AI Glossary
The complete dictionary of Artificial Intelligence
Do-calculus Criterion
Set of algebraic rules developed by Judea Pearl for manipulating causal graphs and determining whether a causal quantity can be identified from observational data.
Average Treatment Effect on the Treated (ATT)
Average effect of treatment calculated only on the subgroup of individuals who actually received the treatment, relevant for assessing impact on a target population.
Randomization
Process of randomly assigning units to different treatment groups, aiming to eliminate confounding biases and isolate the pure causal effect.
Causal Moderation
Phenomenon where the effect of a cause on an outcome varies depending on the levels of a third variable, called a moderator, indicating causal interaction.
Causal Instrument
Variable used in instrumental variables, correlated with the treatment but not directly with the outcome, allowing estimation of a causal effect in the presence of unmeasured confounding.
Stabilization Weights
Weighting technique used in causal inference to reduce the variance of estimators by adjusting the inverse probability of treatment weights.