Invariant Causal Prediction
A causal inference method that identifies true causal variables by finding predictive relationships that remain stable across multiple distinct environments. Its core assumption is that genuine causal mechanisms are environment-independent, whereas spurious correlations driven by confounders vary across environments.
Timeline (last 90 days)
Invariant Causal Prediction (ICP) and causal representation learning are research directions that leverage causal mechanisms to address distributional shifts.
The core idea of the invariant causal prediction (ICP) method is to identify predictive relationships that remain stable across multiple different environments
All Facts (2)
Invariant Causal Prediction (ICP) and causal representation learning are research directions that leverage causal mechanisms to address distributional shifts.
50%UnverifiedThe core idea of the invariant causal prediction (ICP) method is to identify predictive relationships that remain stable across multiple different environments
50%