Stabilizing Adversarial Nets With Prediction Methods
This addresses a critical problem for researchers and practitioners in machine learning by making adversarial networks more stable and easier to train, though it is incremental as it builds on existing optimization methods.
The paper tackles the instability and convergence issues in training adversarial neural networks by proposing a modified stochastic gradient descent method that reliably converges to saddle points, resulting in faster training with larger learning rates and reduced likelihood of collapse.
Adversarial neural networks solve many important problems in data science, but are notoriously difficult to train. These difficulties come from the fact that optimal weights for adversarial nets correspond to saddle points, and not minimizers, of the loss function. The alternating stochastic gradient methods typically used for such problems do not reliably converge to saddle points, and when convergence does happen it is often highly sensitive to learning rates. We propose a simple modification of stochastic gradient descent that stabilizes adversarial networks. We show, both in theory and practice, that the proposed method reliably converges to saddle points, and is stable with a wider range of training parameters than a non-prediction method. This makes adversarial networks less likely to "collapse," and enables faster training with larger learning rates.