一作ACL 2026Poster· Natural Language Processing
Is the Attention Matrix Really the Key to Self-Attention in Multivariate Long-Term Time Series Forecasting?
Xinyu Li · Kexi Chen · Jiajie Shen · Ying Zheng · Hong Lu · Jin Zhao · Xin Wang
摘要
In multivariate long-term time series forecasting, it is widely believed that the effectiveness of self-attention arises from its attention matrix. We challenge this assumption with a counterintuitive finding: our experiments, conducted on three classic and three latest Transformer models, show t...