Backpropagation

When the output is wrong, how much is each weight to blame?

A tiny network, one training example, and every number shown. The table at the bottom is the heart of how all neural networks learn: for each weight, backprop multiplies a few local slopes together to get "how much does the error change if I nudge this weight?"

The network after 2346 updates

w1 = 0.319 w2 = 0.538 w3 = 0.417 w4 = 0.635 w5 = −3.410 w6 = −3.382 w7 = 2.382 w8 = 2.443 i1 0.050 i2 0.100 h1 0.603 h2 0.607 o1 0.029 o2 0.971 target 0.01 target 0.99 bias b1 = 0.35 output bias b2 = 0.6

Train it

0.5
0.000357total error E
0.0290, 0.9713outputs (want 0.01, 0.99)
0.00000.0755870.1491,1730.2241,7600.2982,346updates

1. Forward pass: compute the answer

net_h1 = w1·i1 + w2·i2 + b1 = 0.3188·0.05 + 0.5377·0.10 + 0.35 = 0.419708
h1 = σ(net_h1) = 0.603413    h2 = σ(0.434367) = 0.606916
net_o1 = w5·h1 + w6·h2 + b2 = −3.510039    o1 = 0.029028
net_o2 = w7·h1 + w8·h2 + b2 = 3.520091    o2 = 0.971254
E = ½(t1 − o1)² + ½(t2 − o2)² = 0.000181 + 0.000176 = 0.000356737

What to notice

Output weights are easy. w5's blame is just three numbers multiplied: how wrong o1 is, how steep o1's sigmoid is, and how big h1 was (a big input means the weight mattered more).

Hidden weights reuse that work. w1 can't see the error directly. Its blame arrives through both outputs: output 1's blame times w5, plus output 2's blame times w7. Then the same local slopes. That reuse is what makes backprop fast enough for billions of weights.

The check column agrees. "Nudge and measure" gives the same numbers without any calculus, but it needs two extra forward passes per weight. For an LLM that would be billions of passes per step; backprop needs one.

2. Backward pass: blame for every weight

WeightLocal slopes multiplied together (chain rule)∂E/∂wCheck (nudge)New value = w − 0.50 × ∂E/∂w
w5 (o₁ − t₁) 0.01903 × o₁(1 − o₁) 0.02819 × h₁ 0.60341 0.0003236 0.0003236 −3.40952 → −3.40968254
w6 (o₁ − t₁) 0.01903 × o₁(1 − o₁) 0.02819 × h₂ 0.60692 0.0003255 0.0003255 −3.38216 → −3.38232550
w7 (o₂ − t₂) −0.01875 × o₂(1 − o₂) 0.02792 × h₁ 0.60341 −0.0003158 −0.0003158 2.38200 → 2.38216037
w8 (o₂ − t₂) −0.01875 × o₂(1 − o₂) 0.02792 × h₂ 0.60692 −0.0003176 −0.0003176 2.44310 → 2.44326299
w1 δo₁·w5 + δo₂·w7 −0.00308 × h₁(1 − h₁) 0.23931 × i₁ 0.05000 −0.0000368 −0.0000368 0.31883 → 0.31885039
w2 δo₁·w5 + δo₂·w7 −0.00308 × h₁(1 − h₁) 0.23931 × i₂ 0.10000 −0.0000736 −0.0000736 0.53766 → 0.53770078
w3 δo₁·w6 + δo₂·w8 −0.00309 × h₂(1 − h₂) 0.23857 × i₁ 0.05000 −0.0000369 −0.0000369 0.41747 → 0.41748516
w4 δo₁·w6 + δo₂·w8 −0.00309 × h₂(1 − h₂) 0.23857 × i₂ 0.10000 −0.0000738 −0.0000738 0.63493 → 0.63497032

At the start with learning rate 0.5 you should see w5 → 0.35891648 and w1 → 0.14978072, the same as the original worked example. δo means (o − t)·o(1 − o), the output neuron's blame.

The code that just ran

Everything on this page came from calling this code, which lives at the top of this same file.

Show or hide the source
final class TinyNet
{
    public array $w = ['w1' => 0.15, 'w2' => 0.20, 'w3' => 0.25, 'w4' => 0.30,
                       'w5' => 0.40, 'w6' => 0.45, 'w7' => 0.50, 'w8' => 0.55,
                       'b1' => 0.35, 'b2' => 0.60];   // biases stay fixed, as in the original example
    public array $in = [0.05, 0.10];
    public array $target = [0.01, 0.99];

    public function forward(): array
    {
        $w = $this->w; [$i1, $i2] = $this->in; [$t1, $t2] = $this->target;
        $nh1 = $w['w1'] * $i1 + $w['w2'] * $i2 + $w['b1'];  $h1 = sigmoid($nh1);
        $nh2 = $w['w3'] * $i1 + $w['w4'] * $i2 + $w['b1'];  $h2 = sigmoid($nh2);
        $no1 = $w['w5'] * $h1 + $w['w6'] * $h2 + $w['b2'];  $o1 = sigmoid($no1);
        $no2 = $w['w7'] * $h1 + $w['w8'] * $h2 + $w['b2'];  $o2 = sigmoid($no2);
        $E1 = 0.5 * ($t1 - $o1) ** 2;
        $E2 = 0.5 * ($t2 - $o2) ** 2;
        return compact('i1', 'i2', 't1', 't2', 'nh1', 'h1', 'nh2', 'h2', 'no1', 'o1', 'no2', 'o2', 'E1', 'E2') + ['E' => $E1 + $E2];
    }

    public function loss(): float { return $this->forward()['E']; }

    // BACKPROPAGATION. Each gradient is a product of simple local slopes:
    //   (o − t)      how the error changes with the output
    //   o·(1 − o)    how the sigmoid output changes with its input
    //   h or i       how that input changes with the weight
    // Hidden weights get the SUM of the blame flowing back from both outputs.
    public function gradients(): array
    {
        $f = $this->forward(); $w = $this->w;
        $do1 = ($f['o1'] - $f['t1']) * $f['o1'] * (1 - $f['o1']);   // δ of output 1
        $do2 = ($f['o2'] - $f['t2']) * $f['o2'] * (1 - $f['o2']);   // δ of output 2
        $back1 = $do1 * $w['w5'] + $do2 * $w['w7'];                 // blame reaching h1
        $back2 = $do1 * $w['w6'] + $do2 * $w['w8'];                 // blame reaching h2
        $s1 = $f['h1'] * (1 - $f['h1']);
        $s2 = $f['h2'] * (1 - $f['h2']);

        $g = fn($factors) => ['factors' => $factors, 'value' => array_product(array_column($factors, 1))];
        return [
            'w5' => $g([['(o₁ − t₁)', $f['o1'] - $f['t1']], ['o₁(1 − o₁)', $f['o1'] * (1 - $f['o1'])], ['h₁', $f['h1']]]),
            'w6' => $g([['(o₁ − t₁)', $f['o1'] - $f['t1']], ['o₁(1 − o₁)', $f['o1'] * (1 - $f['o1'])], ['h₂', $f['h2']]]),
            'w7' => $g([['(o₂ − t₂)', $f['o2'] - $f['t2']], ['o₂(1 − o₂)', $f['o2'] * (1 - $f['o2'])], ['h₁', $f['h1']]]),
            'w8' => $g([['(o₂ − t₂)', $f['o2'] - $f['t2']], ['o₂(1 − o₂)', $f['o2'] * (1 - $f['o2'])], ['h₂', $f['h2']]]),
            'w1' => $g([['δo₁·w5 + δo₂·w7', $back1], ['h₁(1 − h₁)', $s1], ['i₁', $f['i1']]]),
            'w2' => $g([['δo₁·w5 + δo₂·w7', $back1], ['h₁(1 − h₁)', $s1], ['i₂', $f['i2']]]),
            'w3' => $g([['δo₁·w6 + δo₂·w8', $back2], ['h₂(1 − h₂)', $s2], ['i₁', $f['i1']]]),
            'w4' => $g([['δo₁·w6 + δo₂·w8', $back2], ['h₂(1 − h₂)', $s2], ['i₂', $f['i2']]]),
        ];
    }

    // Sanity check without any calculus: nudge the weight up and down a hair
    // and see how much the loss moves. Should match backprop almost exactly.
    public function numericalGradient(string $name, float $eps = 1e-5): float
    {
        $keep = $this->w[$name];
        $this->w[$name] = $keep + $eps; $up = $this->loss();
        $this->w[$name] = $keep - $eps; $down = $this->loss();
        $this->w[$name] = $keep;
        return ($up - $down) / (2 * $eps);
    }

    public function step(float $lr): void
    {
        foreach ($this->gradients() as $name => $g) $this->w[$name] -= $lr * $g['value'];
    }
}