Feature-aligned N-BEATS with Sinkhorn divergence

arXiv:2305.15196v3 Announce Type: replace
Abstract: We propose Feature-aligned N-BEATS as a domain-generalized time series forecasting model. It is a nontrivial extension of N-BEATS with doubly residual stacking principle (Oreshkin et al. [45]) into a representation learning framework. In particular, it revolves around marginal feature probability measures induced by the intricate composition of residual and feature extracting operators of N-BEATS in each stack and aligns them stack-wise via an approximate of an optimal transport distance referred to as the Sinkhorn divergence. …

abstract arxiv cs.ai cs.lg divergence domain extension feature forecasting framework generalized math.oc math.pr probability representation representation learning residual series time series time series forecasting type