Mathematics —
From Fundamentals
to Advanced Topics
Taught by a PhD-level mathematician who uses math daily in nuclear engineering and ML research.
Algebra & Pre-Calculus
Pre-calculus builds the toolkit for everything that follows. Mastery here pays dividends through Calc III and beyond. Students who struggle in calculus are often missing a pre-calc concept, not a calculus one.
Topics Covered
- Functions: domain, range, composition, inverse
- Polynomials: factoring, synthetic division, remainder theorem
- Rational functions: asymptotes, holes, end behavior
- Exponentials and logarithms: change of base, natural log
- Trigonometry: unit circle, identities, inverse trig
- Complex numbers: rectangular and polar form, Euler's formula
Key Relationships to Memorize
- ln(e^x) = x and e^(ln x) = x
- sin²(θ) + cos²(θ) = 1
- sin(2θ) = 2sin(θ)cos(θ)
- cos(2θ) = cos²(θ) − sin²(θ)
- e^(iπ) + 1 = 0 (Euler's identity)
- log_b(xy) = log_b(x) + log_b(y)
Common Errors
- Log rules: log(x + y) ≠ log(x) + log(y). The product rule is log(xy) = log x + log y
- Trig signs by quadrant: memorize ASTC (All, Sin, Tan, Cos) for signs in quadrants I–IV
- Function composition direction: f(g(x)) means apply g first, then f — not the other way
- Inverse functions: f²(x) means [f(x)]², but f−¹(x) means the inverse function, not 1/f(x)
Show Full Solution
Show Full Solution
Show Full Solution
Limits & Derivatives
Calculus I introduces the two central ideas of calculus: the limit (what a function approaches) and the derivative (rate of change). These concepts underpin all of physics, engineering, and modern data science.
d/dx[f·g] = f'g + fg' (Product Rule)
d/dx[f/g] = (f'g − fg') / g² (Quotient Rule)
d/dx[f(g(x))] = f'(g(x)) · g'(x) (Chain Rule)
d/dx[e^x] = e^x d/dx[ln x] = 1/x d/dx[sin x] = cos x
Key Concepts
- Formal limit definition: lim[x→a] f(x) = L
- L'Hôpital's rule for 0/0 and ∞/∞ indeterminate forms
- Continuity: limit exists, f(a) is defined, and they are equal
- Implicit differentiation for equations not solved for y
- Related rates: differentiate both sides with respect to time
- Critical points, concavity, and the First and Second Derivative Tests
Applied Uses Dr. Preston Uses Daily
- Criticality calculations: neutron flux gradients in reactor cores
- ML optimization: gradient descent uses the derivative of a loss function
- Signal processing: rate of change of detected radiation counts
- Related rates: neutron population growth as a function of time
Common Errors
- Chain rule: d/dx[sin(x²)] = cos(x²) · 2x — forgetting the "inner derivative" 2x is the most common Calc I mistake
- Product rule: d/dx[fg] ≠ f'g'. Both terms are required: f'g + fg'
- Implicit differentiation: when differentiating y², the result is 2y · dy/dx, not just 2y
- L'Hôpital's rule applies only to 0/0 or ∞/∞ — verify the form before applying
Show Full Solution
Show Full Solution
Show Full Solution
Integration & Series
Calculus II covers the integral (accumulation of change) and infinite series. This is where many STEM students hit their first significant wall — a systematic approach to technique selection makes the difference.
∫ f(g(x)) g'(x) dx → let u = g(x) (u-Substitution)
∫ sin²(x) dx = x/2 − sin(2x)/4 + C (Power Reduction)
Taylor: f(x) = ∑ f^(n)(a)/n! (x−a)^n (Taylor Series at x=a)
Key Techniques
- Antiderivatives and the Fundamental Theorem of Calculus
- u-substitution: reverse chain rule
- Integration by parts: ∫ u dv = uv − ∫ v du
- Trig substitution: x = a sinθ, a tanθ, or a secθ
- Partial fractions for rational functions
- Improper integrals: limits at infinity or discontinuities
- Taylor and Maclaurin series with convergence tests
Convergence Tests
- Divergence test: if lim a_n ≠ 0, the series diverges
- Integral test: compare ∑ a_n to ∫ f(x) dx
- Ratio test: L = lim |a_{n+1}/a_n|; converges if L < 1
- Alternating Series Test: |a_{n+1}| ≤ |a_n| and lim = 0
- Comparison and limit comparison tests
Common Errors
- Forgetting +C for indefinite integrals — worth partial credit on most exams and essential for ODEs
- Wrong u-sub choice: u should simplify the integral, not complicate it
- Alternating series sign errors: ensure the series alternates sign and terms decrease to zero
- Applying the Ratio Test to alternating series without taking absolute value
Show Full Solution
Show Full Solution
Show Full Solution
Multivariable Calculus
Multivariable calculus extends single-variable ideas into multiple dimensions. This is the mathematical language of physics, fluid dynamics, electromagnetism, and reactor theory.
Topics
- Partial derivatives: ∂f/∂x holds all other variables constant
- Gradient: ∇f points in the direction of steepest ascent
- Divergence: ∇·F measures outward flux per unit volume
- Curl: ∇×F measures rotational tendency of a field
- Double and triple integrals with change of coordinates
- Stokes’ theorem: &iint; (∇×F)·dS = ∮ F·dr
Real Applications
- Reactor neutron diffusion: the Laplacian ∇²φ appears in the diffusion equation
- Gradient descent in ML: ∇L guides parameter updates
- Maxwell’s equations: ∇×E and ∇·B govern electromagnetic fields
- Heat equation: ∂T/∂t = α∇²T describes thermal diffusion
Common Errors
- Mixing up the gradient (∇f, a vector) and the Laplacian (∇²f, a scalar) — the Laplacian is the divergence of the gradient
- Integration order: when switching order of integration, the limits must be re-derived from the region, not just swapped
- Forgetting the Jacobian when changing coordinates: dA = r dr dθ in polar, not just dr dθ
Show Full Solution
Show Full Solution
Linear Algebra
Linear algebra is the language of data science, quantum mechanics, and structural engineering. The concepts here — eigenvalues, SVD, transformations — are the mathematical backbone of modern ML.
Topics
- Matrix operations: addition, multiplication, transpose, inverse
- Determinants: Laplace expansion, properties
- Systems of equations: row reduction and RREF
- Eigenvalues: det(A − λI) = 0
- Eigenvectors: (A − λI)v = 0
- Singular Value Decomposition (SVD): A = UΣV¹
- Vector spaces: span, basis, dimension, rank
- Linear transformations: kernel and image
Why It Matters
- SVD underlies PCA (dimensionality reduction) in ML
- Eigenvalues of the neutron transport matrix determine reactor criticality
- Neural network weight matrices are linear transformations
- Least squares regression: solve (A¹A)x = A¹b
- Quantum states are vectors in Hilbert space
Common Errors
- Matrix multiplication order: AB ≠ BA in general. Multiplication is not commutative
- Eigenvalue vs. eigenvector: the eigenvalue λ is a scalar; the eigenvector v is the nonzero vector satisfying Av = λv
- Singular vs. invertible: a matrix is singular (non-invertible) if and only if its determinant is zero
- Rank-nullity theorem: rank(A) + nullity(A) = number of columns — often forgotten on exams
Show Full Solution
Show Full Solution
Show Full Solution
Differential Equations
Differential equations are the mathematical description of how things change. Every physics law of motion, every reactor kinetics model, every population growth equation is a differential equation.
First-order linear: multiply by integrating factor μ = e^(∫ P dx)
2nd-order homogeneous (const coeff): try y = e^(rx) → characteristic equation
Undetermined coefficients: guess y_p based on form of g(x)
Laplace: L{y'} = sY − y(0) L{y''} = s²Y − sy(0) − y'(0)
Topics
- First-order: separable, linear, exact equations
- Second-order homogeneous with constant coefficients
- Method of undetermined coefficients
- Variation of parameters
- Laplace transforms and inverse transforms
- Systems of first-order ODEs
- Phase plane analysis and stability
Real-World Applications
- Reactor point kinetics: dN/dt = (ρ − β)/Λ N + ∑ λ_i C_i
- Newton’s law of cooling: dT/dt = −k(T − T_∞)
- RLC circuits: L d²q/dt² + R dq/dt + q/C = V(t)
- Population dynamics: logistic equation dP/dt = rP(1 − P/K)
Common Errors
- Particular vs. homogeneous solution: the general solution is y = y_h + y_p. Students often forget to add both parts
- Sign errors in Laplace transforms: L{e^(at)} = 1/(s−a), not 1/(s+a)
- Characteristic equation: for y'' + py' + qy = 0, the equation is r² + pr + q = 0. Sign errors here propagate through the entire problem
- Initial conditions are applied to the general solution, not to y_h or y_p separately
Show Full Solution
Show Full Solution
Show Full Solution
Mathematics Textbooks
Dr. Preston's field-tested recommendations. Amazon links use the fissionlab-20 affiliate tag — your purchase supports free content at no additional cost.
The standard university calculus text. Clear explanations, excellent problem sets from routine to challenging. Used in Calc I through III at most universities.
View on Amazon →The best theoretical linear algebra book available. Builds intuition for vector spaces and linear maps before matrices. Highly recommended for CS and math majors.
View on Amazon →Exceptional treatment of probability with real intuition, story proofs, and connections to statistics and ML. The accompanying Harvard Stat 110 lectures are free on YouTube.
View on Amazon →The standard reference for discrete math: logic, proofs, graph theory, combinatorics, and number theory. Essential for CS majors and anyone studying algorithms.
View on Amazon →Study With Other STEM Students
Join the FissionLab Discord to ask math questions, share solutions, and get support from peers and Dr. Preston.
Join the Discord →Ready for 1:1 Math Tutoring?
Dr. Preston works with a limited number of students on calculus, linear algebra, and differential equations. Sessions are customized to your course, your exam schedule, and the specific problems you are stuck on.
Book Free Intro Session →