Volume I · Computational laboratories

Seven controlled pressure tests.

The laboratories make the computational test layer legible: each isolates a question about reliability, repair, viability or continuity under disturbance.

Lab I

Recursive Self-Model Reliability Improves Viability Under Self-Relevant Perturbation

Question: Does recursive self-model reliability support viable behaviour when disturbance reaches the system’s own state and model?
Purpose: Compares agents under self-relevant perturbation to isolate the role of recursive reliability in maintaining viability.

Canonical record ↗

Lab II

Channel-Decomposed Reliability and the Limits of Reliability Tracking Alone

Question: Is tracking reliability across separate channels sufficient when the system cannot act effectively on those estimates?
Purpose: Separates measurement from intervention to test the limits of reliability tracking without an adequate repair mechanism.

Canonical record ↗

Lab III

Actionable Reliability and Cost-Sensitive Repair in Viability-Constrained Agents

Question: How should a system choose repair when reliability, consequence, cost and viability compete?
Purpose: Places repair selection under explicit cost and viability constraints so reliability must become actionable.

Canonical record ↗

Lab IV

Minimal Reliability Thresholds in Viability-Constrained Agents

Question: What minimum reliability conditions are compatible with continued viable behaviour?
Purpose: Searches the lower boundary at which reliability remains sufficient for viability-constrained action.

Canonical record ↗

Lab V

Recursive Reliability Under Estimator Corruption

Question: What happens when the mechanism estimating reliability is itself damaged or corrupted?
Purpose: Stresses the recursive layer directly to examine whether repair remains possible when confidence estimation becomes unreliable.

Canonical record ↗

Lab VI

Neuro-Symbolic Persistence in Language-Like Agents

Question: Can symbols, memory and action remain coordinated in a language-like neuro-symbolic setting?
Purpose: Moves the persistence problem into a language-like environment where symbolic commitments must survive changing context.

Canonical record ↗

Lab VII

Autobiographical Continuity Under Memory Injection and Goal Drift

Question: Can an agent preserve autobiographical continuity when memories are injected and goals change?
Purpose: Places memory and goal structure under direct pressure to test continuity across altered histories and priorities.

Canonical record ↗

Publications preserves the compact DOI ledger; this page explains the experimental map.

Open the laboratory ledger →