GenDICE

Implementation: crl.estimators.gen_dice.GenDICEEstimator

Assumptions

• Sequential ignorability
• Markov property

Requires

• TrajectoryDataset with state_space_n

Algorithm

Generalized density-ratio estimation with linear features and regularization.

Uncertainty

• CI reported but may be unreliable for density-ratio estimators; interpret cautiously.

Failure modes

• Sparse state-action coverage can destabilize ratio estimates.

Minimal example

from crl.estimators.gen_dice import GenDICEEstimator

report = GenDICEEstimator(estimand).estimate(dataset)

References

• Yang et al. (2020)