Diffusion-Augmented Neural Processes
This work addresses issues in NPs for applications like healthcare and climate sciences, offering an incremental improvement over existing methods.
The paper tackles limitations in Neural Processes (NPs) for uncertainty estimation in data-scarce domains by proposing a diffusion-based approach that conditions on noised datasets, achieving state-of-the-art performance.
Over the last few years, Neural Processes have become a useful modelling tool in many application areas, such as healthcare and climate sciences, in which data are scarce and prediction uncertainty estimates are indispensable. However, the current state of the art in the field (AR CNPs; Bruinsma et al., 2023) presents a few issues that prevent its widespread deployment. This work proposes an alternative, diffusion-based approach to NPs which, through conditioning on noised datasets, addresses many of these limitations, whilst also exceeding SOTA performance.