JA EN
Student mode › Grade 12
★ MEMBER ·Grade 12

Derivatives

Where this is used

A derivative answers one question: how fast is this changing right now, at this exact instant? That single idea turns out to be the most useful tool in modern science and engineering. It is how a neural network figures out which way to nudge its weights so it makes fewer mistakes, how physics writes down the motion of anything that moves, and how a company finds the price that makes it the most money. Once you can differentiate, you can find the top of a curve, the bottom of a cost, and the direction of any change.

Start gently

Differentiation is the tool for finding the slope at a single instant. It is the steepness of the speed graph you drew in primary school, and the rate of change of a linear function from junior high, sharpened down to one point on a curve.

Key points

The definition of a derivative

f'(x) is the slope of the function at the single point x — its steepness at that instant. You get there by taking the average slope over a small step Δx and then letting that step shrink toward zero.

Differentiating a power of x

This is the formula you will use more than any other. To differentiate x to the power n, bring the exponent down in front as a multiplier and then drop the exponent by one.

What the derivative tells you (slope, maximums, and minimums)

The sign of the derivative reads the shape of the graph: f'(x) > 0 means the function is climbing, f'(x) < 0 means it is falling, and f'(x) = 0 marks a spot where it has levelled off — a candidate for a peak or a valley. Setting the derivative of a profit function to zero and solving is exactly how "maximize profit" is done in mathematics.

The product rule and the quotient rule

These two rules handle a product or a quotient of functions, which comes up constantly in practice — differentiating a revenue function built as price times quantity, for example.

Differentiating a function inside a function (the chain rule)

Split the expression into an outer function and an inner one, differentiate each, and multiply the results. Backpropagation — the algorithm that trains every neural network — is nothing more than this rule applied layer after layer.

See it drawn

The slope at one point on a curve is the tangent line
xyy = x²tangent at x = 2: y = 4x − 4(2, 4)

Since y' = 2x, the slope at x = 2 is 4. The tangent is the line through that point with that slope.

Jobs that use this

AI researcher$220k

Control systems engineer$118k

Economist / econometrician$130k

§

Members-only from here

The practice questions and full career details are for members. $4.99/mo, cancel anytime.

Comments

Sign in to comment

All grades