# Multi-objective training of Generative Adversarial Networks with multiple discriminators

arXiv: https://arxiv.org/abs/1901.08680 · HTML: https://mlsanity.com/paper/1901.08680/

## Key facts

- Cited by 24 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 24 Jan 2019.

## Built on by (most cited first, 24 of 24)

- [Projected GANs Converge Faster](https://mlsanity.com/paper/2111.01007/) (2021)
- [Ensembling Off-the-shelf Models for GAN Training](https://mlsanity.com/paper/2112.09130/) (2021)
- [Fair Generative Models via Transfer Learning](https://mlsanity.com/paper/2212.00926/) (2022)
- [NitroFusion: High-Fidelity Single-Step Diffusion through Dynamic Adversarial Training](https://mlsanity.com/paper/2412.02030/) (2024)
- [Multi-Objective Learning to Predict Pareto Fronts Using Hypervolume Maximization](https://mlsanity.com/paper/2102.04523/) (2021)
- [Rethinking Robustness Assessment: Adversarial Attacks on Learning-based Quadrupedal Locomotion Controllers](https://mlsanity.com/paper/2405.12424/) (2024)
- [Complex Momentum for Optimization in Games](https://mlsanity.com/paper/2102.08431/) (2021)
- [MCL-GAN: Generative Adversarial Networks with Multiple Specialized Discriminators](https://mlsanity.com/paper/2107.07260/) (2021)
- [Self-Supervised GANs with Label Augmentation](https://mlsanity.com/paper/2106.08601/) (2021)
- [Brain Multigraph Prediction using Topology-Aware Adversarial Graph Neural Network](https://mlsanity.com/paper/2105.02565/) (2021)
- [Hierarchical Semantic Regularization of Latent Spaces in StyleGANs](https://mlsanity.com/paper/2208.03764/) (2022)
- [A Spectral Enabled GAN for Time Series Data Generation](https://mlsanity.com/paper/2103.01904/) (2021)
- [Multi-objective Deep Learning: Taxonomy and Survey of the State of the Art](https://mlsanity.com/paper/2412.01566/) (2024)
- [Multi-objective evolutionary GAN for tabular data synthesis](https://mlsanity.com/paper/2404.10176/) (2024)
- [DuelGAN: A Duel Between Two Discriminators Stabilizes the GAN Training](https://mlsanity.com/paper/2101.07524/) (2021)
- [OPT-GAN: A Broad-Spectrum Global Optimizer for Black-box Problems by Learning Distribution](https://mlsanity.com/paper/2102.03888/) (2021)
- [Generating Adversarial Examples with Task Oriented Multi-Objective Optimization](https://mlsanity.com/paper/2304.13229/) (2023)
- [Towards convergence to Nash equilibria in two-team zero-sum games](https://mlsanity.com/paper/2111.04178/) (2021)
- [Stylized Projected GAN: A Novel Architecture for Fast and Realistic Image Generation](https://mlsanity.com/paper/2307.16275/) (2023)
- [The Bid Picture: Auction-Inspired Multi-player Generative Adversarial Networks Training](https://mlsanity.com/paper/2403.13866/) (2024)
- [HMAMP: Hypervolume-Driven Multi-Objective Antimicrobial Peptides Design](https://mlsanity.com/paper/2405.00753/) (2024)
- [Towards Fair and Robust Face Parsing for Generative AI: A Multi-Objective Approach](https://mlsanity.com/paper/2502.04391/) (2025)
- [Auxiliary Discrminator Sequence Generative Adversarial Networks (ADSeqGAN) for Few Sample Molecule Generation](https://mlsanity.com/paper/2502.16446/) (2025)
- [Curated Synthetic Data Doesn't Have to Collapse: A Theoretical Study of Generative Retraining with Pluralistic Preferences](https://mlsanity.com/paper/2605.07724/) (2026)

## Abstract (arXiv)

Recent literature has demonstrated promising results for training Generative Adversarial Networks by employing a set of discriminators, in contrast to the traditional game involving one generator against a single adversary. Such methods perform single-objective optimization on some simple consolidation of the losses, e.g. an arithmetic average. In this work, we revisit the multiple-discriminator setting by framing the simultaneous minimization of losses provided by different models as a multi-objective optimization problem. Specifically, we evaluate the performance of multiple gradient descent and the hypervolume maximization algorithm on a number of different datasets. Moreover, we argue that the previously proposed methods and hypervolume maximization can all be seen as variations of multiple gradient descent in which the update direction can be computed efficiently. Our results indicate that hypervolume maximization presents a better compromise between sample quality and computational cost than previous methods.