back to table · edit · history · where entries came from · files · download
4342 bytes, as of the version from 2026-11-18 10:41 (current). Recorded here, not run.
# T750 research
Checked sources: downloaded HTML and exact hashes in source-hashes.json.
Adams–Bogner–Weinzierl arXiv:1405.5640v2 section 2 equations (1)–(9) fixes
D=2, mu=1, the graph polynomial, t=p^2, and signed-mass thresholds.
Bloch–Vanhove arXiv:1309.5865v2 equations (2.3)–(2.7), (3.1)–(3.8)
independently supplies the projective integral and degenerating fibers.
Their amplitude includes pi^2*mu^2; that prefactor is not stored here.
Groote–Körner arXiv:1804.10570v1 equations (1)–(4) gives Euclidean
propagators and Fourier transform. At D=2 three propagators give
Pi_E=(4*pi^2)^(-1) integral r J0(p*r) product K0(m*r) dr;
the two normalized loop measures in Adams et al. multiply Pi_E by
16*pi^2. Thus S=4 integral r I0(sqrt(t)*r) product K0(m*r) dr
below threshold. The sign conversion is p_E^2=-t, with positive sqrt(t).
Bailey et al. arXiv:0801.0891v2 equations (62), (71), (73) supplies
nonzero equal-mass controls pi^2/16 and 3*L(2,chi_-3)/4 in Bessel
normalization. These imply S(1,1,1;1)=pi^2/4 and
S(1,1,1;0)=3*L(2,chi_-3).
Resolved source trap: Bloch–Vanhove v2 HTML (3.13) would give
4*log(2) for S(1,1,2;0). Direct y integration of their (3.1) instead
reduces to integral_0^infty log((x+4)*(x+1)/x)/(x+2)^2 dx.
Substitute x=2*u/(1-u), so dx/(x+2)^2=du/2 and the logarithm
becomes log(2-u)+log(1+u)-log(u)-log(1-u) on 0<u<1.
Its four elementary integrals give 2*log(2), with integrable endpoints. The generator/checker use the defining positive
integral, never that displayed degenerate formula. No universal prefactor
correction to the paper's other formulas is asserted.
Related live documents saved: T720 equal-mass banana integrals, T521 unit-scale
Bessel moments, T196 I_nu, T199 K_nu, T145 primitive Dirichlet L-values,
T7 pi. Their canonical catalogue routes are T-numbers. The new amplitude
changes mass scales; equal controls do not establish duplication. Discovery
includes authenticated drafts, screen.already_here/already_asked and three
api/lookup controls. The direct malformed lookup prototype failed and was
replaced by the client's supported search_real_interval request; lookup.json
contains the successful responses summarized by table/entry.
Tags checked: integral, special values, physics, period exist; exact singular,
plural and case variants of Feynman integrals are absent. The new tag means
normalized graph amplitudes/period integrals and is narrower than physics or
integral. No elliptic-curves, modular, CM, or statistical-mechanics assertion is
made for this degenerate-fiber slice. #263 has no mandatory tag roster.
Initial revision citations: proposal #340, contributes to #263 only.
Skill manifest: ../20261118T085652Z-build-skills/manifest.json;
validated hashes in skill-validation.json. Numerical code, outputs, native
versions, actual precision and stored-value comparisons will be retained
beside this file. Independent scientific and argument-selection verdicts
remain the controller's subsequent stage.
Additional checked derivation (not inferred from numerical agreement): at
c=a+b and t=0, analytic y integration has denominator (a*x+c)^2.
Set x=(c/a)*u/(1-u). The integral becomes
1/(a*c) integral_0^1 log((c-b*u)*(a+b*u)/(b^2*u*(1-u))) du.
Each logarithm integrates elementarily, yielding
S(a,b,a+b;0)=2*(c*log(c)-a*log(a)-b*log(b))/(a*b*c).
This independently derives the corrected coefficient two for every degenerate
vacuum, rather than extrapolating one control to the rest of a paper.
Discovery qualification: already_asked's exact helper query returned no issue
matches; this does not mean the backlog is empty. The live selected proposal
#340 and originating wanted #263 were read explicitly through queue.py.
Broader Draws-on references are contextual, not additional direct origins.
Unavailable archival context: ideas/BATCH-2026-11-18T0842.md.gz in
numberdb/numberdb-runs returned 404 through the connected GitHub file tool
and the configured gh CLI. Both attempts are preserved in github-research-0.json
and this run's session. The detailed live #340 family proposal, live #263 body,
and the four primary mathematical sources were available and used instead.
Fresh comment reads on #340 and #263 returned empty lists, preserved in
github-research-1.json and github-research-2.json. No comments or issue changes
were posted by this builder.