{"id":"research-index","title":"Research Index","subtitle":"Annotated catalog of every research, planning, and presentation document.","category":"foundations","tags":["index","overview"],"source":"articles/docs/research-index.md","lang":"en","words":256,"readMinutes":1,"toc":[{"depth":2,"text":"Multi-fidelity, UQ & ensembles","id":"multi-fidelity-uq-ensembles"},{"depth":2,"text":"Error prediction & topology","id":"error-prediction-topology"},{"depth":2,"text":"Benchmarking & physical models","id":"benchmarking-physical-models"},{"depth":2,"text":"Theory & information","id":"theory-information"},{"depth":2,"text":"Program context","id":"program-context"}],"html":"<h1 id=\"deep-research-reports-index\">Deep-Research Reports Index</h1><blockquote>\n<p>⚠️ <strong>Partly stale.</strong> This file was pruned 2026-06-02 because it catalogued a\n<code>glim/</code>-era layout and pointed to root research docs (<code>deep-research-report.md</code>,\n<code>ancillary-research-opps.md</code>, <code>foundational-research.md</code>, <code>example-research-papers.md</code>)\nthat <strong>no longer exist</strong>. What remains is genuinely live: the external <strong>deep-research\nreviews</strong> under <code>docs/</code>.</p>\n<p>For the current science map, use <a href=\"#/read/navigation\"><code>docs/navigation.md</code></a>. For\nengineering, use <a href=\"./engineering/README.md\" target=\"_blank\" rel=\"noopener\" class=\"ll-raw-source\"><code>docs/engineering/README.md</code></a>.</p>\n</blockquote>\n<p>These are imported technical reviews (each distilled from a longer external report) that\ninform the GLIM / Lupine Science program. They are background reading, not primary claims —\nthe claims live in <a href=\"#/read/conjecture-ledger\"><code>conjectures/ledger.md</code></a>.</p>\n<h2 id=\"multi-fidelity-uq-amp-ensembles\">Multi-fidelity, UQ &amp; ensembles</h2><ul>\n<li><a href=\"#/read/glimmer-multifidelity-uq\"><code>multi_fidelity_uq_glimMER_report.md</code></a> — cross-potential\nmeta-analysis and systematic bias correction; the &quot;glimMER&quot; PCA-of-errors correction idea.</li>\n<li><a href=\"#/read/bayesian-active-learning\"><code>bayesian_active_learning_report.md</code></a> — Gaussian-process\nsurrogates and active learning for potential selection at scale.</li>\n<li><a href=\"#/read/weather-climate-ensembles\"><code>weather_climate_ensembles_report.md</code></a> — multi-model\nensemble-weighting strategies ported from climate science.</li>\n</ul>\n<h2 id=\"error-prediction-amp-topology\">Error prediction &amp; topology</h2><ul>\n<li><a href=\"#/read/gnn-error-prediction\"><code>gnn_error_prediction_report.md</code></a> — GNNs predicting where a\npotential fails from local chemical environment.</li>\n<li><a href=\"#/read/tda-error-landscapes\"><code>tda_error_landscapes_report.md</code></a> — persistent homology of\nhigh-dimensional error surfaces.</li>\n</ul>\n<h2 id=\"benchmarking-amp-physical-models\">Benchmarking &amp; physical models</h2><ul>\n<li><a href=\"#/read/phonon-benchmarking\"><code>phonon_benchmarking_report.md</code></a> — phonon spectra as a\n&quot;gold standard&quot; validation; executive summary in <a href=\"#/read/key-findings\"><code>KEY_FINDINGS_SUMMARY.md</code></a>.</li>\n<li><a href=\"#/read/rg-coarsegraining\"><code>rg_coarsegraining_report.md</code></a> — renormalization-group coarse-graining\nfor deriving effective potentials.</li>\n</ul>\n<h2 id=\"theory-amp-information\">Theory &amp; information</h2><ul>\n<li><a href=\"#/read/sloppy-models\"><code>sloppy_models_report.md</code></a> — Fisher-information eigenvalue analysis;\nstiff vs. sloppy directions (the mathematical background for the hyper-ribbon — see\n<a href=\"#/read/error-geometry-objects\"><code>science/objects.md</code></a>).</li>\n<li><a href=\"#/read/info-theoretic\"><code>info_theoretic_report.md</code></a> — Kolmogorov complexity, rate-distortion,\nand entropy applied to model selection.</li>\n</ul>\n<h2 id=\"program-context\">Program context</h2><ul>\n<li><a href=\"#/read/funding-landscape\"><code>funding_landscape_report.md</code></a> — federal materials-informatics /\nUQ funding landscape (2025–2026).</li>\n</ul>\n<blockquote>\n<p>Note: three <code>docs/EXTRACTION_*.md</code> files are one-time content-extraction <em>process logs</em> (with\ndead <code>/sessions/...</code> paths) that are still listed in the public catalog\n(<code>library-site/scripts/catalog.js</code>). Whether they belong on the public site is a content\ndecision, flagged in <a href=\"./decisions/0002-documentation-architecture.md\" target=\"_blank\" rel=\"noopener\" class=\"ll-raw-source\"><code>decisions/0002-documentation-architecture.md</code></a> §5,\nnot changed here.</p>\n</blockquote>\n"}