State
Numerical representation of the system at a point in time.
QGI architecture
From language interface to quantitative cognition.
QGI treats quantitative state, computation, models, uncertainty, experiments and evidence as native cognitive primitives rather than external utilities.
QGI cognition visual
A quantitative field moving through an intelligence architecture.
The canonical QGI visual maps cognitive functions rather than biological lobes: observe, model, simulate, predict, optimise, experiment, verify and remember.
LIVE QUANTITATIVE FIELDQUANTITATIVE COGNITIONObjective → measurable state → evidence-bearing decision
HUMAN / MACHINE OBJECTIVE
↓Semantic interpretation
language can remain an interface
↓language can remain an interface
QGI CONTROL PLANE
HypothesisPlanningExperiment designState estimation
↓QUANTITATIVE MODEL PLANE
LQMSolverSimulatorInstrument / lab
↓EVIDENCE PLANE
UncertaintyProvenanceReproducibilityVerification
↓SCIENTIFIC MEMORY
↺ next hypothesis / next experimentEight native primitives
A QGI system should reason in the language of measurable systems.
Dynamics
Models of how state changes under time, actions and disturbances.
Constraints
Physical, financial, scientific and operational boundaries.
Objectives
Explicit functions to predict, optimise, validate or discover.
Uncertainty
Calibrated representation of what the system does not know.
Experiments
Selection of measurements, simulations or tests that reduce uncertainty.
Evidence
Lineage from observations and methods to quantitative conclusions.
Memory
Persistent hypotheses, results, failures and unresolved questions.
Native federation
QGI should choose and compose quantitative engines rather than pretend one model is universal.
STATISTICAL
Estimation & forecasting
PHYSICAL
Simulation & digital twins
OPTIMISATION
Search & constrained decisions
SCIENTIFIC ML
Surrogates & learned operators
FORMAL
Proofs & hard verification
LAB / SENSOR
Measured-world feedback
Architecture rule
The language model may propose. The quantitative system must establish the number.
QGI can use language models for interaction, literature synthesis, code generation and hypothesis formation. Numerical authority remains with explicit quantitative computation, measurement, uncertainty and evidence.