OAI/GroupTheory/PolycyclicRecognition/Algebra/SequenceSubsequences.lean
3 declarations, 0 contain sorry, 0 declared as axioms
Passing all profile data to the same sequence of good scales
- theorem
OAI.PolycyclicRecognition.HasSubexponentialLowerBoundSeq.compline 13Source documentation
Subexponential losses are unchanged on a subsequence tending to infinity.
theorem HasSubexponentialLowerBoundSeq.comp {scale f : ℕ → ℝ} {v : ℕ → ℕ} (hf : HasSubexponentialLowerBoundSeq scale f) (hv : Tendsto v atTop atTop) : HasSubexponentialLowerBoundSeq (scale ∘ v) (f ∘ v) := …Mentions: OAI.PolycyclicRecognition.HasSubexponentialLowerBoundSeq
- theorem
OAI.PolycyclicRecognition.HasSubexponentialUpperBoundSeq.compline 18 - theorem
OAI.PolycyclicRecognition.exists_increasing_good_scale_subsequenceline 26Source documentation
The eventual cell and slack conditions can hold at every retained index, while the physical scales are strictly increasing. No interpolation of the profiles or of the group laws is needed.
theorem exists_increasing_good_scale_subsequence {scale : ℕ → ℝ} (hscale : Tendsto scale atTop atTop) {P : ℕ → Prop} (hP : ∀ᶠ n in atTop, P n) : ∃ u : ℕ → ℕ, StrictMono u ∧ StrictMono (scale ∘ u) ∧ ∀ n, P (u n) := by …