurons is sigmoid, and the weight distribution of the synapses is normal between -1.0 and 1.0.

Draw a neural network

net.to_svg(custom_net)

Network Drawing

With its svg drawing, we can see what a neural network looks like. The color of each neuron depends on its activation function while the transparency of the synapses depends on their weight.

Measure the difference between the expected and predicted values

import gleam_synapses/stats

fn exp_and_pred_vals() -> Iterator(#(List(Float), List(Float))) {
  [
    #([0.0, 0.0, 1.0], [0.0, 0.1, 0.9]),
    #([0.0, 1.0, 0.0], [0.8, 0.2, 0.0]),
    #([1.0, 0.0, 0.0], [0.7, 0.1, 0.2]),
    #([1.0, 0.0, 0.0], [0.3, 0.3, 0.4]),
    #([0.0, 0.0, 1.0], [0.2, 0.2, 0.6])
  ]
  |> iterator.from_list
}

Root-mean-square error

stats.rmse(exp_and_pred_vals())
// 0.6957010852370435

Classification accuracy score

stats.score(exp_and_pred_vals())
// 0.6

Import the `Codec` module

import gleam_synapses/codec.{type Codec}

One hot encoding is a process that turns discrete attributes into a list of 0.0 and 1.0.
Minmax normalization scales continuous attributes into values between 0.0 and 1.0.

fn setosa() -> Dict(String, String) {
  [
    #("petal_length", "1.5"),
    #("petal_width", "0.1"),
    #("sepal_length", "4.9"),
    #("sepal_width", "3.1"),
    #("species", "setosa")
  ]
  |> dict.from_list
}

fn versicolor() -> Dict(String, String) {
  [
    #("petal_length", "3.8"),
    #("petal_width", "1.1"),
    #("sepal_length", "5.5"),
    #("sepal_width", "2.4"),
    #("species", "versicolor")
  ]
  |> dict.from_list
}

fn virginica() -> Dict(String, String) {
  [
    #("petal_length", "6.0"),
    #("petal_width", "2.2"),
    #("sepal_length", "5.0"),
    #("sepal_width", "1.5"),
    #("species", "virginica")
  ]
  |> dict.from_list
}

fn dataset() -> Iterator(Dict(String, String)) {
  iterator.from_list([setosa(), versicolor(), virginica()])
}

You can use a Codec to encode and decode a data point.

Create a `Codec` by providing the attributes and the data points

let cdc = codec.new([
      #("petal_length", False),
      #("petal_width", False),
      #("sepal_length", False),
      #("sepal_width", False),
      #("species", True))
    ],
    dataset()
)

The first parameter is a list of pairs that define the name and the type (discrete or not) of each attribute.
The second parameter is an iterator that contains the data points.

Get the json of the codec

let codec_json = codec.to_json(cdc)
// "[
//   {\"Case\" : \"SerializableContinuous\",
//    \"Fields\" : [{\"key\" : \"petal_length\",\"min\" : 1.5,\"max\" : 6.0}]},
//   {\"Case\" : \"SerializableContinuous\",
//    \"Fields\" : [{\"key\" : \"petal_width\",\"min\" : 0.1,\"max\" : 2.2}]},
//   {\"Case\" : \"SerializableContinuous\",
//    \"Fields\" : [{\"key\" : \"sepal_length\",\"min\" : 4.9,\"max\" : 5.5}]},
//   {\"Case\" : \"SerializableContinuous\",
//    \"Fields\" : [{\"key\" : \"sepal_width\",\"min\" : 1.5,\"max\" : 3.1}]},
//   {\"Case\" : \"SerializableDiscrete\",
//    \"Fields\" : [{\"key\" : \"species\",\"values\" : [\"virginica\",\"versicolor\",\"setosa\"]}]}
// ]"

Create a codec by providing its json

codec.from_json(codec_json)

Encode a data point

let encoded_setosa = codec.encode(cdc, setosa())
// [0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 1.0]

Decode a data point

codec.decode(cdc, encoded_setosa)
|> dict.to_list
// [
//   #("species", "setosa"),
//   #("sepal_width", "3.1"),
//   #("petal_width", "0.1"),
//   #("petal_length", "1.5"),
//   #("sepal_length", "4.9")
// ]