Skip to content

Adaptive Stationary-Nonstationary Attention (ASNA)

Home / Concepts / Adaptive Stationary-Nonstationary Attention (ASNA)

Adaptive Stationary-Nonstationary Attention (ASNA)

Auto-generated stub. Edit this file to add more details.

An attention mechanism that adaptively fuses dependencies derived from normalized and raw time series sequences based on instance-level non-stationarity.

Why It Matters

It addresses the fundamental tension between modeling common global patterns and instance-specific local dynamics in non-stationary time series, which is the primary contribution of the paper.

Evidence

At its core is Adaptive Stationary-Nonstationary Attention (ASNA), which captures common dependencies from normalized sequences and specific dependencies from raw sequences, and adaptively fuses them according to instance-level non-stationarity.

Metadata & Links