Generative models, particularly Generative Adversarial Networks (GANs), often suffer from a lack of output diversity, frequently generating similar samples rather than a wide range of variations. This paper introduces a novel generalization of the GAN loss function based on Dempster-Shafer theory of evidence, applied to both the generator and discriminator. Additionally, we propose an architectural enhancement to the generator that enables it to predict a mass function for each image pixel. This modification allows the model to quantify uncertainty in its outputs and leverage this uncertainty to produce more diverse and representative generations. Experimental evidence shows that our approach not only improves generation variability but also provides a principled framework for modeling and interpreting uncertainty in generative processes.
Figures & tables
Figure 1Figure 2
Model
Time (sec)
Standard GAN
75.91 ± 1.79
Epistemic GAN
77.10 ± 1.86
Table 1: Training time per epoch on CelebA (averaged over 5 runs).
CelebA
Cifar-10
Food101
FID ↓
Vendi Score ↑
FID ↓
Vendi Score ↑
FID ↓
Vendi Score ↑
Reference
-
6.95
-
8.48
-
16.29
Standard GAN
18.5
5.70
25.9
4.25
33.76
12.78
Epistemic GAN
17.3
5.86
24.1
4.53
29.1
13.82
Table 2: Performance of Standard GAN and Epistemic GAN on Celeb-A dataset. Reference represents the Vendi score (diversity) of the training data.
Figure 5: Generations for Standard GAN (left) and Epistemic GAN(right) for Celeb-A dataset.
Appendix figures & tables4 assets
Supplementary material from the paper’s appendix.
Appendix
γ \ β
0
0.5
1
2
0
17.4 (5.35)
17.1 (5.11)
20.1 (6.20)
23.7 (6.30)
0.5
18.9 (5.43)
17.5 (5.71)
18.4 (5.98)
20.2 (5.87)
1
19.1 (5.05)
19.71 (5.70)
17.3 (5.86)
22.6 (6.01)
2
21.6 (5.01)
23.1 (5.33)
19.6 (5.77)
21.7 (5.92)
Appendix
Table 3: Ablation study for interval precision weight ( γ ) and variance promotion weight ( β ) on CelebA. Results are reported as [FID ↓ (Vendi Score ↑ )]. The ”sweet spot” at ≈1 confirms the theoretical necessity of balancing epistemic spread with prediction confidence.
λ
0
0.5
1
2
Score
19.4 (5.88)
17.4 (5.70)
17.3 (5.90)
20.1 (6.01)
Appendix
Table 4: Ablation study for the belief regularization weight λ on the CelebA dataset. Reported as FID (Vendi Score) . λ=1 provides sufficient constraint without hampering training dynamics.
Configuration
FID (Vendi Score)
Evidential Discriminator Only
15.1 (5.15)
Evidential Generator Only
19.2 (5.97)
Full Epistemic GAN
17.3 (5.86)
Appendix
Table 5: Ablation for architectural components on CelebA. The full model strikes the necessary balance between quality and diversity.
Figure 6: Epistemic GAN generations for road scenarios on Citiscape dataset.
Department of Computer Science and Engineering, The Chinese University of Hong Kong · Department of Information Engineering, The Chinese University of Hong Kong