Deep Learning / 04
Forward Propagation
Follow inputs through weighted transformations and activations to a network output.
Prediction moves forward. Learning begins when the error moves back.
In this gradient-based training picture, backpropagation computes gradients and the optimizer uses them to update parameters.
This is a knowledge index. Published titles open their articles; the remaining entries outline planned coverage.
Subtopics
- 04.01
What is Forward Propagation?
Planned guide: What is Forward Propagation. Focus on intuition, assumptions, and a worked example.
- 04.02
Input Transformation
Planned guide: Input Transformation. Focus on intuition, assumptions, and a worked example.
- 04.03
Weighted Sum
Planned guide: Weighted Sum. Focus on intuition, assumptions, and a worked example.
- 04.04
Activation
Planned guide: Activation. Focus on intuition, assumptions, and a worked example.
- 04.05
Layer-to-Layer Flow
Planned guide: Layer-to-Layer Flow. Focus on intuition, assumptions, and a worked example.
- 04.06
Output Computation
Planned guide: Output Computation. Focus on intuition, assumptions, and a worked example.
- 04.07
Batch Processing
Planned guide: Batch Processing. Focus on intuition, assumptions, and a worked example.
- 04.08
Tensor Shapes
Planned guide: Tensor Shapes. Focus on intuition, assumptions, and a worked example.
- 04.09
Forward Pass Example
Planned guide: Forward Pass Example. Focus on intuition, assumptions, and a worked example.
- 04.10
Computational Graph of a Forward Pass
Planned guide: Computational Graph of a Forward Pass. Focus on intuition, assumptions, and a worked example.