From Generalization Analysis to Optimization Designs for State Space Models

arXiv:2405.02670v1 Announce Type: new
Abstract: A State Space Model (SSM) is a foundation model in time series analysis, which has recently been shown as an alternative to transformers in sequence modeling. In this paper, we theoretically study the generalization of SSMs and propose improvements to training algorithms based on the generalization results. Specifically, we give a \textit{data-dependent} generalization bound for SSMs, showing an interplay between the SSM parameters and the temporal dependencies of the training sequences. Leveraging the generalization bound, …

abstract algorithms alternative analysis arxiv cs.lg designs foundation foundation model improvements modeling optimization paper series space ssm ssms state state space model state space models study time series training transformers type