Latent State Inference in a Spatiotemporal Generative Model
This work addresses the challenge of unsupervised latent state inference in spatiotemporal systems like weather prediction, offering a method that improves prediction accuracy and derives causal factors, though it is incremental as it builds on existing DISTANA architecture.
The paper tackled the problem of inferring hidden causal factors from spatiotemporal time series data without supervision, using an enhanced DISTANA model with active tuning to achieve more accurate predictions and reliably derive location-specific hidden factors, such as inferring a land-sea mask from temperature dynamics to improve weather forecasts.
Knowledge about the hidden factors that determine particular system dynamics is crucial for both explaining them and pursuing goal-directed interventions. Inferring these factors from time series data without supervision remains an open challenge. Here, we focus on spatiotemporal processes, including wave propagation and weather dynamics, for which we assume that universal causes (e.g. physics) apply throughout space and time. A recently introduced DIstributed SpatioTemporal graph Artificial Neural network Architecture (DISTANA) is used and enhanced to learn such processes, requiring fewer parameters and achieving significantly more accurate predictions compared to temporal convolutional neural networks and other related approaches. We show that DISTANA, when combined with a retrospective latent state inference principle called active tuning, can reliably derive location-respective hidden causal factors. In a current weather prediction benchmark, DISTANA infers our planet's land-sea mask solely by observing temperature dynamics and, meanwhile, uses the self inferred information to improve its own future temperature predictions.