https://github.com/cgarciae/neuraljs

Neural Networks in JavaScript, inspired by PyBrain.
https://github.com/cgarciae/neuraljs

Last synced: 10 days ago
JSON representation

Neural Networks in JavaScript, inspired by PyBrain.

• Host: GitHub
• URL: https://github.com/cgarciae/neuraljs
• Owner: cgarciae
• License: mit
• Created: 2014-06-30T03:27:11.000Z (over 10 years ago)
• Default Branch: master
• Last Pushed: 2014-07-18T18:15:11.000Z (over 10 years ago)
• Last Synced: 2024-11-07T09:39:06.989Z (2 months ago)
• Language: JavaScript
• Homepage:
• Size: 357 KB
• Stars: 2
• Watchers: 3
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
NeuralJS

========

Neural Networks in JavaScript, inspired by PyBrain.

### fullConnection

**Signature**


```

fullConnection :: Layer -> Layer -> Layer

```

**Expression**


```

fullConnection (la, lb) == lb

```

**Description**


`fullConnection` connects all neurons in `la` to all neurons in `lb`, and returns `lb`. More specifically, for all `Neuron` `na` 

in `la.neurons`, and for all `Neuron` `nb` in `lb.neurons`, `Weight (na, nb)` belongs to `na.targets` and `nb.sources`. 

This is the classical connection between layers.

**Usage**

Since `fullConnection (la, lb)` returns `lb`, this function admits for chaining

```

fullConnection(la, lb);

fullConnection(lb, lc);

```

is equivalent to

```

fullConnection (fullConnection(la, lb), lc);

```

### fullConnectTo

**Signature**


```

(Layer l) => l.fullConnectTo :: Layer -> Layer

```

**Expression**


```

la.fullConnectTo (lb) == fullConnection (la, lb) == lb

```

**Description**


Same as `fullConnection` but wrapped as a method for `Layer`s.

**Usage**

`fullConnectTo` improves the chaining capacities of `fullConnection`, given that

```

fullConnection(la, lb);

fullConnection(lb, lc);

```

is equivalent to

```

la.fullConnectTo (lb).fullConnectTo (lc);

```