Training GANs (including cGANs), however, are known to be often hard and highly unstable [46]. Numerous techniques have thus been proposed to tackle the issue from different angles, e.g., improving architectures [32, 56, 7], losses and regularizers [16, 38, 20] and other training heuristics [46, 51, 8].

We adjust this regularization term so that it concentrates on increasing the entropy of the particular pairs of binary vectors that are not correlated in high-dimensional space.

Reinforcement learning is well suited for optimizing policies of recommender systems.

IS (( 11.095.20 0.008 Default (nodebiasing) 5.1658.70

To this end, we propose an analogous score function called the "Concretescore",ageneralizationofthe(Stein)scorefordiscretesettings.

Although suchmodern methods allowedlearning successful semantic segmentation systems, their training requires large-scale labeled datasets with usually a need for pixel-level annotations.

Generative adversarial networks (GANs) typically require ample data for training in order to synthesize high-fidelity images.

For example, transactions infinancial systems, serverlogs, and user activity traces can allnaturally be represented as discrete events in continuous time.

In this document, we provide details and supplementary materials that cannot fit into the main manuscript due to the page limit. The specific form ofcenter distribution isunknown, but we can still train a generatorG to approximate it. IfR(G,D,T)),wechooseλ=0, i.e., no restriction onR(G,D,T)), to obtain the minimal cost. IfR(G,D,T)) >, then a large λshould be applied as apenalization. According to the derivation of Eq. (3), we obtain arelaxed versionoftheintractableEq.(2),expressedasfollows: min Inknowledge distillation, student models arecrafted using unlabeled datasets, where only thesoft targets from teachers are utilized.

Inparticular, GANs generate samples of minor groups with low fidelity or even fail to generate such samples, exhibitingthemodecollapse[43].