Dec 2, 2021 · We propose a novel framework for learning policies that generalize to the target population. For this, we characterize the difference between the training data ...

May 20, 2022 · This paper learns the problem of learning policies that generalize to the target population. They characterize the difference between training ...

We prove that, if the uncertainty set is well-specified, our policies generalize to the target population as they can not do worse than on the training data.

We briefly explain how Theorem 1 is proven in the case in which we do not have access to the true nuisance functions (using the results from Athey and Wager [ ...

An efficient algorithm based on a convex-concave procedure is derived and it is proved that, if the uncertainty set is well-specified, policies generalize ...

To address this challenge, we propose a novel framework for learning policies that generalize to the target population. For this, we characterize the difference ...

Apr 23, 2023 · In this work, we aim to identify and learn policies that are robust to certain failures of external validity, namely sampling bias in the ...

In. Section 3 we present a new categorization of learning methods that is useful for characterizing their behavior under sample selection bias and study how a ...

However, in general bv might be a biased estimator for v, and the bias might not even go down with the number of samples n in our dataset. For instance, if we ...

Here we generalize eligibility traces to off-policy learning, in which one learns about a policy dif- ferent from the policy that generates the data. Off-policy ...