gointel

---

gointel is a Go implementation of common machine learning models.

Constraints:

Constraints are used to limit the search space given the initial variables and domains. Local Constraints provide path consistency while Global Constraints provide absolute consistency.

Local Constraints imply that local consistency tends to global consistency. Used for monotone solutions.
Ex: sumLessThan(), localAllDiff()

Global Constraints imply that the entire sequence is needed to know consistency.
Ex: isName(), equals()

Reusable Constraints accomplishes both local and global consistency. These can both prune the search space while exploring and be reused as global constraints at the end.
Ex: min(), max()

n-Queens

(n=16, solutions=14772512, time=14.78s)

Feed Forward Neural Network

(xor problem, training set=4, data={(0,0),(0)}, {(0,1),(1)}, {(1,0),(1)}, {(1,1),(0)})

Sample Code

In your project run:

go mod download github.com/mtresnik/goutils
go mod download github.com/mtresnik/gomath 
go mod download github.com/mtresnik/gointel 

Your go.mod file should look like this:

module mymodule

go 1.23.3

require github.com/mtresnik/gointel v1.1.12

Then in your go files you should be able to run different common models:

package main

import "github.com/mtresnik/gointel/pkg/gointel"

func main() {
	// Common XOR / Hello World for Neural Networks
	layers := []int{2, 4, 1}
	nn := gointel.NewNeuralNetwork(layers, 0.1, 0.9, 0.01) // 0.01 error threshold

	trainingData := []struct {
		inputs  []float64
		targets []float64
	}{
		{[]float64{0, 0}, []float64{0}},
		{[]float64{0, 1}, []float64{1}},
		{[]float64{1, 0}, []float64{1}},
		{[]float64{1, 1}, []float64{0}},
	}

	maxEpochs := 10000
	for i := 0; i < maxEpochs; i++ {
		_, shouldStop := nn.TrainEpoch(trainingData)
		if shouldStop {
			println("Reached error threshold at epoch:", i)
			break
		}
	}
}