Instrumental Variable
In causal inference, a tool that isolates confounding through a variable affecting the outcome only via the treatment variable; it is used to identify causal effects when unmeasured confounding is present.
Timeline (last 90 days)
In offline reinforcement learning, unmeasured confounding makes the behavior policy correlated with outcomes, so directly using logged data for policy evaluation or optimization produces bias, requiring auxiliary structures such as instrumental variables, negative controls or proxy variables to achieve causal identification.
The paper combines transferability theory with instrumental variables, negative controls and proxy variables to build a unified framework for identifying causal quantities in the target domain.
In research on technology adoption and employment change, causal identification requires an exogenous shock as an instrumental variable, but in reality the exclusion restriction is very hard to satisfy.
Methods such as instrumental variables (IV), regression discontinuity design (RDD), and difference-in-differences (DID) are widely used to identify causal relationships from observational data.
All Facts (4)
In offline reinforcement learning, unmeasured confounding makes the behavior policy correlated with outcomes, so directly using logged data for policy evaluation or optimization produces bias, requiring auxiliary structures such as instrumental variables, negative controls or proxy variables to achieve causal identification.
50%UnverifiedThe paper combines transferability theory with instrumental variables, negative controls and proxy variables to build a unified framework for identifying causal quantities in the target domain.
50%UnverifiedIn research on technology adoption and employment change, causal identification requires an exogenous shock as an instrumental variable, but in reality the exclusion restriction is very hard to satisfy.
50%UnverifiedMethods such as instrumental variables (IV), regression discontinuity design (RDD), and difference-in-differences (DID) are widely used to identify causal relationships from observational data.
50%