REVIEW 4 cited by
Koopman Neural Forecaster for Time Series with Temporal Distribution Shifts
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Temporal distributional shifts, with underlying dynamics changing over time, frequently occur in real-world time series and pose a fundamental challenge for deep neural networks (DNNs). In this paper, we propose a novel deep sequence model based on the Koopman theory for time series forecasting: Koopman Neural Forecaster (KNF) which leverages DNNs to learn the linear Koopman space and the coefficients of chosen measurement functions. KNF imposes appropriate inductive biases for improved robustness against distributional shifts, employing both a global operator to learn shared characteristics and a local operator to capture changing dynamics, as well as a specially-designed feedback loop to continuously update the learned operators over time for rapidly varying behaviors. We demonstrate that \ours{} achieves superior performance compared to the alternatives, on multiple time series datasets that are shown to suffer from distribution shifts.
Forward citations
Cited by 4 Pith papers
-
MetaKoopman: Bayesian Meta-Learning of Koopman Operators for Modeling Structured Dynamics under Distribution Shifts
MetaKoopman meta-learns a Bayesian prior over Koopman operators and updates it online with recent data, improving multi-step forecasting and uncertainty estimates under distribution shift.
-
Deep Koopman operator framework for causal discovery in nonlinear dynamical systems
A deep Koopman framework called Kausal discovers causal direction and magnitude in nonlinear dynamical systems by comparing joint versus marginal prediction errors in learned observable spaces.
-
A Minimal Interpretable Architecture for Zero-Shot Reconstruction of Dynamical Systems
A two-parameter nearest-neighbor recurrence, DynaBase, matches large foundation models at zero-shot dynamical-system reconstruction and unifies context parroting with chaotic dynamics as two ends of one parameter.
-
Wavelet-based Disentangled Adaptive Normalization for Non-stationary Times Series Forecasting
WDAN uses wavelet decomposition to split series into trend and residual, normalizes them separately, and predicts future statistics to improve non-stationary time series forecasting.
Discussion (0). Continue with ORCID to comment.