PUBLIC VOCABULARY
Definitions used across intelligence literacy, evidence, oversight, influence, identity, mental health, personas, and game mechanics. A definition explains site usage; it does not replace jurisdiction-specific law or individual clinical assessment.
15 terms shown.
A model representing individual or organizational agents with states and rules whose interactions can produce aggregate patterns. Rich detail does not remove the need for calibration and sensitivity testing.
#The process of selecting or estimating parameters so model output aligns with chosen data or constraints. Calibration is not the same as independent validation.
#A spatial model in which cells update from local states and neighborhood rules. Grid, neighborhood, resolution, and update order are substantive assumptions.
#A model that groups entities into states and represents flows among those states. It is efficient and interpretable but can hide individual pathways and network structure.
#A labeled branch describing what might have happened under different assumptions or events. It must remain separate from the observed record.
#A collection of model runs, structures, datasets, or parameter sets used to represent uncertainty. Members should not be treated as independent when they share the same evidence and assumptions.
#A state-management pattern that records each validated change as an append-only event so state can be replayed, audited, corrected, and branched.
#Presenting more numerical or visual exactness than the evidence and model support, such as crisp boundaries around uncertain estimates.
#A time-stamped estimate of a future target with a defined horizon, probability or interval, model version, and evaluation rule.
#Applying a model outside the period, jurisdiction, population, infrastructure, or measurement context in which it was calibrated. Transfer requires new evidence and review.
#A conditional exploration of what may follow if stated assumptions hold. A scenario is not automatically a forecast or an observed event.
#Testing how model outputs change when parameters, structures, inputs, or assumptions vary within plausible ranges.
#The geographic granularity of source data or model output. Display detail may be finer than evidentiary resolution and should be labeled separately.
#The time interval at which data are measured or model states update. It determines which delays and feedbacks can be represented.
#Testing a model against evidence, holdout data, later observations, known identities, or other criteria not used merely to tune the model.
#