Conscious contents as inferred states at the continuous–discrete interface (Active Inference conjecture)

Establish whether the contents of consciousness—including exteroceptive and interoceptive experiences and the brain’s internal states—correspond to inferred hidden states at the interface between continuous sensory computations and discrete counterfactual policy selection processes within active inference models.

Background

In proposing a minimal theory of consciousness grounded in active inference, the authors link conscious contents to inferred hidden states produced by generative models. They emphasize a specific interface: continuous sensory computations feeding into discrete, temporally deep, counterfactual policy selection mechanisms that support report and goal-directed behavior.

By formulating this as a conjecture in the theory’s core, the authors invite empirical and modeling work to confirm or refute whether conscious contents systematically map onto these inferred states at the specified interface.