When drawing 1,000 z from the priorp(z)of the latent space learned by PLAS, only4%of the samples are decoded as high-return actions, while inLAPO,45%ofthedecoded actions arehigh-return actions.

Alldatausedispublic.] (b) Did you specify all the training details (e.g., data splits, hyperparameters, how they werechosen)? A.1 TrainingDetails In our experiments, the classifierfθ is a 8-layer MLP with 128 hidden dimensions per layer.

LICD has 577 features and labels. The dataset provides several ML research opportunities; we discuss two demonstration experiments.

Everyattemptwasmade to state the full set of assumptions behind each calculation (e.g., the footnote on page5).

Figure 1: The thick line in (a) indicates minimal pointsofC.Thethicklinein(b)indicatesweakly minimalpointsofC.

Themodel outputs the normal distribution for the observations, conditional on hidden stateh(t). Since only some features are observed at atime, we mask out the missing values when calculatingLpre. We denote our predicted distribution withppre,and predicted distribution after updating the state with ppost.

These representations are concatenated and passed toaone-hidden-layer feed-forward network for binary classification.

Yes (c) Did you report error bars (e.g., with respect totherandom seed after running experiments multiple times)?