Download lean/OAI/Geometry/Perimeter/Regularity/EllipticPointwise.lean from roseking/openAI-math-dataset: direct link, hf CLI and curl.
- Browser
- Download file 2.37 kB
-
https://huggingface.co/datasets/roseking/openAI-math-dataset/resolve/main/lean/OAI/Geometry/Perimeter/Regularity/EllipticPointwise.lean
- Command line
-
hf download hf://datasets/roseking/openAI-math-dataset/lean/OAI/Geometry/Perimeter/Regularity/EllipticPointwise.lean
-
curl -L -o EllipticPointwise.lean https://huggingface.co/datasets/roseking/openAI-math-dataset/resolve/main/lean/OAI/Geometry/Perimeter/Regularity/EllipticPointwise.lean
2.37 kB
| import OAI.Geometry.Perimeter.Regularity.EllipticIteration | |
| import OAI.Geometry.Perimeter.Regularity.EllipticGradientLimit | |
| namespace OAI | |
| noncomputable section | |
| open Set Filter MeasureTheory | |
| open scoped Topology ENNReal ContDiff InnerProductSpace | |
| namespace RPISingularHarnack | |
| variable {n : ℕ} | |
| theorem linear_pointwise_iteration {θ δ K a B H R : ℝ} | |
| (hθ : 0<θ) (hθ1 : θ<1) (hK : 1≤K) | |
| (hc : LinearContractionProperty n θ δ K) | |
| (ha : 0≤a) (hB : 0≤B) (hH : 0≤H) (hR : 0<R) | |
| (hsmall : 32*K^2*a≤1) (hforce : 16*K*B≤H) | |
| {A : Euc n → Euc n →L[ℝ] Euc n} (hA : Continuous A) | |
| {f : Euc n → ℝ} (hf : ContDiff ℝ 1 f) {F : Euc n → Euc n} (hF : Continuous F) | |
| (p : Euc n) (h : WeakLinearEquation A f F (Metric.ball p R)) | |
| (hnear : ∀ x∈Metric.ball p R, ‖ContinuousLinearMap.id ℝ (Euc n)-A x‖≤δ) | |
| (hosc : ∀ j x, x∈Metric.ball p (θ^j*R) → | |
| ‖A p-A x‖≤a*(1/2)^j ∧ ‖F x-F p‖≤B*(1/2)^j) | |
| (hinit : scaledGradientEnergy f p R 0≤H^2) : | |
| ∃ q : ℕ → Euc n, (∀ j, ‖q j-gradient f p‖≤2*K*H*(1/2)^j) ∧ | |
| (∀ j, scaledGradientEnergy f p (θ^j*R) (q j)≤H^2*((1/2)^j)^2) ∧ | |
| ‖gradient f p‖≤2*K*H := by | |
| obtain ⟨q,hq0,hqb,hqE,hqinc⟩ := linear_energy_iteration hθ hθ1 hK hc ha hB hH hR | |
| hsmall hforce hA hf hF p h hnear hosc hinit | |
| have hdist (j : ℕ) : dist (q j) (q (j+1))≤(K*H)*(1/2:ℝ)^j := by | |
| rw [dist_eq_norm,norm_sub_rev]; exact hqinc j | |
| obtain ⟨w,hw⟩ := cauchySeq_tendsto_of_complete | |
| (cauchySeq_of_le_geometric (1/2) (K*H) (by norm_num) hdist) | |
| have hwf : gradient f p=w := gradient_eq_of_scaled_energy hf p | |
| (fun j ↦ by positivity) | |
| (by simpa only [zero_mul] using (tendsto_pow_atTop_nhds_zero_of_lt_one hθ.le hθ1).mul_const R) | |
| (by simpa only [zero_pow (by norm_num : (2:ℕ)≠0),mul_zero] using | |
| ((tendsto_pow_atTop_nhds_zero_of_lt_one (by norm_num : (0:ℝ)≤1/2) (by norm_num)).pow 2).const_mul (H^2)) | |
| hw hqE | |
| have herror (j : ℕ) : ‖q j-gradient f p‖≤2*K*H*(1/2)^j := by | |
| have hh := dist_le_of_le_geometric_of_tendsto (1/2) (K*H) (by norm_num) hdist hw j | |
| rw [dist_eq_norm,←hwf] at hh | |
| convert hh using 1 | |
| ring | |
| refine ⟨q,herror,hqE,?_⟩ | |
| have hh := herror 0 | |
| simpa only [hq0,zero_sub,norm_neg,pow_zero,mul_one] using hh | |
| end RPISingularHarnack | |
| end | |
| end OAI | |