Canonical Artifact Name: Education OS Kernel v1.0 — Human Outcome Physics Kernel
Artifact Class: Executable Human Systems Diagnostic Kernel
Artifact Version: v1.0
Artifact Authority: eduKate Singapore
Canonical ID: EducationOS-Kernel-v1.0
Block 1 — System Boundary
System: A Primary-level class (or tuition group) showing a sustained plateau in English reading comprehension over a 10–12 week term.
Included: vocabulary knowledge, comprehension strategies, inference ability, attention during reading, question analysis, feedback cycles, and weekly practice routine.
Excluded: composition writing outcomes, other subjects, and long-term outcomes beyond the term.
Block 2 — D/L/T Score Table (0–5)
| Axis | Score | Justification |
|---|---|---|
| D (Depth) | 2/5 | Students can understand surface meaning but struggle with vocabulary nuance and inference. |
| L (Load) | 3/5 | In normal classwork, performance is stable; under exam conditions, weaker students decline. |
| T (Transfer) | 1/5 | Strategies taught (highlighting, summarising, clue-finding) do not transfer to unseen passages. |
Block 3 — Binding Constraints (e)
- Time constraint: limited curriculum time per week for deep reading practice.
- Attention constraint: inconsistent home reading habits reduce repeated exposure to varied texts.
Block 4 — Trajectory Physics (t)
Rise: Early improvements occur when passages are familiar or guided.
Stall: Scores plateau when texts change genre, topic, or difficulty.
Regression: Under timed practice, students revert to guessing or copying lines without inference.
Lock-in risk: Repeated “practice without transfer” trains shallow pattern-matching and hardens the plateau.
Block 5 — Failure Signature
Primary failure: T-FAIL (Transfer failed first)
Reason: Students can perform strategies with guidance but cannot reliably apply them to unseen texts under time and novelty.
Block 6 — Recovery Mode
Selected: Transfer Repair.
Block 7 — Retest Probe
Probe A – Unseen passage transfer
Weekly: one unseen passage (new topic + genre), strict time limit.
Pass: ≥70% of students improve by at least 1 band (or +10% score) within 6 weeks, without teacher hints.
Probe B – Inference integrity
Weekly: 5 inference questions that require evidence-based reasoning.
Pass: Students must cite the exact phrase/sentence that supports their answer (not copy random lines). ≥70% accuracy.
Probe C – Vocabulary-in-context transfer
Weekly: 8 target words taught in class, tested only inside a new passage (not isolated definitions).
Pass: ≥70% of students correctly infer meaning from context in 6 weeks.
Re-run rule:
If probes pass for 3 consecutive weeks, raise T by +1 and re-execute the kernel.
Canonical ID: EducationOS-run-003-v1.0