Sampling Strategies for GAN Synthetic Data

TL;DR

This paper introduces novel sampling strategies for GAN-generated synthetic data, leveraging GAN training parameters and reinforcement learning to improve the quality and diversity of data used for training deep CNNs, leading to better performance.

Contribution

It presents simple, effective sampling methods that utilize GAN discriminator confidence and RL-based subset selection to enhance CNN training with synthetic data.

Findings

Sampling synthetic data improves CNN accuracy.

Using discriminator confidence enhances data quality.

Reinforcement learning effectively selects meaningful synthetic examples.

Abstract

Generative Adversarial Networks (GANs) have been used widely to generate large volumes of synthetic data. This data is being utilized for augmenting with real examples in order to train deep Convolutional Neural Networks (CNNs). Studies have shown that the generated examples lack sufficient realism to train deep CNNs and are poor in diversity. Unlike previous studies of randomly augmenting the synthetic data with real data, we present our simple, effective and easy to implement synthetic data sampling methods to train deep CNNs more efficiently and accurately. To this end, we propose to maximally utilize the parameters learned during training of the GAN itself. These include discriminator's realism confidence score and the confidence on the target label of the synthetic data. In addition to this, we explore reinforcement learning (RL) to automatically search a subset of meaningful synthetic examples from a large pool of GAN synthetic data. We evaluate our method on two challenging face attribute classification data sets viz. AffectNet and CelebA. Our extensive experiments clearly demonstrate the need of sampling synthetic data before augmentation, which also improves the performance of one of the state-of-the-art deep CNNs in vitro.