A spin glass model for the loss surfaces of generative adversarial networks
We present a novel mathematical model that seeks to capture the key design feature of generative adversarial networks (GANs). Our model consists of two interacting spin glasses, and we conduct an extensive theoretical analysis of the complexity of the model’s critical points using techniques from Ra...
Main Authors: | Baskerville, N, Keating, J, Mezzadri, F, Najnudel, J |
---|---|
Format: | Journal article |
Language: | English |
Published: |
Springer
2022
|
Similar Items
-
The loss surfaces of neural networks with general activation functions
by: Baskerville, N, et al.
Published: (2021) -
Universal characteristics of deep neural network loss surfaces from random matrix theory
by: Baskerville, N, et al.
Published: (2022) -
Triple‐loss driven generative adversarial network for pansharpening
by: Bo Huang, et al.
Published: (2024-01-01) -
Generation of synthetic ground glass nodules using generative adversarial networks (GANs)
by: Zhixiang Wang, et al.
Published: (2022-11-01) -
Geophysical model generation with generative adversarial networks
by: Vladimir Puzyrev, et al.
Published: (2022-08-01)