(4.2), p. 32 — S_κ(a) = (1 − κ/|a|)₊ a for a ≠ 0
ProvedBoydADMM.Prox.softThreshold_eq_shrinkp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-bookp2o-v1shrinkagesoft-thresholding
For and every ,
This formula shows that soft thresholding is a shrinkage operator: it multiplies by a factor in , moving it toward zero.
Formalization Note The hypothesis is the book's ("for "); it is kept explicitly rather than relying on Lean's convention . The hypothesis is implicit in the book () and is stated.
Preamble
import Mathlib import Definitions.Def_BoydADMM_Prox_Basic open Matrix
Formal statement
namespace BoydADMM.Prox
theorem softThreshold_eq_shrink (κ : ℝ) (hκ : 0 ≤ κ) (a : ℝ) (ha : a ≠ 0) :
softThreshold κ a = max (1 - κ / |a|) 0 * a := by sorry
end BoydADMM.Prox
Source
Boyd, Parikh, Chu, Peleato, Eckstein, Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers, Found. Trends Mach. Learn. 3(1) (2011), p. 32, §4.4.3, (4.2)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.