Critical Attention Systems

IRAM-Ω-Q Papers

A sequence of preprints developing the IRAM-Ω-Q framework for uncertainty regulation, anticipatory control, hysteresis, and adaptive artificial agency.

Paper 1

IRAM-Ω-Q: A Computational Architecture for Uncertainty Regulation in Artificial Agents

Introduces the IRAM-Ω-Q architecture, including entropy regulation, coherence-gap dynamics, susceptibility maps, and the basic regulation-first versus disturbance-first comparison.

FrameworkArtificial AgencyEntropy Regulation
Read Paper 1 on arXiv
Paper 2

When Regulation Has Memory: Hysteresis and Control Burden in Artificial Agency

Examines how regulation timing changes the burden of control in an adaptive artificial agent, with emphasis on hysteresis, path dependence, and anticipatory stabilization.

HysteresisRegulatory BurdenControl Timing
Read Paper 2 on arXiv
Paper 3

Intermittent Control Is Not Diluted Control: A Switching Effect in Artificial Agency

Studies switching between anticipatory and reactive control, showing that intermittent anticipatory access can produce a nonlinear switching effect rather than a simple average of fixed-control regimes.

Intermittent ControlSwitchingArtificial Agents
Read Paper 3 on arXiv
Paper 4

Regulation-Induced Instability in IRAM-Ω-Q

Explores the case where regulation itself becomes a source of disturbance, distinguishing external noise, internal fluctuations, and control-induced instability.

Regulatory CostNoiseInstability
Coming soon