TSFMAudit: 时间序列基础模型中的数据污染审计
TSFMAudit: Data Contamination Auditing in Forecasting Time Series Foundation Models
AI总结：
针对时间序列基础模型（TSFMs）预训练数据污染问题，提出基于探针适应动力学的审计方法TSFMAudit，通过检测微调后损失下降更快且骨干网络移动更小的异常现象来识别污染数据集。
摘要：
时间序列基础模型（TSFMs）越来越多地在大型语料库上进行预训练，这引发了评估数据集可能在预训练期间被暴露从而导致过于乐观的性能估计的担忧。在时间序列中审计此类污染具有挑战性，因为信号是连续且异质的，并且通常缺乏语料库文档。据我们所知，这是第一个研究TSFMs预训练污染审计的工作。我们形式化了TSFMs的预训练污染审计问题，并提出了TSFMAudit，一种基于探针适应动力学的方法。我们的关键直觉是，污染表现为异常高效的适应：在微调探针后，受污染的数据集往往表现出更快的损失减少和更小的骨干网络移动。我们在6个TSFMs和187个数据集上评估了TSFMAudit，使用文档化的训练来源证据作为监督，并与从LLM文献中改编的10个竞争基线进行了比较。
英文摘要：
Time series foundation models (TSFMs) are increasingly pretrained on large corpora, raising concerns that evaluation datasets may have been exposed during pretraining and thus yield overly optimistic performance estimates. Auditing such contamination is challenging in time series because signals are continuous and heterogeneous, and often lack corpus documentation. To the best of our knowledge, this is the first work to study pretraining contamination auditing for TSFMs. We formalize the problem of pretraining contamination auditing for TSFMs and propose TSFMAudit, a method based on probe adaptation dynamics. Our key intuition is that contamination manifests as unusually efficient adaptation: after a fine tuning probe, contaminated datasets tend to exhibit faster loss reduction with smaller backbone movement. We evaluate TSFMAudit on 6 TSFMs and 187 datasets using documented training source evidence as supervision, and compare against 10 competitive baselines adapted from the LLM literature.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
