What Uncertainties Do We Need for Dynamical Systems?
Provides a conceptual framework for uncertainty modeling in dynamical systems, which is less studied than in supervised learning.
The paper clarifies sources of uncertainty in dynamical systems, distinguishing aleatoric and epistemic types, and discusses how uncertainty representation and quantification objectives vary across tasks.
The distinction between aleatoric and epistemic uncertainty has received considerable attention in machine learning research, mainly in the context of supervised learning but also in other settings such as generative modeling. In this paper, we offer a machine learning perspective on uncertainty modeling for dynamical systems, which has been studied much less so far. In particular, we ask: what uncertainties do we need for dynamical systems? We discuss sources of uncertainty, clarify their nature (aleatoric or epistemic), and consider how the objectives of representing and quantifying uncertainty vary across different tasks.