Search papers, labs, and topics across Lattice.
The paper addresses the problem of time series domain generalization when dealing with structurally heterogeneous dynamical systems, where global alignment can lead to negative transfer. They propose a Structurally Stratified Calibration Framework (SSCF) that first identifies structurally consistent sample clusters and then performs amplitude calibration only within these compatible clusters. This approach mitigates spurious correspondences and improves generalization performance in zero-shot settings.
Forget global alignment – calibrating time series *within* structurally similar clusters dramatically boosts cross-domain generalization.
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.
