We list the best hyperparameter configurations for our methods in each dataset. In Tables 11 to 16, we list all results. The reason why we include the sub-optimal versions ofIT-GANis that one can easily decrease the attack success score by selecting sub-optimal models (and thereby, potentially sacrificing the synthesis quality). As shown in Table 23,IT-GAN(L)shows the lowest attack success scores overall. Wealso compare with various sub-optimal versions ofIT-GAN.While trainingIT-GAN,wecould obtain several sub-optimal versions around the epoch where we obtainIT-GAN and call them as IT-GAN(Sk)withk {1,2,3}.

We summarize the statistics of our datasets as follows: 1. Adult We list the best hyperparameter configurations for our methods in each dataset. In Tables 11 to 16, we list all results. We report the average of the earth mover's distance (EMD) We report more tables about the parameter sensitivity in Tables 17 to 22. One easy way to protect sensitive information in the original data is to use these sub-optimal versions. However, they typically fail to obtain a good trade-off between synthesis quality and information protection.

Tabular data synthesis has received wide attention in the literature.

Remarkable progress has been achieved in synthesizing photo-realistic images with generative adversarial networks (GANs). Recently, GANs are utilized as the training sample generator when obtaining or storing real training data is expensive even infeasible. However, traditional GANs generated images are not as informative as the real training samples when being used to train deep neural networks. In this paper, we propose a novel method to synthesize Informative Training samples with GAN (IT-GAN). Specifically, we freeze a pre-trained GAN model and learn the informative latent vectors that correspond to informative training samples. The synthesized images are required to preserve information for training deep neural networks rather than visual reality or fidelity. Experiments verify that the deep neural networks can learn faster and achieve better performance when being trained with our IT-GAN generated images. We also show that our method is a promising solution to dataset condensation problem.

Tabular data synthesis has received wide attention in the literature. This is because available data is often limited, incomplete, or cannot be obtained easily, and data privacy is becoming increasingly important. In this work, we present a generalized GAN framework for tabular synthesis, which combines the adversarial training of GANs and the negative log-density regularization of invertible neural networks. The proposed framework can be used for two distinctive objectives. First, we can further improve the synthesis quality, by decreasing the negative log-density of real records in the process of adversarial training. On the other hand, by increasing the negative log-density of real records, realistic fake records can be synthesized in a way that they are not too much close to real records and reduce the chance of potential information leakage. We conduct experiments with real-world datasets for classification, regression, and privacy attacks. In general, the proposed method demonstrates the best synthesis quality (in terms of task-oriented evaluation metrics, e.g., F1) when decreasing the negative log-density during the adversarial training. If increasing the negative log-density, our experimental results show that the distance between real and fake records increases, enhancing robustness against privacy attacks.