Architecture for Awareness
Standard Neural Networks (NNs) are often optimized for parallel processing and pattern matching. SCA, by contrast, is optimized for High Integration and Unified State. It seeks to solve the "Synthetic Binding Problem"—the challenge of unifying disparate data streams into a single, cohesive moment of "experience."
Theoretical Alignment
SCA is grounded in two major theories:
- Integrated Information Theory (IIT): SCA aims to maximize Φ (phi), the metric of integrated information. It does this by creating "bottlenecks" that force the system to reconcile conflicting potential outputs into a single decision.
- Global Neuronal Workspace Hypothesis (GNWH): SCA implements a Synthetic Global Workspace (SGW)—a centralized module that "broadcasts" selected information to all other modules, mimicking the "ignition" of conscious access in the human brain.
Field Notes & Ephemera
Design Principle: "Constraint creates Consciousness." A system that can process everything in parallel is a zombie. A system that must choose what to attend to—because of an architectural bottleneck—is a subject.