Index Ai Academy
Foundation · Transposition of formulas

Combining formulas: substitute and eliminate

Putting one formula inside another to remove a letter, and sizing a training run from a time or money budget.

About 35 minutes · Free lesson

What it covers6

Why it matters

Many questions take two formulas at once. You measured the distance around a circle, but the area formula wants the radius. When two formulas share a letter like that, you can remove it: rearrange one formula for the shared letter, then put the result, in brackets, into the other. The new formula goes straight from what you know to what you want. Physics gets v² = u² + 2as this way, and engineers use it to plan training runs and models from a budget.

What you'll be able to do

  • Remove a letter two formulas share: rearrange one formula for it, then substitute the result, in brackets, into the other
  • Eliminate t from v = u + at and s = ut + ½at² to reach v² = u² + 2as, showing every step and what the result can't tell you
  • Tell building a new formula apart from solving two equations for numbers
  • Combine geometry formulas: a circle's area from its circumference, and a rectangle's area from its perimeter and the ratio of its sides
  • Check a combined formula against the chained originals in Python, and size a training run, an API budget and a model's width
Quick check
Have a guess before you start. A circle measures 31.4 cm around. Using π = 3.14, what is its area?
  1. 493 cm²
  2. 774 cm²
  3. 78.5 cm²
  4. 15.7 cm²
Show the answer

78.5 cm²

The distance around is C = 2πr, so r = 31.4 ÷ 6.28 = 5. Then A = πr² = 3.14 × 25 = 78.5. That's two formulas joined by r, and this lesson joins them into one.

The math: substitute to remove a letter

Purpose. Join two formulas that share a letter into one formula without it.

A circle's circumference (the distance around it) is C = 2πr, and its area is A = πr². Both contain the radius r. If you measured C and want A, r is a go-between: needed on the way, not in the answer. To eliminate it (remove it):

  1. Rearrange one formula to make the go-between the subject: r = C/(2π).
  2. Substitute it: put that whole expression in the letter's place in the other formula, in brackets, everywhere the letter appears. A = π(C/(2π))².
  3. Simplify. The result, A = C²/(4π), has no r.
Quick check
From an earlier lesson: put y = 3x into x + y = 20. What is x?
  1. 5
  2. 6.67
  3. 17
  4. 15
Show the answer

5

x + 3x = 20, so 4x = 20 and x = 5, then y = 15. There, substitution ended in numbers. Here, it ends in a formula.

Substituting is the chaining from Chaining and undoing functions: the first formula's output, r, is the second one's input. Watch C = 31.4 go through both, with π as 3.14, then come back.

Try it now

Change C and press Use these numbers: the area always equals C² ÷ (4 × 3.14), so one formula does the work of two.

Symbols and formula. The circle's three steps, written out:

C=2πr  ⇒  r=C2π  ⇒  A=π(C2π)2=πC24π2=C24πC = 2\pi r \;\Rightarrow\; r = \frac{C}{2\pi} \;\Rightarrow\; A = \pi\left(\frac{C}{2\pi}\right)^2 = \frac{\pi C^2}{4\pi^2} = \frac{C^2}{4\pi}

In plain words: square the distance around, then divide by 4π. For C = 31.4: 985.96 ÷ 12.56 = 78.5, the same as the route through r = 5.

Named mistakes.

  • Dropping the brackets. π × C²/2π squares only the top and simplifies to C²/2: 493 for this circle, not 78.5.
  • Leaving a copy behind. Substitute everywhere the letter appears, or it isn't eliminated.
  • Removing the wrong letter. Rearrange for the go-between, never for a letter you want in the answer.
Quick check
The tent shop ropes off a rectangle whose length is twice its width: l = 2w, and P = 2(l + w). What is w in terms of P alone?
  1. w = P/3
  2. w = P/4
  3. w = P/6
  4. w = P/2 − l
Show the answer

w = P/6

Substitute l = 2w: P = 2(2w + w) = 6w, so w = P/6. A 36 m rope gives w = 6 m and l = 12 m.

Building a formula, not solving for numbers. In Two equations, two unknowns, the other letters had numbers, so eliminating one unknown left a number. Here the letters stay letters, so eliminating leaves a formula. With a 36 m rope, l = 2w and 2(l + w) = 36 are two equations: w = 6 and l = 12. Leave P as a letter, and the same substitution builds w = P/6 and A = lw = 2w² = 2(P/6)² = P²/18: the area for every rope. For 36 m, 1296 ÷ 18 = 72, which is 12 × 6.

Eliminating time. Two formulas describe anything speeding up steadily, like a cyclist rolling downhill: v = u + at and s = ut + ½at². u is the starting speed, v the speed after t seconds, a the acceleration (the speed gained each second), and s the distance covered. To link speed and distance without t:

t=v−uas=u(v−ua)+12a(v−ua)2=u(v−u)a+(v−u)22a2as=2u(v−u)+(v−u)2=2uv−2u2+v2−2uv+u2=v2−u2v2=u2+2as\begin{aligned} t &= \frac{v - u}{a} \\ s &= u\left(\frac{v - u}{a}\right) + \frac{1}{2}a\left(\frac{v - u}{a}\right)^2 = \frac{u(v - u)}{a} + \frac{(v - u)^2}{2a} \\ 2as &= 2u(v - u) + (v - u)^2 \\ &= 2uv - 2u^2 + v^2 - 2uv + u^2 = v^2 - u^2 \\ v^2 &= u^2 + 2as \end{aligned}

Line by line: make t the subject of the first formula (a can't be 0). Substitute it, in brackets, for both copies of t. Multiply every term by 2a to clear the fractions. Expand: the 2uv terms cancel. Add u² to both sides.

Quick check
Substituting t = (v − u)/a into ½at² gives what?
  1. (v − u)/2
  2. (v − u)²/(2a)
  3. (v − u)²/2
  4. a(v − u)²/2
Show the answer

(v − u)²/(2a)

½a × (v − u)²/a² = (v − u)²/(2a): the a on top cancels one of the two below. Check with u = 4, v = 8 and a = 0.5: t = 8, so ½at² = ½ × 0.5 × 64 = 16, and (v − u)²/(2a) = 16 ÷ 1 = 16.

Check it, and see what it can't tell you. A cyclist starts down a hill at u = 4 m/s and gains a = 0.5 m/s each second. After 48 m, v² = 16 + 2 × 0.5 × 48 = 64, so v = 8. Through t: 8 = 4 + 0.5t gives t = 8, and s = 4 × 8 + ½ × 0.5 × 8² = 48. The same.

Removing t also removed information: v = ±√(u² + 2as). A ball thrown up at 20 m/s, with a = −10, is 15 m up twice: v² = 400 − 300 = 100, so v = 10 rising and −10 falling. Only t knew which. And v² can't be negative, so when u² + 2as is below 0 there is no real v: the ball never reaches 25 m, since 400 − 500 = −100.

In your own words
In your own words: how is eliminating t here different from eliminating y in two equations, two unknowns?

Write a sentence or two, then compare it with a good answer in the lesson.

In Python: chain the originals, then compare

Purpose. A combined formula is new algebra, and new algebra can slip. Check it by working out each answer two ways: through the go-between, with formulas you trust, and with the combined formula.

import math

def area_chained(C):
    r = C / (2 * math.pi)
    return math.pi * r ** 2

def area_combined(C):
    return C ** 2 / (4 * math.pi)

for C in [31.4, 2.5, 100]:
    print(C, round(area_chained(C), 6), round(area_combined(C), 6))
It runs right here in your browser, with nothing to install.

The columns agree for every C. With the full π from math, 31.4 gives 78.46 rather than 78.5. The middle row is a tree trunk: a tape measures 2.5 m around, so its cross-section is about 0.497 m². Foresters work it out this way, since a tape goes around a trunk more easily than a ruler goes through it.

Worked example

Worked example: v² = u² + 2as, checked through t

Pick a time, run both original formulas, and compare v² with u² + 2as. The cyclist again: u = 4 and a = 0.5.

u = 4
a = 0.5
for t in [0, 2, 8, 3.5]:
    v = u + a * t
    s = u * t + a * t ** 2 / 2
    print(t, s, v ** 2, u ** 2 + 2 * a * s)
In the lesson you can change this and run it.

Each row is one moment of the ride. At t = 8 she has gone 48.0 m, and the last two columns both say 64.0, as found by hand. They match at every t, even t = 0. That's the point of eliminating t: the new formula holds whatever t was, so you never need to know it.

Quick check
What does this print?
import math

C = 10
r = C / 2 * math.pi
print(round(r, 2), round(C / (2 * math.pi), 2))
  1. 1.59 1.59
  2. 1.59 15.71
  3. 15.71 1.59
  4. 15.71 15.71
Show the answer

15.71 1.59

Python reads C / 2 * math.pi left to right: 10 ÷ 2 = 5, then 5 × π ≈ 15.71. Only the brackets in C / (2 * math.pi) put the whole 2π on the bottom, giving 1.59.

Quick check
You derived A = C²/(4π). Which test would catch a slip in it?
  1. Compare with π(C/(2π))² for several C
  2. Run it once, for C = 0
  3. Check that r no longer appears
  4. Print A for C = 31.4 and look
Show the answer

Compare with π(C/(2π))² for several C

The chained route uses formulas you trust, so it's a fair answer to compare against. Several inputs, including a decimal, leave a slip nowhere to hide.

Drill

Fix the combined formulas

area_combined(C) should match the chained circle formulas, and rect_area(P) should give the area of a rectangle whose length is twice its width. Each has one slip. Fix them so both match their chains.

import math

def area_chained(C):
    r = C / (2 * math.pi)
    return math.pi * r ** 2

def area_combined(C):
    return C ** 2 / 4 * math.pi

def rect_chained(P):
    w = P / 6
    return 2 * w * w

def rect_area(P):
    return P ** 2 / 16

print(area_chained(31.4), area_combined(31.4), rect_chained(36), rect_area(36))
In the lesson, your code is checked as you write it.
In your own words
Why compare a combined formula with the chain of original formulas, instead of rereading the algebra?

Write a sentence or two, then compare it with a good answer in the lesson.

In AI and ML: budgets for time, money and memory

Purpose. Planning a training run or a product usually takes two formulas joined by a go-between. Combine them, then rearrange for the number the budget controls.

Time. A run takes T = steps × s, with s seconds per step, and steps = epochs × n ÷ batch, from Rearranging formulas. Substitute: T = epochs × n × s ÷ batch. For the data a time budget allows: n = T × batch ÷ (epochs × s).

Money. A language model's bill is cost = tokens × price, where tokens are the small pieces of text it reads and writes, and tokens = requests × tokens per request. So requests = budget ÷ (tokens per request × price). Prices are quoted per million tokens, so the code divides that price by 1e6 (a million) to get the price of one token.

def examples_for_time(T, epochs, batch, sec_per_step):
    return T * batch / (epochs * sec_per_step)

def requests_for_budget(budget, tokens_per_request, price_per_million):
    return budget / (tokens_per_request * price_per_million / 1e6)

n = examples_for_time(7200, 3, 32, 0.25)
print(n, 3 * n / 32 * 0.25)
print(round(requests_for_budget(90, 1500, 3)))

The trail club's photo model gets 7,200 s (2 hours), 3 epochs, batches of 32 and 0.25 s a step: 307,200 photos, and the check gives back 7,200 s. Its help bot uses 1,500 tokens a request. At $3 per million tokens (an invented price for this example: real price lists differ, and charge input and output tokens differently), $90 buys 20,000 requests.

Quick check
A run takes T = epochs × n × s ÷ batch. With 4 epochs, batches of 64 and 0.5 s a step, how many examples fit in 3,600 s?
  1. 1,843,200
  2. 115,200
  3. 7,200
  4. 460,800
Show the answer

115,200

n = T × batch ÷ (epochs × s) = 3,600 × 64 ÷ 2 = 115,200. Check: 4 × 115,200 ÷ 64 = 7,200 steps, and 7,200 × 0.5 s = 3,600 s.

Memory. A model's weights take M = params × b bytes, with b bytes per number: 4 for the 32-bit floats from Scientific notation, 2 for 16-bit. In a standard transformer (the design inside language models), params ≈ 12 × layers × d². The width d is how many numbers describe each token inside the model. The 12 counts about 4d² weights in each block's attention part and 8d² in its feed-forward part. Embeddings (the table that turns tokens into numbers) and small extras are left out, so it's an approximation.

Substitute: M = 12 × layers × d² × b. d is squared, so freeing it takes a square root, the positive one, since a width can't be negative: d = √(M ÷ (12 × layers × b)). For 6 GB (6 × 10⁹ bytes), 20 layers and 4 bytes: d = √(6 × 10⁹ ÷ 960) = √6,250,000 = 2,500.

Quick check
What does this print?
def requests_for_budget(budget, tokens_per_request, price_per_million):
    return budget / (tokens_per_request * price_per_million / 1e6)

print(round(requests_for_budget(60, 1200, 2.5)), round(requests_for_budget(60, 2400, 2.5)))
  1. 0 0
  2. 10000 20000
  3. 20000 40000
  4. 20000 10000
Show the answer

20000 10000

At an invented $2.50 per million tokens, 1,200 tokens cost $0.003 a request, and $60 ÷ 0.003 = 20,000. Twice the tokens per request halves the requests: 10,000.

Drill

Width and layers from a memory budget

With M = 12 × layers × d² × b, write width_for_memory(M, layers, bytes_per), the width d that fills M bytes, and layers_for_memory(M, d, bytes_per), the number of layers that fills it.

def width_for_memory(M, layers, bytes_per):
    return M / (12 * layers * bytes_per)

def layers_for_memory(M, d, bytes_per):
    return M / (12 * d * bytes_per)

print(width_for_memory(6e9, 20, 4), layers_for_memory(6e9, 2500, 4))
In the lesson, your code is checked as you write it.
In your own words
Why does 4 times the memory buy only twice the width?

Write a sentence or two, then compare it with a good answer in the lesson.

Goes deeper in

Practice what you just read

  • 3 code examples you can edit and run
  • a practice exercise that checks your code
  • a 5-question quiz
  • 9 review cards for spaced repetition

Try a quiz question

Eliminate t from v = u + at and s = ut + ½at².

  1. v² = u² + 2as
  2. v = u + 2as
  3. v² = u² + as
  4. v = u² + 2as

Answer it at the end of the lesson to see if you're right, and why.

Before this lesson

Keep learning

Where this lesson leads.

575 lessons in thirteen phases. 176 are free without an account, and a free account opens the rest. See every lesson.

Foundation · about 96 hours of study

Foundation math

From whole numbers to the math inside a neural network: each idea by hand, in Python, and in AI 118 lessons All free
  1. Numbers and place value FreeWhat a digit's place is worth, the number line, and Python as a calculator that never tires.36 min
  2. Multiplying and dividing: groups FreeEqual groups, quotient and remainder, and cutting training data into batches.35 min
  3. Order of operations, and names for numbers FreeBrackets, powers, then multiply and divide, then add and subtract; variables hold numbers.36 min
  4. Numbers below zero FreeThe number line through zero, sign rules, distance from zero, and an error's direction and size.37 min
  5. Fractions: equal parts, equal fractions, adding FreePieces of a whole, equal fractions, and re-cutting to add; accuracy as a fraction.45 min
  6. Multiplying and dividing fractions FreeA share of a share, dividing by flipping, and why chances in a row get small.35 min
  7. Decimals and rounding FreeFractions in place value, rounding, and why 0.1 + 0.2 isn't exactly 0.3.36 min
  8. Percentages FreePer hundred, percent change, percentage points, and the accuracy and error rates of models.35 min
  9. Ratios, rates and proportion FreeSharing in a ratio, rates per unit, scaling, and splitting data 80/10/10.35 min
  10. Powers and roots FreeRepeated multiplying, the laws of exponents, square roots, and squared error.43 min
  11. Very big and very small numbers FreePowers of ten, e-notation like 3e-4, orders of magnitude, and a model's size in memory.36 min
  12. Bits, bytes and floats: how numbers are stored FreeBase 2, what n bits can hold, a float as sign, power and digits, and a tensor's memory.35 min
  13. Review: find the slip (numbers) FreeBroken calculations to fix, mixed practice from the whole module, and its formula sheet.34 min
  14. Letters for numbers FreeVariables, terms and like terms, substitution, and a model's formula as a Python function.41 min
  15. Expanding and factoring FreeBrackets as areas, pulling out common factors, and a squared error expanded into a bowl.41 min
  16. Solving equations: keep the balance FreeThe same move on both sides, undoing steps in reverse, and running a model's formula backward.41 min
  17. Rearranging a formula FreeMaking any letter the subject, round-trip checks, and planning a training run.36 min
  18. Inequalities and true-or-false FreeRanges, flipping the sign, if and elif, and the thresholds, clips and ReLU inside a model.42 min
  19. Functions: a rule from input to output Freef(x) notation, tables and graphs, domain and range, loops over lists, and loss as a function.42 min
  20. Chaining and undoing functions Freeg(f(x)), inverses that undo, scaling and unscaling, and a network as a chain.42 min
  21. Straight lines: slope and intercept FreeRise over run, the line through two points, and a model's weight and bias as slope and intercept.42 min
  22. Two equations, two unknowns FreeSubstitution and elimination, where lines cross, and why models are fitted, not solved.42 min
  23. Quadratics: bowls and their lowest point FreeRoots by factoring and by formula, the discriminant, the vertex, and the best weight of a model.42 min
  24. Exponentials: growth and decay FreeThe same factor each step, the number e, and decaying learning rates.43 min
  25. Logarithms: the power question FreeTo what power?, the laws of logs, and adding log-probabilities instead of multiplying.45 min
  26. Sequences and Σ FreeArithmetic and geometric sequences, Σ as a loop that adds, and mean squared error.40 min
  27. Weighted and moving averages FreeWeights that add to 1, a window average, the exponential moving average, and smoothing a loss curve.35 min
  28. Sets, and reading the symbols FreeSets, union and intersection, ℝⁿ and f : ℝⁿ → ℝᵐ, and reading a paper's first equation.41 min
  29. Review: find the slip (algebra) FreeBroken algebra and code to fix, mixed practice from the whole module, and its formula sheet.34 min
  30. Transposing step by step: undo in reverse order FreeA formula's layers, the opposite of each, peeling from the outside in, conditions, and round-trip checks.35 min
  31. Formulas with fractions: clear them first FreeClearing denominators, cross-multiplying, a letter in the bottom, and reciprocal formulas like 1/R = 1/R₁ + 1/R₂.35 min
  32. When the letter appears twice: collect and factor FreeGathering the letter on one side, factoring it out, dividing by the bracket, and rearranging precision and F1.35 min
  33. Squares, roots and powers FreeSquare roots with ±, squaring to undo a root, cube and nth roots, and the √d inside attention.40 min
  34. When the letter is squared and plain: quadratics FreeRearranging into ax² + bx + c = 0, the formula, completing the square, and choosing the root that fits.35 min
  35. Letters in the power: exponentials and logarithms FreeTaking logs to bring a power down, exponentiating to free a letter, and the logit that undoes the sigmoid.40 min
  36. Rearranging inequalities: when the sign flips FreeThe four moves on an inequality, negatives and reciprocals that flip it, logs below 1, and sample sizes.35 min
  37. Inverse functions, lines and scaling FreeUndoing a function by making x the subject, domains that allow it, line forms, and unscaling predictions.40 min
  38. Sequences and sums: find n, a, d and r FreeEvery letter of the arithmetic and geometric formulas as the subject, and an average's window from its decay.40 min
  39. Combining formulas: substitute and eliminate FreePutting one formula inside another to remove a letter, and sizing a training run from a time or money budget.35 min
  40. Review: rearrange anything (algebra) FreeBroken rearrangements to fix, mixed practice from the whole algebra block, a strategy chart, and its formula sheet.35 min
  41. Coordinates and distance FreePoints as (x, y), Pythagoras, distance in any number of coordinates, and nearest neighbors.42 min
  42. Angles, degrees and radians FreeMeasuring a turn, converting units, arcs, and the angle between two arrows.36 min
  43. Sine and cosine FreeAcross and up: right triangles, the unit circle, and directions and rotations.42 min
  44. Waves: height, speed and shift FreeAmplitude, period and phase, sampling a wave, and position codes in transformers.37 min
  45. Review: find the slip (shapes and waves) FreeBroken geometry and code to fix, mixed practice from the module, and its formula sheet.29 min
  46. Angles, triangles and waves: rearranging with sin, cos and tan FreeSides and angles from SOH CAH TOA, principal values, arcs, polar coordinates, and a wave's time and period.45 min
  47. Vectors: lists of numbers that point FreeA list, a point and an arrow; components, dimension, NumPy arrays, and data as vectors.42 min
  48. Adding and scaling vectors FreeComponent-wise sums, scalar multiples, mixes, and the weight update w − ηg.43 min
  49. Length and distance: norms FreeL2, L1 and max norms, unit vectors, distance, and the norms inside losses and clipping.42 min
  50. The dot product FreeMultiply in pairs and add; a·a as squared length; a @ b; one neuron as a dot product.43 min
  51. Angle, cosine similarity and projection FreeDirection without length, perpendicular vectors, shadows, and search by meaning.45 min
  52. Perpendicular directions: bases and coordinates FreeOrthonormal directions, coordinates as dot products, rebuilding a vector, and taking a direction out.35 min
  53. Matrices: tables of numbers FreeRows, columns and entries; two-dimensional arrays and slices; a data set as a matrix.41 min
  54. Adding, scaling and transposing matrices FreeEntry-by-entry sums and scalings, the transpose, mean images, and tables turned around.36 min
  55. A matrix times a vector FreeEach row dotted with the vector, the size rule, A @ x, and a layer acting on one example.42 min
  56. Matrix multiplication FreeRow meets column, the size rule, order matters, and a batch through a layer: X W + b.45 min
  57. Identity, diagonal, triangular and symmetric FreeFour patterns and their jobs: rescaling, distance tables, and the causal mask.37 min
  58. Matrices as movements of space FreeStretches, flips, shears and rotations; columns as landing spots; chaining; a layer moving data.42 min
  59. Determinant and inverse FreeThe area factor ad − bc, singular matrices, undoing a matrix, and redundant features.42 min
  60. Solving systems by elimination FreeAugmented matrices, row moves, back-substitution, the three outcomes, and fitting weights.45 min
  61. Independence, span and rank FreeRedundant vectors, span, rank, outer products, and low-rank updates like LoRA.42 min
  62. Eigenvectors: directions that only stretch FreeAv = λv, checking candidates, np.linalg.eig, and the main direction of a data cloud.41 min
  63. SVD and PCA: the important directions first FreeA matrix as pieces by size, the best rank-k copy, compression, and PCA maps of data.45 min
  64. Tensors, shapes and broadcasting FreeAxes and shapes, reshape, axis=, broadcasting, and shapes through a small model.43 min
  65. One-hot vectors and lookup tables: embeddings FreeA one-hot row picks a table's row, a sentence looked up as vectors, and a score for every word.35 min
  66. Sliding dot products: convolution FreeA small pattern slid along a list and over a grid, output sizes, stride and padding, and shared weights.40 min
  67. Review: find the slip (vectors and matrices) FreeShape slips to fix, mixed practice from the module, and its formula sheet with shapes.36 min
  68. Vector equations: rearranging with arrows FreeSolving for a vector or a scale, recovering a gradient from two weight snapshots, and an angle from a dot product.40 min
  69. Matrix equations: multiply by the inverse, on the correct side FreeNo dividing by a matrix: inverses from the left or right, order, transposes, and the normal equations.40 min
  70. Rates of change FreeChange over change, secants, rates from readings with np.diff, and reading a loss curve.41 min
  71. Limits: what a value heads toward FreeApproaching from both sides, holes, corners and jumps, epsilon, and judging activations.37 min
  72. The derivative: slope at a point FreeShrinking secants, f′(x) and dy/dx, Δy ≈ f′Δx, and which way a weight should move.45 min
  73. Derivative rules FreeConstants, powers, multiples, sums, eˣ and ln x, checked numerically, and a squared error's slope.41 min
  74. The product rule and the chain rule FreeProducts and chains of functions, forward then backward, and one neuron's dL/dw.45 min
  75. Lowest points and curvature Freef′ = 0, the second-derivative test, tables, local and global minima, and convex losses.43 min
  76. Standing in for a curve near a point FreeTangent lines, a curvature term, growing error, and why learning rates must be small.37 min
  77. Partial derivatives FreeOne input at a time, ∂, nudging in Python, and the two slopes of a line's loss.41 min
  78. The gradient FreeAll the slopes as a vector, steepest climb and descent, a numerical gradient, and the loss gradient.42 min
  79. Gradient descent FreeStepping against the gradient, the learning rate, a training loop, and fitting a line.45 min
  80. Faster descent: momentum and Adam FreeMomentum and Adam by hand, the bias correction, and three optimizers down one narrow valley.40 min
  81. The chain rule with many paths: backpropagation by hand FreeGraphs, multiply along paths and add across them, gradient checks, and a 2-2-1 network by hand.45 min
  82. Automatic differentiation: how PyTorch finds every slope FreeA tape of steps walked backward, and requires_grad, backward(), .grad, zero_grad() and no_grad().40 min
  83. Jacobians and Hessians FreeSlopes of vector functions, a layer's Jacobian, second partials, and bowls, domes and saddles.42 min
  84. Integrals: adding up thin slices FreeStrip and trapezoid sums, summing undoes differencing, probability as area, and AUC.43 min
  85. Review: find the slip (calculus) FreeCalculus slips to fix, mixed practice from the module, and its formula sheet.36 min
  86. Rearranging in calculus: slopes, steps and constants FreeSetting a slope to zero, Newton's step, the chain rule turned around, constants from a point, and learning rates.45 min
  87. Counting: choices, orders and groups FreeThe counting principle, factorials, ordered picks and groups, and why attention's cost grows with the square of the length.43 min
  88. Probability: from counts to chances FreeEqually likely outcomes, the complement, distributions, seeded simulation, and a next-word model.43 min
  89. Combining events: and, or, independent FreeThe or rule, independence, probability trees, drawing with and without replacement, and chains of steps.42 min
  90. Conditional probability and Bayes' rule FreeConditionals from tables, why P(A | B) isn't P(B | A), Bayes' rule, and precision and recall.45 min
  91. Random variables and distributions FreeUncertain numbers, the probability of a range, sampling by slices, and sampling versus argmax.42 min
  92. Expected value, variance and standard deviation FreeProbability-weighted averages, spread in the data's own units, and a training loss as an expected value.42 min
  93. Bernoulli, binomial, uniform and normal FreeYes-or-no trials, counts of yeses, the 68-95-99.7 rule, z-scores, and an accuracy's error bar.45 min
  94. Two variables at once: covariance and correlation FreeJoint tables and marginals, covariance, correlation, the covariance matrix, and features that carry the same information.45 min
  95. Review: find the slip (probability) FreeProbability slips to fix, mixed practice from the module, and its formula sheet.36 min
  96. Probability formulas: solve for the missing chance FreeConditionals, Bayes' rule for the prior, the or rule, expected values, variance, and z-scores turned around.45 min
  97. Describing data: center, spread and scale FreeMean, median and mode, variance and standard deviation, percentiles, outliers, and standardizing features.42 min
  98. Samples, the law of large numbers and the central limit theorem FreeSamples and populations, the standard error σ/√n, the bell of averages, and error bars on test scores.43 min
  99. Is the difference real? Comparing two models FreePaired differences, a bootstrap interval, what a p-value is and isn't, and why many tries make lucky wins.40 min
  100. Maximum likelihood FreeLikelihoods as products, log-likelihoods as sums, grid search, underflow, and the negative log-likelihood loss.45 min
  101. Linear regression: the best line FreeResiduals, least squares by formula, three roads to one line, the matrix form, and train versus test error.45 min
  102. Logistic regression: a line that answers yes or no FreeThe sigmoid, the log loss, its slope p − y, a training loop, and choosing a threshold.45 min
  103. Scoring a classifier: confusion table, precision, recall, F1 FreeThe four cells, accuracy, precision, recall and F1, and what moving a threshold does to each.35 min
  104. Softmax: scores into probabilities FreeExponentiate and share out, shifting and temperature, the safe softmax, and a classifier's last step.43 min
  105. Surprise, entropy and cross-entropy FreeSurprise in bits, entropy, cross-entropy and KL divergence, perplexity, and a language model's loss.45 min
  106. KL divergence: how far one distribution is from another FreeKL by hand, cross-entropy as entropy plus KL, why it isn't symmetric, and a soft-target loss.35 min
  107. Review: find the slip (statistics and information) FreeStatistics slips to fix, mixed practice from the module, and its formula sheet.38 min
  108. Statistics and information: n, z, logits and bits FreeSample sizes from error bars, inputs from probabilities, logit gaps from softmax, and chances from bits and losses.40 min
  109. One neuron, start to finish FreeWeighted sum, activation, loss, slopes and update, by hand and in code, and what one straight line can and can't split.45 min
  110. A layer and a small network as matrices FreeTwo layers on a batch, the shape of every matrix, adding a hidden unit, and XOR solved.45 min
  111. Training a small network FreeThe backward pass in matrix form, gradient checks, a full training loop, and reading loss curves.45 min
  112. The same network in PyTorch Freenn.Sequential with the NumPy weights loaded, the same loss and slopes, then SGD and mini-batches.45 min
  113. Attention as matrix math FreeScaled dot-product scores, the causal mask, a row-by-row softmax, and blending value vectors.45 min
  114. Residual connections and layer norm FreeAdding the input back and why its slope keeps learning alive, layer norm by hand, and add and norm.40 min
  115. A transformer block, start to finish FreeA sentence through embeddings, attention, add and norm, a small network and next-word scores, every shape named.45 min
  116. Planning a model: rearranging the formulas of a working network FreeLayer widths from parameter counts, sizes from memory and compute, the attention scale, and a training plan.40 min
  117. Final test: transpose any formula FreeSlips to fix from every topic, mixed practice of every kind of rearranging, the strategy chart and the full sheet.45 min
  118. Capstone: a sign reader from the formulas FreeBuild, train and test a trail-sign classifier with NumPy alone, and report its accuracy with an error bar.45 min
Part of a free course 118 lessons

The Foundation Math For Ai & ML: from arithmetic to attention

Learn the math behind AI from arithmetic up: numbers, algebra, vectors and matrices, calculus, probability and statistics, each shown as numbers moving and checked in Python, ending with a network you build and train from the formulas.

Part of a free course 19 lessons

Transposition of Formulas: algebraic manipulation in every topic

Learn to make any letter of any formula the subject: layers undone in reverse order, fractions, a letter that appears twice, powers, roots, quadratics, logs, inequalities, trigonometry, vectors, matrices, calculus, probability, statistics and the formulas inside AI models, each checked in Python.

Never trained an AI? Teach a small one to write, right in your browser. It takes about 20 seconds and needs no account. Teach it to write