Skip to content

AI & ML · Machine Learning Math

Automatic differentiation

76 problems in this skill, each with a solver-verified answer key. Answers are checked the moment you submit — with a hint if you want one and a worked solution afterwards.

What the problems look like

A few statements from this skill. Answers, hints and worked solutions live in the practice session.

  1. Warm-up

    During backprop through an L=7L = 7-layer network with no stored activations, every layer's forward pass must be recomputed, costing 55 multiplications per layer. Checkpointing stores the activations of 33 of the layers, so only the remaining layers are recomputed. How many multiplications does checkpointing save?

  2. Warm-up

    During backprop through an L=10L = 10-layer network with no stored activations, every layer's forward pass must be recomputed, costing 88 multiplications per layer. Checkpointing stores the activations of 55 of the layers, so only the remaining layers are recomputed. How many multiplications does checkpointing save?

  3. Warm-up

    During backprop through an L=10L = 10-layer network with no stored activations, every layer's forward pass must be recomputed, costing 77 multiplications per layer. Checkpointing stores the activations of 22 of the layers, so only the remaining layers are recomputed. How many multiplications does checkpointing save?

Problems shown under their original terms: Generated (template, solver-computed).

Practice this free, right now

Unlimited math practice is on the free plan — no card, no trial clock.

Practice freeLog in and practise this skill

See every Machine Learning Math skill for ai & ml