Self-Gating Attention for Efficient Time Series Forecasting
Transformer architectures have shown strong potential in time series forecasting, where multi-head self-attention is widely used to capture temporal dependencies across historical timestamps. However, standard self-attention has quadratic time and memory complexity with respect to the look-back length. This cost may limit its use in resource-constrained or high-throughput forecasting systems, where fast and memory-efficient inference is important. Through qualitative and quantitative analyses, we observe that self-attention maps in time series forecasting often contain redundant patterns acros
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 54%amazon-science/chronos-forecasting →
- PossiblePossibly related (embedding) · 47%Nixtla/statsforecast →
- LinkedLinked via arxiv author · 85%Dezheng Wang →
“Self-Gating Attention for Efficient Time Series Forecasting”
- LinkedLinked via arxiv author · 85%Tong Chen →
“Self-Gating Attention for Efficient Time Series Forecasting”
- LinkedLinked via arxiv author · 85%Wei Yuan →
“Self-Gating Attention for Efficient Time Series Forecasting”
- LinkedLinked via arxiv author · 85%Congyan Chen →
“Self-Gating Attention for Efficient Time Series Forecasting”
- LinkedLinked via arxiv author · 85%Shihua Li →
“Self-Gating Attention for Efficient Time Series Forecasting”
- LinkedLinked via arxiv author · 85%Hongzhi Yin →
“Self-Gating Attention for Efficient Time Series Forecasting”