TU-GAN: A Cooperative Game Theory Framework with Transferable Utility for Stable GAN Training

Q: What does this research mean for the field?

A cooperative game theory framework with transferable utility (TU-GAN) enables more stable training of Generative Adversarial Networks compared to standard approaches. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

Hamidreza Navidi; Hossein Yarahmadi

doi:10.5281/zenodo.22151208

August 29, 2026Open Access

TU-GAN: A Cooperative Game Theory Framework with Transferable Utility for Stable GAN Training

View Full Paper

Ask AI

Bookmark

Share

Key Points

To introduce and evaluate a cooperative game-theoretic framework using transferable utility to stabilize the training dynamics of generative adversarial networks.
Implemented standard GAN and Wasserstein GAN (WGAN) alongside cooperative variants (TU-GAN and TU-WGAN).
Evaluated sample quality and diversity using Fréchet Inception Distance (FID) and Inception Score (IS) evaluation pipelines.
Conducted multi-seed statistical analyses reporting mean, standard deviation, t-tests, and Cohen's d effect sizes.
Formulated cooperative objective functions enabling transferable utility payoffs between the generator and discriminator.
Demonstrated measurable stability improvements across multiple random seeds relative to non-cooperative GAN and WGAN baselines.

Ask about this paper

Answers are grounded in this paper and cite their sources.

Implication

Computational study demonstrates stabilized training in generative adversarial networks via transferable utility, suggesting cooperative game theory enhances generative modeling.

Abstract

This repository contains the official implementation of the paper "TU-GAN: A Cooperative Game Theory Framework with Transferable Utility for Stable GAN Training". It includes:- Standard GAN and Cooperative GAN (TU-GAN) implementations- WGAN and Cooperative WGAN (TU-WGAN) implementations- FID/IS evaluation scripts- Multi‑seed statistical analysis (mean, std, t‑test, Cohen's d)- All hyperparameters and results reported in the paper

Discussion

No takes yet. Share an insight, caveat, or question.

Cite This Study

Navidi et al. (2026) studied this question.

synapsesocial.com/papers/6a93f0ce6c1a8fb52e79d562 https://doi.org/10.5281/zenodo.22151208

View Full Paper

Ask AI

Bookmark

Share

TU-GAN: A Cooperative Game Theory Framework with Transferable Utility for Stable GAN Training

Key Points

Ask about this paper

Implication

Abstract

Discussion

Cite This Study

TU-GAN: A Cooperative Game Theory Framework with Transferable Utility for Stable GAN Training

Key Points

Ask about this paper

Implication

Abstract

Discussion

Cite This Study

Keep reading

Keep reading