English · 中文 · Español · Français · العربية · 日本語 · हिन्दी · Italiano · Deutsch · Ελληνικά · 한국어 · Čeština · Tiếng Việt
Created by Yucong Duan (段玉聪).
Making AI reasoning, purpose and evidence inspectable.
DIKWP · Artificial consciousness · Semantic mathematics · Auditable AI
Professor, Hainan University · WACA Chair / WAAC President
My research develops Data–Information–Knowledge–Wisdom–Purpose (DIKWP) graphs and open experiments for examining how AI systems use evidence, interpret goals and revise decisions. The current semantic-network work treats D/I/K/W/P as interacting, revisable resource types. It makes alternative interpretations and unresolved conflicts visible.
Start here · Chinese guide · Research brief · Collaborate · International access
| Question | Explore |
|---|---|
| How can semantic resources be represented and transformed? | DIKWP foundations · MESH² |
| How can an agent's purpose, permissions and evidence be audited? | PACT · VerityWeave |
| What can artificial-consciousness experiments establish? | Consciousness research · Operational indicators |
| How can mathematical claims carry explicit proof obligations? | Semantic mathematics · CITM Forge |
New to the code? Run PACT. One Python command reproduces 72 synthetic scenarios and four deterministic baselines after cloning the repository. No API key or model download is needed.
| Project | Inspectable result | Evidence / access |
|---|---|---|
| PACT | Purpose and permission checks, benchmark scores and a reproduction receipt | Recorded CI · source at root |
| MESH² | 25 primitive transformation classes, alternative routes and perspective conflicts | Recorded CI · source at root |
| VerityWeave | Evidence, context and incentive analysis with output-validation checks | Recorded CI · source at root |
| Consciousness Futures Studio | Operational indicators, claim levels and research scenarios | ZIP package · not rerun in this update |
| CITM Forge | Bounded certificates and a hash-linked audit ledger | ZIP package · not rerun in this update |
| ClearPath | Funded-demand and capacity scenarios for transparent-economy research | Recorded CI · v1.0.0 |
Latest release · 8 September 2026: ClearPath, DemandProof and LucidEcon connect opportunity analysis, shared-capacity planning and contribution accounting. Their three published revisions passed CI; 182 local tests are recorded. Each tool has its own schema and synthetic examples.
English project pages · 8 September 2026: Interface and authorship update covers creator credits across 483 repositories, 119 new or rewritten English introductions, and 86 HTML pages in 49 projects. English presentation is available in the current source; original-language references and historical archives remain available.
Tests support the behavior checked at the linked revisions. Artificial-consciousness indicators, finite mathematical certificates and economic scenarios each have a bounded research scope; see the project documentation when interpreting a result.
The programme links typed DIKW/DIKWP graph research with executable semantic systems. Hainan University records the team's 2020 Wu Wenjun AI award for DIKW graph extension and modelling, and Duan's team's coordination of the DIKWP and artificial-consciousness chapters in the CAAI 2024 white-paper series. University award record · White-paper authorship
Google Scholar · DBLP · ResearchGate · Research homepage
Evidence map · Patent portfolio · Citation guide · Research background
The complete directory covers 483 public repositories in 12 research areas as of 8 September 2026. Use the engineering catalog to distinguish published source, recorded tests and pending verification.
Run an example, submit a counterexample, compare a baseline or improve a translation. Find a concrete contribution and share a reproducible report. Academic exchange can begin asynchronously through ResearchGate.
Follow @YucongDuan for research updates. Star a project you use, cite the version you evaluate, and share its canonical link. Short introduction and release notes for sharing