60 lines
2.5 KiB
TeX
60 lines
2.5 KiB
TeX
%% Vol I, Chapter 5 — Experimental Results
|
|
%% SOURCE: papers/James_Steady-State_Convergence_Adaptive_Runtime.pdf (SSRN, published)
|
|
%% docs/working/archive/legacy/phase-1/Reference/physics_experiment/ (raw data)
|
|
%% docs/working/papers/FORMAL_CLAIM_TABLE.md
|
|
%% docs/working/papers/NEGATIVE_RESULTS.md
|
|
%% docs/working/papers/SENSITIVITY_ANALYSIS.md
|
|
%% docs/working/experiments/02-experiments/physics-optimization/execution-guide.md
|
|
|
|
\chapter{Experimental Results}
|
|
\label{vol1:chap:results}
|
|
|
|
%% TODO(bob): this chapter is the narrative companion to the published SSRN
|
|
%% paper (cite james:2025:ssrn). Do NOT duplicate the paper verbatim; instead
|
|
%% provide context, reproducibility guidance, and formally structured claims.
|
|
|
|
\section{Experimental Design}
|
|
\label{vol1:sec:results:design}
|
|
|
|
%% SOURCE: docs/working/experiments/02-experiments/physics-optimization/execution-guide.md
|
|
%% TODO(bob): describe the 2^7 factorial DoE design (90-run subset), the
|
|
%% workload-shape taxonomy (arithmetic-heavy, control-heavy, mixed), and
|
|
%% the triple-blind validation protocol.
|
|
|
|
\section{Primary Result: Steady-State Convergence}
|
|
\label{vol1:sec:results:primary}
|
|
|
|
%% SOURCE: james:2025:ssrn
|
|
%% TODO(bob): state the main result in one sentence. Then cite the paper.
|
|
%% Key figure: 0\% algorithmic variance (CV) across 90 independent runs.
|
|
|
|
\section{Claim Table}
|
|
\label{vol1:sec:results:claims}
|
|
|
|
%% SOURCE: docs/working/papers/FORMAL_CLAIM_TABLE.md
|
|
%% TODO(bob): promote the claim-evidence-reproducibility table here as a
|
|
%% formal LaTeX longtable or a series of claim-numbered paragraphs.
|
|
%% Each claim must carry its evidence source and a reproducibility verdict.
|
|
|
|
\section{Negative Results and Limitations}
|
|
\label{vol1:sec:results:negative}
|
|
|
|
%% SOURCE: docs/working/papers/NEGATIVE_RESULTS.md
|
|
%% TODO(bob): promote the honest failure-mode catalog from NEGATIVE_RESULTS.md.
|
|
%% This section is essential for peer-review defensibility.
|
|
|
|
\section{Sensitivity Analysis}
|
|
\label{vol1:sec:results:sensitivity}
|
|
|
|
%% SOURCE: docs/working/papers/SENSITIVITY_ANALYSIS.md
|
|
%% TODO(bob): describe the parameter-sweep design for testing robustness of
|
|
%% the 0\% CV claim against perturbations in tuning knobs.
|
|
|
|
\section{Reproducing the Results}
|
|
\label{vol1:sec:results:reproduce}
|
|
|
|
%% SOURCE: docs/working/papers/REPRODUCIBILITY.md
|
|
%% docs/working/experiments/02-experiments/physics-optimization/reproduce.md
|
|
%% TODO(bob): one-command reproduction recipe with exact expected output.
|
|
%% Cross-reference Appendix A (build instructions).
|