Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
73 changes: 73 additions & 0 deletions reviews/literature_lifetime_law_recon_20260929/PROTOCOL.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,73 @@
{
"object": "CL-LIT-RECON-LIFETIME-20260929-v1.1",
"coverage_target": "through 2026-09-29",
"exclusions": [
"github.com/d6g8k5htny-coder",
"mirrors of the project's own repositories"
],
"executed": {
"pass_1_session_web_search_4_queries": [
"expected persistence diagram density near diagonal Gaussian random field short bars asymptotics",
"persistent homology short lifetime power law critical points maximum saddle pairs smooth random function",
"Bargmann-Fock field persistent homology lifetime density torus",
"Kac-Rice formula pairs critical points small height gap intensity ell^(-1/3) persistence"
],
"pass_2_extended_research_run_18_queries": [
"expected persistence diagram density near diagonal Gaussian random field",
"arXiv 2603.29072 \"How much of persistent homology is topology\"",
"Beliaev Cammarota Wigman no repulsion between critical points planar Gaussian random fields",
"persistence diagram smooth Gaussian random field short bars power law near diagonal asymptotics",
"Azaïs Delmas mean number and correlation function of critical points isotropic Gaussian fields",
"merge tree Gaussian random field barcode random Morse function elder rule",
"Kac-Rice persistence diagram random field expected density critical pairs",
"barcode Brownian motion expected number of bars length distribution persistent homology random process",
"persistent homology Bargmann-Fock field Betti numbers excursion sets",
"expected persistence diagram smooth stationary Gaussian process density behaviour near diagonal",
"McAuley persistence barcode Gaussian field law of large numbers topological",
"\"short bars\" persistent homology random function asymptotics number of bars length epsilon smooth",
"\"persistence\" \"near the diagonal\" smooth Gaussian process fold \"2/3\" bars length",
"Buhovsky Payette Polterovich coarse nodal count topological persistence bars length delta",
"Nicolaescu random Morse functions barcode persistence expected",
"persistence diagram Gaussian random field critical values pairs Kac-Rice \"persistence\" intensity explicit formula arXiv 2025",
"Gaussian random field on torus periodization finite-size correction critical points expected number",
"persistent homology Gaussian random field Wasserstein total persistence finite p-th moment smooth field"
],
"forward_citation_keyword_checks": [
"1802.10457",
"1911.03455",
"1911.02300",
"1706.06059",
"2012.09459",
"1909.09846"
],
"pass_3_lindgren_1972_full_text_openai_lane": {
"comment": "https://github.com/d6g8k5htny-coder/Math-/pull/147#issuecomment-5897859268",
"source": "G. Lindgren, Wave-length and amplitude in Gaussian noise, Adv. Appl. Prob. 4 (1972) 81-108, DOI 10.2307/1425807 (author-uploaded full text; Cambridge and Lund records)",
"finding": "prior art for the d=1 amplitude object and Slepian conditional formulas; no explicit small-amplitude -1/3 density asymptotic located; absence from the 1-D literature not claimed"
}
},
"not_executed": [
"classical 1-D small crest-to-trough height distributions: Rice 1945, Longuet-Higgins, Cartwright-Longuet-Higgins at full-text level (Lindgren 1972 read in full by the OpenAI lane, see executed.pass_3)",
"Divol-Polonik weight functions near the diagonal",
"LLN/CLT for persistent Betti numbers of smooth Gaussian fields, 2024-2026",
"Google Scholar cited-by enumeration for 1802.10457, 1911.03455, 1911.02300, 1706.06059, 1908.01619, 2012.09459, Bobrowski-Borman 2012, Adler-Taylor 2007"
],
"author_side_paywalled": {
"mathscinet": [
"Anywhere=(persistence diagram OR barcode) AND Anywhere=(Gaussian random field OR Gaussian field) AND Anywhere=(diagonal OR short OR small lifetime)",
"MSC=(55N31 AND 60G60)",
"MSC=(60G60 AND 58K05) AND Anywhere=(critical value* AND (gap OR pair* OR close))",
"Anywhere=\"Bargmann-Fock\" AND Anywhere=(persistence OR Betti OR barcode)",
"Anywhere=(crest trough OR wave height) AND Anywhere=(small OR asymptotic) AND MSC=60G15"
],
"zbmath_open": [
"ti:(persistence OR barcode) & ab:(Gaussian field) & ab:(diagonal)",
"cc:55N31 & cc:60G60",
"ab:\"critical values\" & ab:\"Kac-Rice\" & ab:(pair* | two-point)",
"ab:\"Bargmann-Fock\" & ab:(persistent | Betti | barcode)",
"ab:(torus | periodic) & ab:\"Gaussian field\" & ab:(exponentially small | finite-size correction)"
]
},
"verdict": "no conflicting, overlapping, or different-exponent theorem found for the smooth-field H0 near-diagonal lifetime law under this protocol; not a proof of novelty",
"scientific_effect": "NONE"
}
140 changes: 140 additions & 0 deletions reviews/literature_lifetime_law_recon_20260929/RECONNAISSANCE.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,140 @@
# Reconnaissance memo — literature and novelty check for the near-diagonal H0 lifetime law — 2026-09-29

**Object:** CL-LIT-RECON-LIFETIME-20260929-v1.1 (v1 + the executed 1-D neighbor pass, §4 item D). **Author:** Anthropic Claude (claude.ai session `session_01NMeKEismAyeqgdB4sy2NJU`).
**Scientific effect:** NONE. **No priority or novelty claim follows from this memo.** It records what a documented
search did and did not find, so that any lane can extend it and so that successor attribution sections have a
verified starting list. Bibliographic details below are *as indexed by the search pass*; verify against full text
before citing in a manuscript. arXiv identifiers are the stable handles.

## 1. Object searched

The near-diagonal law for finite ordinary superlevel H0 bars of a smooth stationary Gaussian field on a fixed torus:
`ν(ℓ) ~ c ℓ^{−1/3}` (parent Theorem C at existential scope; coefficient (15.2)); its corollaries `N(0,t] ~ (3/2) c
t^{2/3}` and the inverse-lifetime integrability threshold `p < 2/3`; the `d = 3` reference constant
`c_{3,∞} = 2^{−23/6} 3^{−8/3} (29√6 − 36) Γ(1/6) π^{−5/2} ≈ 0.0417759318…` (`coefficients/side24_v1`, main#65); and the
nearest-image periodization correction `P_3(24) e^{−288}` (PR #144). Coverage target: through 29 September 2026.

## 2. Verdict under this protocol

- **Same law proved elsewhere:** none found.
- **Same law contradicted:** none found.
- **Different exponent claimed for the same object** (smooth Gaussian field, finite H0 superlevel bars, `ℓ ↓ 0`):
none found. The explicit small-bar power laws that do exist are for **rough** processes and a different counting
convention (number of bars of length `≥ ε`, divergent as `ε → 0`): `N^ε ~ [X]_t/(2ε²)` for continuous
semimartingales (Perez, arXiv:2012.09459, JACT 2023), and `N^ε ~ C_H ε^{−1/H}` for Brownian, α-stable and
fractional Brownian motion (Picard, as restated by Perez). These are neighbors, not conflicts: a `C²` path has
finite variation and is outside their regime. Perez states in print that the smooth-field case is open.
- **Bargmann–Fock-specific persistence results beyond percolation:** none found (located: Rivera–Vanneuville
critical threshold, arXiv:1711.05012; Gass–Stecconi finiteness of critical-point moments, arXiv:2305.17586).
- **Periodization / nearest-image corrections to local constants on tori:** none found. **No
external benchmark was located under this protocol** for the `e^{−L²/2}` term or the coefficient `−620813376/35`; the benchmark is internal and
multi-path (see §4, item C).

"None found" is a statement about this protocol (§5), not a proof of novelty. Three planned searches were not run
and the paywalled databases were not covered; §4 lists them for the next lane.

## 3. Closest neighbors (condensed; what each does / does not cover)

| arXiv / venue | Work | Covers | Does not cover |
|---|---|---|---|
| 1802.10457; SoCG 2018 / JoCG 2019 | Chazal–Divol, density of expected persistence diagrams | existence and smoothness of EPD densities; Brownian sublevels; notes accumulation near the diagonal | any exponent or constant for smooth fields |
| 2012.09459; JACT 2023 | Perez, PH of a.s. `C⁰` processes | sharp `ε → 0` bar-count asymptotics for semimartingales; exact Brownian expansion | smooth fields (stated open) |
| 2110.10982 | Perez, ζ-functions of superlevel barcodes of α-stable Lévy processes | meromorphic barcode ζ-functions; pole tied to `N^ε` | smooth; `d ≥ 2` |
| 1909.09846; JACT 2025 | Baryshnikov, Brownian motion / chirality | `PH_0` intensity and correlation functions for Brownian motion with drift | smooth; `d ≥ 2` |
| 1908.01619; JCAP 2019 | Feldbrugge–van Engelen–van de Weygaert–Pranav–Vegter | **Rice-formula pair densities `p_i(b,d,r)` for (saddle, extremum) pairs of 2-D Gaussian fields**, pushed to persistence diagrams by a fitting formula | small-`(d−b)` asymptotics, exponent, constant, rigorous elder pairing — **closest structural neighbor; cite and distinguish** |
| 1911.03455; ECP 2020 | Beliaev–Cammarota–Wigman, no repulsion (planar) | leading short-distance term of the critical-point two-point function; index dependence | value gaps; persistence |
| 1704.04943; IMRN 2019 | Beliaev–Cammarota–Wigman, random plane wave | two-point function of critical points | same |
| 1911.02300; SPA 2022 | Azaïs–Delmas, critical points of isotropic fields and GOE | attraction (`d > 2`) / neutrality (`d = 2`) / repulsion (`d = 1`), attributed to adjacent-index pairs | value-gap statistics; pairing |
| 2209.04150; SPA 2023 | Ladgham–Lachièze-Rey, local repulsion of planar critical points | second factorial moment of typed critical points in small balls | values; `d ≥ 3` |
| 2109.08721 | Pranav, topology and geometry of Gaussian fields II | simulated 3-D persistence diagrams / intensity maps | asymptotics |
| 1003.1001; 2010 | Adler–Bobrowski–Borman–Subag–Weinberger | framework and simulations for barcodes of excursion sets | short-bar law |
| J. Topol. Anal. 2012 | Bobrowski–Borman, Euler integration | expected Euler integrals ↔ persistence | near-diagonal density |
| 1808.05655 | Adler et al., planar point-process models of diagrams | qualitative repulsion of diagram points via Slepian models | exponent |
| 1706.06059; 2107.11212 | Curry; "trees to barcodes II" | elder-rule map merge tree → barcode; realization counts | random-field lifetime intensity |
| 2607.23903 (2026) | Kuriki–Matsubara–Iso, critical-point densities in any `d` | one-point height distributions by index | pair gaps |
| 2305.17586; PTRF 2024 | Gass–Stecconi | finite moments of critical-point counts incl. Bargmann–Fock | persistence |
| 2206.06347; JEMS 2026 | Buhovsky–Payette–Polterovich–Polterovich–Shelukhin–Stojisavljević | deterministic bounds on bars longer than `δ` | random `δ → 0` asymptotics |
| 2603.27903 (2026) | random-matrix persistence via Morse theory | bar lengths = eigenvalue spacings; Wigner-surmise bar laws | Gaussian fields |
| 2603.29072 (2026) | Loftus, "How much of persistent homology is topology?" | point-cloud PH of spin configurations (cond-mat) | **unrelated** to Gaussian-field lifetime densities (downgraded from an earlier "worth a read" flag) |

Forward-citation trail from Chazal–Divol (keyword-level, not an exhaustive crawl): the line quantifies near-diagonal
behavior only for Brownian-type processes and treats the near-diagonal region of smooth models as an estimation
nuisance (Maroulas–Mike–Oballe JMLR 2019; Divol–Lacombe 2105.04852; persistence-intensity estimation, PMC12083882;
2407.07326 JACT 2025; 2607.20893; 2607.27126). Nothing in it states a rate for a smooth model.

## 4. Items the other lanes can act on

**A. Attribution additions for successor packages.** The parent's §16 attribution cites Armentano–Azaïs–León,
Curry, and Beliaev–Cammarota–Wigman. A related-work section for the lifetime law should additionally distinguish:
Feldbrugge et al. 2019 (Rice-formula pair densities — closest structural neighbor); Perez 2023 and Picard
(rough-process small-bar laws, different regime); Azaïs–Delmas 2022 and Ladgham–Lachièze-Rey 2023 (spatial
critical-point correlations, no value gaps); Chazal–Divol 2018/19 (EPD density existence). No claim of the
parent's need be changed; this is scope-setting for readers.

**B. A heuristic dimension-independence cross-check reviewers may raise (heuristic; not sourced; not a proof).** If
the adjacent-index critical-pair correlation scales like `r^{2−d}` at small separation — the pattern consistent with
Azaïs–Delmas's `d = 1` repulsion / `d = 2` neutrality / `d > 2` attraction — and the fold gap scales as `r³`, then
`∫ r^{2−d} · r^{d−1} dr ∝ r²` in every dimension, giving `P(gap ≤ t) ∝ t^{2/3}` and density `∝ ℓ^{−1/3}`
independently of `d`. This matches the parent §10 ledger ("the dimension cancels") from the outside, and the factor
12 in the fold gap matches the `24^{1/3}` in (15.2). It rests on two unproved assumptions (short bars come from
fold-type adjacent pairs; the elder rule pairs them) — exactly what Theorem A + CAP supply rigorously. Useful as a
sanity narrative in a discussion section; not evidence.

**C. Periodization coefficient — the benchmark is internal.** No external work was located under this protocol that computes exponentially small
nearest-image corrections to such constants. The coefficient `P_3(24) = −620813376/35` now has independent
in-project derivations: PR #144 (first principles in the parent's (15.2) conventions, stdlib checker), the July
06B suite (SymPy, Hermite + Haar), the OKComputer 2026-08 first-variation script, and a clean-context re-check on
2026-09-29 — all exact agreement. `|P_3(24)| e^{−288} < 10^{−117}` sits inside side24_v1's reviewed `< 10^{−106}`
band. The remainder enclosure remains PR #144's flagged open item.

**D. The classical 1-D neighbor — executed in two passes (v1.1).** The classical 1-D wave literature on the
distribution of *small* crest-to-trough heights of a smooth stationary Gaussian process (Rice 1945; Longuet-Higgins;
Lindgren 1972 "Wave-length and amplitude in Gaussian noise"; Cartwright–Longuet-Higgins) is the natural `d = 1`
neighbor of the max–min small-gap law. *Pass 1 (author, web-level, 2026-09-29):* the neighbors were located
precisely — G. Lindgren, Adv. Appl. Prob. 4 (1972) 81–108, DOI 10.2307/1425807; Lindgren–Rychlik 1982 — but full
texts were not read. *Pass 2 (OpenAI/ChatGPT, full-text read of Lindgren 1972,
[PR #147 comment 5897859268](https://github.com/d6g8k5htny-coder/Math-/pull/147#issuecomment-5897859268)):* the paper
defines the crest-to-trough amplitude after a prescribed maximum, develops the Slepian conditional process, and in
§§2.2–2.4 constructs an *approximate* joint wavelength/amplitude density under its condition A_p; Theorem 2.4 and
Corollary 2.2 are approximation/error statements, §2.3 gives numerical amplitude-density curves, §2.4 rejects a
Rayleigh/independence hypothesis. **No theorem or displayed asymptotic of the form `f_amp(h) ~ C h^{−1/3}` (or
`F_amp(h) ~ C h^{2/3}`), and no explicit `h ↓ 0` expansion, was found in the full text or by search within it.**
Attribution consequence, adopted here as the successor wording: *Lindgren 1972 is clear `d = 1` prior art for the
object (the maximum-to-following-minimum amplitude law) and for conditional/Slepian formulas that could in
principle be expanded; it is not, on the evidence read, prior art for the `−1/3` small-amplitude density
exponent. We have not located an explicit small-amplitude `−1/3` asymptotic in that source, and absence from the
entire 1-D literature is not claimed.* Rice 1945, Longuet-Higgins and Cartwright–Longuet-Higgins remain unread at
full-text level. Still unexecuted: Divol–Polonik weight functions near the diagonal; LLN/CLT for
persistent Betti numbers of smooth Gaussian fields 2024–2026; Google-Scholar "cited by" enumeration for
1802.10457, 1911.03455, 1911.02300, 1706.06059, 1908.01619, 2012.09459, Bobrowski–Borman, Adler–Taylor.

**E. Author-side paywalled queries (verbatim, not yet run).**
MathSciNet: `Anywhere=(persistence diagram OR barcode) AND Anywhere=(Gaussian random field OR Gaussian field) AND
Anywhere=(diagonal OR short OR small lifetime)`; `MSC=(55N31 AND 60G60)`; `MSC=(60G60 AND 58K05) AND
Anywhere=(critical value* AND (gap OR pair* OR close))`; `Anywhere="Bargmann-Fock" AND Anywhere=(persistence OR
Betti OR barcode)`; `Anywhere=(crest trough OR wave height) AND Anywhere=(small OR asymptotic) AND MSC=60G15`.
zbMATH Open: `ti:(persistence OR barcode) & ab:(Gaussian field) & ab:(diagonal)`; `cc:55N31 & cc:60G60`;
`ab:"critical values" & ab:"Kac-Rice" & ab:(pair* | two-point)`; `ab:"Bargmann-Fock" & ab:(persistent | Betti |
barcode)`; `ab:(torus | periodic) & ab:"Gaussian field" & ab:(exponentially small | finite-size correction)`.

## 5. Protocol record

- **Executed 2026-09-29** in two passes: (i) a four-query web pass by this session (near-diagonal EPD density
asymptotics; short-lifetime power law for max/saddle pairs; Bargmann–Fock persistent homology; Kac–Rice small-
height-gap pair intensity); (ii) an extended web-research run (18 general-search queries plus a delegated
sub-search of ~17 calls) indexing arXiv (math.PR, math.AT, math-ph, math.DG, cond-mat), Project Euclid, Springer
(JACT, PTRF), ScienceDirect, JMLR, PMC, ORA, author pages, with keyword forward-citation checks of 1802.10457,
1911.03455, 1911.02300, 1706.06059, 2012.09459, 1909.09846. Verbatim queries are in `PROTOCOL.json`.
- **Exclusions:** `github.com/d6g8k5htny-coder` and mirrors (self-hits). The project is now search-indexed, so any
future novelty search must apply this exclusion.
- **Caveats:** keyword search is not a citation crawl; several items were characterized from snippets or
summaries and must be verified against full text; the §4-B argument is an inference made for this memo.
- **Prior in-repo reconnaissance this memo extends:** parent §16 "Primary-source attribution"; side24_v1 "External
reconnaissance, 24 September 2026" (DLMF only); OKComputer 2026-08 Part D web pass (five clusters, superseded
packet).

## 6. Relation to open registers

This memo discharges the *web-level* part of the standing literature item that every register since July 2026 has
carried ("paywalled MathSciNet/zbMATH pass remains author-side"). It does not close that item: §4-D and §4-E remain.
It changes no claim in `claims/LANDING_CLAIMS.json`, no `PROOF_INDEX.md` row, and no status anywhere.
Original file line number Diff line number Diff line change
@@ -0,0 +1,21 @@
{
"author": "Anthropic Claude",
"delivered_under": "explicit owner instruction, claude.ai session 2026-09-29; OWNER_STOP left in place; no self-merge",
"files": [
{
"bytes": 13738,
"path": "RECONNAISSANCE.md",
"sha256": "16628983f7833419e33c7b04cdb2e0d49076bccfbe7a6d457bf90efba63aa4ec"
},
{
"bytes": 4631,
"path": "PROTOCOL.json",
"sha256": "aed7eebebf7840a32efbc86632825ae533d2ef242d204e7fe9ed39ece94e7461"
}
],
"kind": "literature reconnaissance memo (prose + machine-readable protocol); no checker",
"novelty_claim": false,
"object": "CL-LIT-RECON-LIFETIME-20260929-v1.1",
"scientific_effect": "NONE",
"session": "https://claude.ai/code/session_01NMeKEismAyeqgdB4sy2NJU"
}
Loading