# Convolution

> ML Prep practice set.

- Stable ID: `convolution`
- Area: machine-learning
- Timebox: 55 minutes

Compute convolution geometry and parameter counts accurately.

## Instructions

1. Write the output-size formula before substituting values.

## Ordered items

1. [Convolution output shape and parameters](https://mlprep.iwase.dev/item/original-ml-convolution/?set=convolution) — `original-ml-convolution` (easy, 6 min)
2. [Convolutional representation and detector design](https://mlprep.iwase.dev/item/mit-6-036-s2018-final-q9/?set=convolution) — `mit-6-036-s2018-final-q9` (medium, 18 min)
3. [CNN shapes, pooling, and backpropagation](https://mlprep.iwase.dev/item/mit-6-036-s2019-final-q4/?set=convolution) — `mit-6-036-s2019-final-q4` (hard, 25 min)

## Completion

Compute shapes and parameter counts without framework assistance.

## Next recommended practice

- [Dimensionality Reduction](https://mlprep.iwase.dev/set/dimensionality-reduction.md)