The Present Model Theory of Consciousness
A testable theory that experience is the operation of a running, self-correcting model of the present.
Self-correcting models of the present
RQVST Research
RQVST Research studies one question: how does a system keep an accurate, live picture of a changing world when its inputs drop out or go wrong? Brains do this constantly. When you glance away from a busy street, the people in it don't vanish; in your head they keep moving, and the next glance corrects you. We study the organization behind that ability, in theory and in simulation, and we build its principles into RQVST's decision-intelligence agents.
How we work
- We state in advance what would count against us. Each theory names the results that would refute it.
- We preregister. Simulation designs, measures and pass criteria are fixed before anything runs.
- We publish everything. Code, data, preregistrations and audit notes for every round are openly archived.
- We say what our results don't show. Every paper states its limits next to its claims.
Publications
Present Model Hypothesis: code, data, preregistrations and audit record
Open code, data, preregistrations and audit notes for every round of the preregistered simulations.
The Present Model Hypothesis: defining, measuring and testing a candidate organization for consciousness
The earlier preprint that introduces the Present Model Hypothesis.
From research to product
The principle we study, parts that each model the whole, carry their picture forward through gaps, and correct one another, also makes software more reliable when data is incomplete or wrong. RQVST's agents are built on it.