arXiv：cs.LG· Van Dai Do, Huu Hiep Nguyen, Minh Hoang Nguyen, Hung Le·· 3 小时前
SteerCast：基于检索的潜空间引导提升 decoder-only 时间序列预测
SteerCast: Retrieval-Based Latent Steering for Decoder-Only Time Series Forecasting
AI 导读
SteerCast 是一种基于检索的潜空间引导方法，能在推理阶段提升 decoder-only 时间序列预测模型，且无需更新其参数。该方法在训练集上构建数据库，存储每个历史窗口的表示及其在模型潜空间中的引导向量，测试时检索近邻并聚合这些向量，注入自回归生成的每一步隐藏状态。在多个多变量基准和预测时域上，SteerCast 的预测精度稳定优于微调后的骨干模型和基于检索的基线。
正文
Abstract:Time series forecasting aims to predict future values from historical observations and auxiliary features. We propose \textbf{SteerCast}, a retrieval-based latent steering method that improves decoder-only forecaster at inference time, without updating its parameters. SteerCast constructs a database from the training set by storing a representation of each history window together with a \emph{steering vector} computed in the forecaster's latent space, defined as the difference between representations induced by the ground-truth continuation and by the model's own prediction. At test time, SteerCast retrieves nearest neighbors for a query history, aggregates their steering vectors, and injects the resulting signal into the forecaster's hidden states at every step of autoregressive generation, guiding predictions toward trajectories consistent with similar training cases. Experiments across diverse multivariate benchmarks and multiple horizons show that SteerCast consistently improves forecasting accuracy over the fine-tuned backbone and retrieval-based baselines, while requiring no additional training beyond the original fine-tuning and using only the training set as a retrieval corpus.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.11229 [cs.LG] |
| (or arXiv:2610.11229v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.11229 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Van Dai Do [view email]
[v1]
Thu, 8 Oct 2026 04:30:54 UTC (1,276 KB)
来源：arXiv：cs.LG · arxiv.org