The zero-shot evaluation of time series foundation models (TSFMs) for classification typically uses a frozen encoder followed by a task-specific classifier. However, this practice violates the training-free premise of zero-shot deployment and introduces evaluation bias due to classifier-dependent training choices. To address this issue, we propose TIC-FM, an in-context learning framework that treats the labeled training set as context and predicts labels for all test instances in a single forward pass, without parameter updates. TIC-FM pairs a time series encoder and a lightweight projection adapter with a split-masked latent memory Transformer. We further provide theoretical justification that in-context inference can subsume trained classifiers and can emulate gradient-based classifier training within a single forward pass. Experiments on 128 UCR datasets show strong accuracy, with consistent gains in the extreme low-label situation, highlighting training-free transfer
Cite
@article{arxiv.2602.00620,
title = {Rethinking Zero-Shot Time Series Classification: From Task-specific Classifiers to In-Context Inference},
author = {Juntao Fang and Shifeng Xie and Shengbin Nie and Yuhui Ling and Yuming Liu and Zijian Li and Keli Zhang and Lujia Pan and Themis Palpanas and Ruichu Cai},
journal= {arXiv preprint arXiv:2602.00620},
year = {2026}
}