The canonical RGBCG block gradient is the Chapter 11 block partial gradient
∇[i] (fun z ↦ (f z).toReal), evaluated at the PiLp point corresponding to the raw block tuple
y. This is the Chapter 13 bridge/view map used when translating the source-facing PiLp
problem into the Chapter 11 raw-product owner.
Instances For
Evaluating the canonical RGBCG block gradient at (i, y) recovers the corresponding block
partial gradient of x ↦ (f x).toReal on PiLp 2 Ei.
Evaluating the RGBCG bridge map at the raw coordinates of x : PiLp 2 Ei recovers the
source-facing block partial gradient at x.
Assumption 13.28: clauses (A)-(D) for the randomized generalized block conditional-gradient
method are expressed directly on the chapter objects f : X → EReal, the block penalties g_i,
the canonical PiLp block gradient ∇[i] (fun z ↦ (f z).toReal). The public owner keeps only
the primitive regularity, domain-compatibility, differentiability, and block-Lipschitz clauses on
the PiLp model; the canonical optimizer set
unconstrained_problem_solutions (composite_model_objective f (PiLp.separableSum g)) and optimal
value generalized_conditional_gradient_optimal_value f (PiLp.separableSum g) are derived below
rather than stored as wrapper parameters. The one-block update and slice are written in the
canonical Chapter 11 textbook form x + 𝒰[i] d.
- block_g_proper (i : ι) : IsProperExtendedRealFunction (g i)
- block_g_closed (i : ι) : LowerSemicontinuous (g i)
- block_g_convex (i : ι) : is_convex_function (g i)
- g_effective_domain_compact (i : ι) : IsCompact (effective_domain (g i))
- f_ne_bot (x : PiLp 2 Ei) : f x ≠ ⊥
- f_convex : is_convex_function f
- g_effective_domain_subset_f_effective_domain : effective_domain (PiLp.separableSum g) ⊆ effective_domain f
- f_effective_domain_open : IsOpen (effective_domain f)
- f_toReal_differentiableOn_effective_domain : DifferentiableOn ℝ (fun (x : PiLp 2 Ei) => (f x).toReal) (effective_domain f)
- block_partial_gradient_lipschitz (i : ι) {x : PiLp 2 Ei} {d : Ei i} (hx : x ∈ effective_domain f) (hxd : x + PiLp.single 2 i d ∈ effective_domain f) : ‖(∇[i] fun (z : PiLp 2 Ei) => (f z).toReal) x - (∇[i] fun (z : PiLp 2 Ei) => (f z).toReal) (x + PiLp.single 2 i d)‖ ≤ ↑(Li i) * ‖d‖
Instances
A generalized block conditional-gradient problem canonically makes the smooth term f a
proper extended-real-valued function.
Each block penalty is proper by the source-facing owner field
block_g_proper.