Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Rethinking Time Series Domain Generalization via Structure-Stratified Calibration

Jinyang Li, Shuhao Mei, Xiaoyu Xiao, Shuhang Li, Ruoxi Yun, Jinbo Sun

TL;DR

This work revisits the problem of cross-domain structural correspondence failure and proposes a structurally stratified calibration framework that explicitly distinguishes structurally consistent samples and performs amplitude calibration exclusively within structurally compatible sample clusters, thereby effectively alleviating generalization failures caused by structural incompatibility.

Abstract

For time series arising from latent dynamical systems, existing cross-domain generalization methods commonly assume that samples are comparably meaningful within a shared representation space. In real-world settings, however, different datasets often originate from structurally heterogeneous families of dynamical systems, leading to fundamentally distinct feature distributions. Under such circumstances, performing global alignment while neglecting structural differences is highly prone to establishing spurious correspondences and inducing negative transfer. From the new perspective of cross-domain structural correspondence failure, we revisit this problem and propose a structurally stratified calibration framework. This approach explicitly distinguishes structurally consistent samples and performs amplitude calibration exclusively within structurally compatible sample clusters, thereby effectively alleviating generalization failures caused by structural incompatibility. Notably, the proposed framework achieves substantial performance improvements through a concise and computationally efficient calibration strategy. Evaluations on 19 public datasets (100.3k samples) demonstrate that SSCF significantly outperforms strong baselines under the zero-shot setting. These results confirm that establishing structural consistency prior to alignment constitutes a more reliable and effective pathway for improving cross-domain generalization of time series governed by latent dynamical systems.

Rethinking Time Series Domain Generalization via Structure-Stratified Calibration

TL;DR

This work revisits the problem of cross-domain structural correspondence failure and proposes a structurally stratified calibration framework that explicitly distinguishes structurally consistent samples and performs amplitude calibration exclusively within structurally compatible sample clusters, thereby effectively alleviating generalization failures caused by structural incompatibility.

Abstract

For time series arising from latent dynamical systems, existing cross-domain generalization methods commonly assume that samples are comparably meaningful within a shared representation space. In real-world settings, however, different datasets often originate from structurally heterogeneous families of dynamical systems, leading to fundamentally distinct feature distributions. Under such circumstances, performing global alignment while neglecting structural differences is highly prone to establishing spurious correspondences and inducing negative transfer. From the new perspective of cross-domain structural correspondence failure, we revisit this problem and propose a structurally stratified calibration framework. This approach explicitly distinguishes structurally consistent samples and performs amplitude calibration exclusively within structurally compatible sample clusters, thereby effectively alleviating generalization failures caused by structural incompatibility. Notably, the proposed framework achieves substantial performance improvements through a concise and computationally efficient calibration strategy. Evaluations on 19 public datasets (100.3k samples) demonstrate that SSCF significantly outperforms strong baselines under the zero-shot setting. These results confirm that establishing structural consistency prior to alignment constitutes a more reliable and effective pathway for improving cross-domain generalization of time series governed by latent dynamical systems.
Paper Structure (52 sections, 12 equations, 4 figures, 18 tables, 2 algorithms)

This paper contains 52 sections, 12 equations, 4 figures, 18 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methods Assuming Global Cross-domain Comparability
  4. Mechanism-aware and Disentangled Representation Learning
  5. Spectral Representations and Frequency-domain Transfer
  6. Preliminaries
  7. Notation.
  8. Cross-domain Setting.
  9. Problem Statement.
  10. Method
  11. Spectral Modeling Assumptions
  12. Definition 1 (Time Series from Latent Dynamical Systems).
  13. Definition 2 (Feature-level Spectral Observation Model).
  14. Structure Stratification and Structural Representation
  15. Reference Anchor Construction
  16. ...and 37 more sections

Figures (4)

  • Figure 1: Overview of structure-consistent domain alignment for time-series domain generalization. (Other) Global alignment ignores structural heterogeneity, causing mismatches between structurally incompatible samples (dashed lines). (Our) We stratify samples by spectral structure and align only within consistent subsets, improving generalization stability.
  • Figure 2: Overview of the proposed SSCF. Time-series data from the source and target domains are first encoded into shallow feature representations and subsequently stratified in the power spectral feature space to capture coarse-grained spectral patterns. Within each structural stratum, a corresponding reference anchor is constructed. For an arbitrary input sample, SSCF first performs structural matching to identify the most compatible stratum and then conducts amplitude calibration exclusively within that stratum, while explicitly preserving phase information.
  • Figure 3: Sensitivity analysis of SSCF. Top: effect of structural granularity $K$. Bottom: impact of anchor matching rank $R$.
  • Figure 4: Spectral visualization of structural stratification in the sleep staging task under $K=8$. Each subfigure corresponds to one structural stratum and shows the normalized power spectral density (PSD) curves of the samples assigned to that stratum, together with their variability ranges. Different strata exhibit clearly distinct spectral profiles in terms of low-frequency energy concentration, dominant frequency bands, and overall spectral decay behavior, while samples within the same stratum demonstrate high spectral consistency.