Loss Computation

Definition

Loss computation measures how wrong the network's output is.

It compares:

network output
target output

and returns a loss value.

Loss computation evaluates the result of the forward pass.

---

Where It Happens

Loss computation happens after the forward pass.

input -> network -> output -> loss

The network first produces an output.

Then the loss function measures how far that output is from what was expected.

Loss is computed after the model produces an output.

---

Simple Example

Suppose the target is 10.

The network outputs 7.

target: 10
output: 7
error: 3

A loss function turns this mismatch into a training signal.

Different loss functions measure mismatch in different ways.

Loss turns output mismatch into a number.

---

Loss Connects Network Structure to Optimization

The network computes output using its parameters.

The loss depends on that output.

So the loss indirectly depends on the parameters.

parameters -> output -> loss

Training changes parameters so future loss becomes smaller.

Loss connects current network behavior to parameter updates.

---

Loss Is Not the Output

The output is what the network predicts.

The loss is a measurement of the output.

output:
model result

loss:
error measurement

They are related, but they are not the same thing.

The model produces output; the loss function judges it.

---

Final Summary

• Loss computation measures output error.
• It compares network output with the target.
• It happens after the forward pass.
• Loss depends on output, which depends on parameters.
• Loss gives optimization a value to reduce.
• Output and loss are related but different.