WinoTS: Wavelet-based Self-Distillation for Time Series Models

Excerpt

arXiv:2609.39337v1 Announce Type: new Abstract: Self-supervised pre-training of time series models is currently dominated by next-token prediction and reconstruction objectives. In continuous-valued domains, these paradigms often waste model capacity on high-frequency, point-wise noise at the expense of learning invariant structure. While invariance-based self-distillation has proven highly effective in computer vision, its application to temporal data remains largely underexplored. Effectively