Skip to content
Hi, Bot

Hi, Bot · First Principles · Phase 4: Many dimensions

Lesson 34 of 48

Jacobians and vector fields

Derivatives when both the input and the output have many dimensions.

The Jacobian is the local linear approximation of a multi-in, multi-out map — the matrix that says what the function looks like if you stand close enough. Chained Jacobians are literally what backprop multiplies.

Do this

Compute the Jacobian of a map from two dimensions to two by hand and verify every entry numerically. Then push a small square near a point through the map and compare it to the Jacobian's prediction.

The question that unlocks the next lesson

For a map from R^3 to R^2, the Jacobian has what shape?

  • A3×3
  • B2×3
  • C3×2
  • D2×2

Start at lesson 1 and work up to this one

48 lessons, one a day. Answer each lesson's question correctly and the next one opens immediately — nothing here is unlocked by waiting.

By submitting, you agree to our Terms and Privacy Policy.