|
|
||
|---|---|---|
| .github/workflows | ||
| artwork | ||
| configs | ||
| docs | ||
| domain-packs | ||
| src/didactopus | ||
| tests | ||
| .gitignore | ||
| Dockerfile | ||
| LICENSE | ||
| Makefile | ||
| README.md | ||
| docker-compose.yml | ||
| pyproject.toml | ||
README.md
Didactopus
Didactopus is a local-first AI-assisted autodidactic mastery platform.
This revision adds an evidence-driven mastery engine on top of the adaptive learner model.
Added in this revision
- evidence record models
- rubric-style evidence scoring
- concept mastery updates from accumulated evidence
- weak-concept resurfacing
- automatic learner state updates from evidence bundles
- project evidence integration
- CLI demonstration of evidence-driven progression
- tests for mastery promotion and resurfacing
Why this matters
Didactopus no longer needs mastery to be supplied only by hand. It can now begin to infer learner state from observed evidence such as:
- explanation quality
- problem-solving performance
- project completion
- transfer-task performance
That is a necessary step toward a genuine mastery engine.