Related papers
Non-autoregressive neural networks outperform autoregressive for faster system identification.
Neural networks can be used to understand how systems work. Researchers compared two types of neural networks for this task: autoregressive and non-autoregressive. They found that non-autoregressive networks are faster and just as accurate as autoregressive ones. In fact, the non-autoregressive network they used was the best at figuring out how systems work compared to other methods.