# Improved generator objectives for GANs

arXiv: https://arxiv.org/abs/1612.02780 · HTML: https://mlsanity.com/paper/1612.02780/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 8 Dec 2016.

## Built on by (most cited first, 12 of 12)

- [An Introduction to Neural Data Compression](https://mlsanity.com/paper/2202.06533/) (2022)
- [One-step Diffusion Models with f-Divergence Distribution Matching](https://mlsanity.com/paper/2502.15681/) (2025)
- [Solution of Physics-based Bayesian Inverse Problems with Deep Generative Priors](https://mlsanity.com/paper/2107.02926/) (2021)
- [Speaker Generation](https://mlsanity.com/paper/2111.05095/) (2021)
- [DigGAN: Discriminator gradIent Gap Regularization for GAN Training with Limited Data](https://mlsanity.com/paper/2211.14694/) (2022)
- [Soft-Di[M]O: Improving One-Step Discrete Image Generation with Soft Embeddings](https://mlsanity.com/paper/2509.22925/) (2025)
- [Addressing GAN Training Instabilities via Tunable Classification Losses](https://mlsanity.com/paper/2310.18291/) (2023)
- [(αD,αG)-GANs: Addressing GAN Training Instabilities via Dual Objectives](https://mlsanity.com/paper/2302.14320/) (2023)
- [An Open Patch Generator based Fingerprint Presentation Attack Detection using Generative Adversarial Network](https://mlsanity.com/paper/2306.03577/) (2023)
- [Text Generation with Deep Variational GAN](https://mlsanity.com/paper/2104.13488/) (2021)
- [TransductGAN: a Transductive Adversarial Model for Novelty Detection](https://mlsanity.com/paper/2203.15406/) (2022)
- [Neural Posterior Regularization for Likelihood-Free Inference](https://mlsanity.com/paper/2102.07770/) (2021)

## Abstract (arXiv)

We present a framework to understand GAN training as alternating density ratio estimation and approximate divergence minimization. This provides an interpretation for the mismatched GAN generator and discriminator objectives often used in practice, and explains the problem of poor sample diversity. We also derive a family of generator objectives that target arbitrary f-divergences without minimizing a lower bound, and use them to train generative image models that target either improved sample quality or greater sample diversity.