Proof of Proposition 6.8, p. 19 — ‖H‖_F ≤ ‖H + K‖_F when ⟨H, K⟩ = 0
ProvedIRLSM.RIP.frobeniusNorm_le_of_traceInner_eq_zerolow-rank-recoveryp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be real matrices with . Then
In the proof of Proposition 6.8 this is the step , "by orthogonality of and ".
Formalization Note. The page writes for the Frobenius norm .
Preamble
import Mathlib import Definitions.Def_IRLSM_Convergence_Basic open HighDimStat.MatrixRank Matrix
Formal statement
namespace IRLSM.RIP
/-- If `⟨H, K⟩ = 0` then `‖H‖_F ≤ ‖H + K‖_F`.
Fornasier–Rauhut–Ward, arXiv:1010.2471v4, proof of Proposition 6.8, p. 19 (second inequality of
"‖H₀‖_* ≤ √(2k)‖H₀‖_F ≤ √(2k)‖H₀ + H₁‖ by orthogonality of H₀ and H₁"; the page writes
`‖H₀ + H₁‖` for the Frobenius norm).
Formalization Note: real matrices; `⟨·,·⟩` is `traceInner`. -/
theorem frobeniusNorm_le_of_traceInner_eq_zero {n p : ℕ} (H K : Matrix (Fin n) (Fin p) ℝ)
(h : traceInner H K = 0) :
frobeniusNorm H ≤ frobeniusNorm (H + K) := by sorry
end IRLSM.RIP
Source
Fornasier, Rauhut, Ward, Low-rank matrix recovery via iteratively reweighted least squares minimization, arXiv:1010.2471v4 (2011), proof of Proposition 6.8, p. 19, second inequality of ‖H0‖_* ≤ √(2k)‖H0‖_F ≤ √(2k)‖H0 + H1‖
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.