Loss Is the Training Signal
Explain loss as the measurable error signal that training tries to minimize.
Loss is the number that tells the network how wrong it was. A neural network needs a way to score how wrong it is. The loss function turns a prediction and the correct label into a number. For a forecast, loss might grow with the distance between predicted and actual values. For classification, loss often penalizes confident wrong answers more than uncertain wrong answers. The training loop then tries to reduce that number across many examples. Loss is not just a reporting metric. It is the steering signal. Gradient descent uses the loss landscape to decide how to nudge the…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in