You are viewing a public preview
Sign in to save your progress, do assignments, join discussions and get feedback.
Flashcards · Vector spaces and linear mappings
9 wordsA nonempty set with addition and real scalar multiplication satisfying ten rules: closure under + and ·, commutativity, associativity, zero vector, additive inverses, two distributive laws, c(du) = (cd)u, 1u = u.
The vector 0 with u + 0 = u for every u. In C(ℝ) it is the function that is 0 for every t (not the constant 1).
ℝⁿ (componentwise), the space M(m × n) of matrices of a fixed shape, and the continuous functions C(ℝ) with pointwise operations.
A subset W is a subspace if it contains 0 and is closed under addition and scalar multiplication.
T(u + v) = T(u) + T(v) and T(cu) = cT(u); equivalently T(au + bv) = aT(u) + bT(v).
Every linear map has T(0) = 0. If T(0) ≠ 0, T is not linear; if T(0) = 0 you still need to check both conditions.
A = [T(e₁) … T(eₙ)]: the j-th column is the image of the j-th standard basis vector; T(x) = Ax.
x ↦ Ax + b. It is linear exactly when b = 0.
A neural-network layer Wx + b is affine.
The matrix of S ∘ T is BA (T acts first). T is invertible iff its matrix A is invertible, and the matrix of T⁻¹ is A⁻¹.
Work through Section 5 of Lab 1 (Linear Algebra with NumPy, MFDS_Lab1_Linear_Algebra.ipynb), attached to the Part I materials lesson at the start of this part.