ConceptE2D
Estimation-to-Decisions
A meta-algorithm framework that reduces a decision problem to the problem of estimating outcome distributions. It estimates distributions through an online regression oracle and then makes decisions based on the estimates, and it applies to sequential decision-making under general function approximation.
Timeline (last 90 days)
Oct 15
E2D.PMSide is designed based on the Estimation-to-Decisions (E2D) meta-algorithm proposed by Foster et al., using an online regression oracle to estimate the outcome distribution based on historical (context, action, feedback symbol) triples, avoiding explicit enumeration over every possible outcome distribution.
Unverified50%