LGDec 10, 2024
Adversarial Autoencoders in Operator Learning
arXiv:2412.07811v1
Originality Synthesis-oriented
AI Analysis
This work addresses improving neural operators for scientific computing, but appears incremental as it adapts existing adversarial techniques to known architectures.
The paper investigates applying adversarial training to two neural operator architectures, DeepONets and Koopman autoencoders, to potentially enhance their performance, though no specific results or numbers are provided.
DeepONets and Koopman autoencoders are two prevalent neural operator architectures. These architectures are autoencoders. An adversarial addition to an autoencoder have improved performance of autoencoders in various areas of machine learning. In this paper, the use an adversarial addition for these two neural operator architectures is studied.