arXiv cs.AI· Sachith Abeywickrama, Peiliang Gong, Emadeldeen Eldele, Min Wu, Chau Yuen·· 10 小时前AI 评分26
PACT:时间序列基础模型的循环域增量适配方法
Recurrent Domain Incremental Adaptation of Time-Series Foundation Models
AI 导读
PACT(Piecewise Adaptation via Chart Transport)通过将冻结的源嵌入空间划分为局部图表、仅为受新工况影响的区域学习轻量残差传输,实现时间序列基础模型的域增量适配。该方法每个新工况仅需 5% 标注校准数据,推理时无需知道工况身份,在四工况轴承故障流上平均 macro-F1 保持 0.95 以上。预训练编码器、分类头与已学传输均保持固定,代码已开源。
正文 · 原文
Abstract:Pretrained Time-Series Foundation Models (TSFMs) provide strong representations for downstream classification tasks. Yet adapting a TSFM under changing operating conditions remains challenging. A model must improve on each newly observed condition without sacrificing conditions it has already learned, often from only a small labeled calibration set and without knowing the operating condition at inference time. We find that this difficulty has an important geometric structure. Under an operating-condition shift, the embedding space of a TSFM does not move uniformly: some local regions change substantially, while others remain nearly unchanged. Consequently, a global adaptation can unnecessarily move embeddings that were already correctly classified. To address this, we propose the novel PACT (Piecewise Adaptation via Chart Transport). PACT partitions the frozen source embedding space into local charts and learns lightweight residual transports for the regions affected by each new condition. An explicit identity path is always retained, and a learned gate softly selects among the identity and accumulated transports for each input. The pretrained TSFM encoder, classifier head, and previously learned transports remain fixed. With only 5% labeled calibration data per arriving condition and no condition identity at test time, PACT maintains an average macro-F1 above 0.95 across a four-condition bearing-fault stream. Under homogeneous tasks and label spaces, similar retention-adaptation behavior is observed across public benchmarks. Code is available at this https URL.
| Comments: | Preprint. Under Review |
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Signal Processing (eess.SP) |
| Cite as: | arXiv:2611.24075 [cs.AI] |
| (or arXiv:2611.24075v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.24075 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Sachith Abeywickrama [view email]
[v1]
Wed, 18 Nov 2026 09:05:22 UTC (6,142 KB)
来源:arXiv cs.AI · arxiv.org