The Latent Autoregressive Neural Operator (LANO) is a framework enabling neural operators to learn and predict from incomplete observational data. It addresses the supervision gap and dynamic spatial mismatch through a mask-to-predict strategy and a Physics-Aware Latent Propagator for autoregressive solution reconstruction.
The Latent Autoregressive Neural Operator (LANO) allows advanced AI models called neural operators to work with incomplete data, which is common in real-world science. It does this by creating artificial training signals and intelligently reconstructing missing information, making these powerful models applicable to more practical problems.
LANO, \ours
Was this definition helpful?