Backpropagation
The error correction mechanism in neural networks.
- Back-Propagation is the phenomenon of a neural network-based function correcting itself
- It looks at the output that its analysis function has calculated and compares it to the actual output
- It then uses the differences to adjust the weights of its analysis function
- Those differences are known as a "loss function"
- Summarily, Back-Propagation explains a process for error correction