实例自适应提示作为时间序列基础模型的上下文

中文摘要 AI 辅助

更长的历史记录可以提升时间序列基础模型（TSFMs）的性能，但会显著增加推理成本。因此，我们探讨是否可以通过一组紧凑的学习嵌入令牌更高效地提供上下文信息。我们引入了PaCTS，它根据可见上下文生成一小部分实例自适应潜在提示，形式为连续嵌入令牌。这些提示作为冻结TSFMs的紧凑上下文替代品。PaCTS从实例特定的全局统计中构建这些提示，并通过段级时间信息进一步细化，捕捉全局特征和局部时间变化。提示模块在异构时间序列上与冻结骨干网络联合训练和部署。大量实验证明了提示作为上下文的有效性，在不同上下文长度和模型架构上持续提升预测性能。在较短的输入上下文下，PaCTS可以超越使用双倍上下文的同一冻结骨干网络，同时需要显著更少的推理计算。与权重空间自适应方法相比，PaCTS实现了更强的改进和更好的分布外泛化。

英文摘要

Longer histories can improve time-series foundation models (TSFMs), but require substantially higher inference cost. We therefore ask whether contextual information can be provided more efficiently through a compact set of learned token embeddings. We introduce PaCTS, which generates a small set of instance-adaptive latent prompts in the form of continuous embedding tokens conditioned on the visible context. These prompts serve as compact context surrogates for frozen TSFMs. PaCTS constructs them from instance-specific global statistics and further refines them with segment-level temporal information, capturing both global characteristics and local temporal variations. The prompt module is jointly trained and deployed across heterogeneous time series with the frozen backbone. Extensive experiments demonstrate the effectiveness of prompts as context, consistently improving forecasting across context lengths and model architectures. With a shorter input context, PaCTS can outperform the same frozen backbone using double context while requiring substantially less inference computation. Compared with weight-space adaptation methods, PaCTS achieves stronger improvements and better out-of-distribution generalization.

发表机构

• LIPADE, Université Paris Cité（巴黎西岱大学LIPADE实验室）

机构由 AI 辅助整理，请以论文原文为准。