Off Policy Evaluation

Tasks

D4RL Currently supports off-policy evaluation on the following tasks:

Policy Name	Environment Name	Returns (undiscounted)
cheetah-medium	halfcheetah-medium-v0	3985
cheetah-random	halfcheetah-random-v0	-199
cheetah	halfcheetah-expert-v0	12330
hopper-medium	hopper-medium-v0	2260
hopper-random	hopper-random-v0	1257
hopper	hopper-expert-v0	3624
walker2d-medium	walker2d-medium-v0	2760
walker2d-random	walker2d-random-v0	896
walker2d	walker2d-expert-v0	4005

Loading Policies

Policies can be downloaded from http://rail.eecs.berkeley.edu/datasets/offline_rl/ope_policies.

Policies are provided in two formats:

• [policy_name]_params.pkl: A pickle file (that can be loaded with joblib) that contains numpy arrays containing the network weights.
• A pair of ONNX files:
  • [policy_name]_params.log_prob.onnx takes as input (observations, gaussian noise) and outputs (actions, mean, log_std). The mean and log_std are the parameters of the gaussian before the tanh is applied.
  • [policy_name]_params.sampler.onnx takes as input (observations, actions) and outputs log_probs.

Metrics

The d4rl.ope module contains metrics for off-policy evaluation. Each metric takes in policy string ID (defined under then polic name column in the Tasks table), and can be computed using discounted or undiscounted returns by passing in a discounted=True/False flag. We provide the following metrics:

• ope.ranking_correlation_metric(policy_ids): Computes Spearmans' ranking correlation coefficient for a list of policy ids.
• ope.precision_at_k_metric(policy_ids, k, n_rel): Computes precision@k.
• ope.recall_at_k_metric(policy_ids, k, n_rel): Computes recall@k.
• ope.value_error_metric(policy_id, value): Computes error in the estimated policy value.
• ope.policy_regret_metric(policy_id, expert_ids): Computes the difference between the best return among the experts and the selected policy's returns.