Probabilistic structure discovery in time series data
This work addresses the issue of unreliable uncertainty estimates in time series structure discovery for researchers and practitioners using Gaussian process models, representing an incremental improvement over prior methods.
The paper tackled the problem of structure discovery in time series data, where existing methods use greedy optimization that restricts solution space and leads to over-confident uncertainty estimates; they introduced a fully Bayesian approach that infers a full posterior over structures, resulting in more reliable uncertainty capture.
Existing methods for structure discovery in time series data construct interpretable, compositional kernels for Gaussian process regression models. While the learned Gaussian process model provides posterior mean and variance estimates, typically the structure is learned via a greedy optimization procedure. This restricts the space of possible solutions and leads to over-confident uncertainty estimates. We introduce a fully Bayesian approach, inferring a full posterior over structures, which more reliably captures the uncertainty of the model.