Reports / English mirror

The report index is a map of review status.

This page tells readers what they can safely do with each report family. Some pages are public reading maps; many project and legacy reports are only source-tracing evidence until reader/editor and academic review pass.

Open full report index Project evidence guide Live publication readiness Redirect indexability GitHub topic readiness Text asset coverage I18N route audit

Report states

Four families, four reading contracts.

491 pages

Project reports

Source-tracing evidence pages for repositories. They are indexable evidence entry points, but their body claims remain review-gated until page-level copy and claim checks pass.

28 pages

Legacy research reports

Older research/project Markdown wrapped as review-gated HTML. Use them to find source trails, not as current conclusions.

5 pages

Survey publication slices

Draft publication slices that still need bilingual same-evidence review and academic claim checks.

2 pages

Paper reports

Paper reading maps and cross-domain synthesis pages. These remain indexable only when they are useful as source-guided reading maps.

QA Audits

High-signal publication gates.

QA · Live publication

Live Publication Readiness Audit

Separates local SEO metadata success from strict HTTPS crawl readiness on the custom domain.

external crawl gate HTTPS BLOCKER Open status

QA · Search Console redirect

Google Redirect Indexability Triage

Classifies Search Console “Page with redirect” rows by checking sitemap final URLs versus HTTP, www, slashless, and legacy GitHub Pages redirect entry points.

redirect gate HTTPS BLOCKER Open redirect triage

QA · GitHub topic

GitHub Topic Indexing Readiness

Checks repository topics, description, homepage, GitHub Search, and topic-page rendering instead of treating Google SEO as GitHub topic evidence.

GitHub discovery TOPIC PASS Open topic status

QA · Text assets

Text Asset Indexability Audit

Answers whether repository text has become indexable public assets by separating public HTML, GitHub README assets, processed-but-unrouted work, raw sources, and external mirrors.

coverage gate INCOMPLETE Open text coverage

QA · I18N routes

Public I18N Route Audit

Checks sitemap routes, HTML language, canonical links, hreflang alternates, the core English path, and Chinese-first long-tail gaps.

bilingual gate INCOMPLETE Open i18n route audit

How to use

Use reports to locate evidence, then verify the claim.

Read report pages as a source trail first, not as a verdict.
Check whether the claim points back to raw GitHub, paper, benchmark, or current project card evidence.
Use review-gated labels as a warning that public-copy review is not finished, even when a page is indexable.
Prefer current project cards and paper pages when legacy report language conflicts with newer evidence.

Paper reading maps

Start with paper maps: they locate sources, not final survey verdicts.

These pages connect cross-domain papers, mechanisms, and source trails. Before citing a claim, return to the original paper, the current paper page, and the limitation notes.

papers · literature map

Self Evolve 跨领域研究图谱

日期：2026-05-22 范围：AutoML + LLM、NAS + LLM、LLM 自我改进、进化计算 + LLM，以及 AI Agent 框架进化。Wiki 检查：当前仓库快照中没有 `wiki/log.md` 内容，因此没有本地历史…

paper source trail not peer-reviewed meta-analysis Read report

papers · literature map

来源索引

- AutoML-GPT：https://arxiv.org/abs/2305.02499 - AutoML-Agent：https://arxiv.org/abs/2410.02958 - Large Language Models Sy…

paper source trail not peer-reviewed meta-analysis Read report

Source-only project evidence

Project evidence pages: trace sources now, cite conclusions later.

These pages keep GitHub sources, classification hints, mechanism links, and limitations discoverable. They are indexable evidence entry points, but page-body claims stay review-gated until public-copy review passes.

projects