Skip to content

Learnable Feature Identity

Home / Concepts / Learnable Feature Identity

Learnable Feature Identity

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

A persistent embedding strategy that encodes intrinsic semantic properties of time series features to provide stable anchors for cross-feature dependency modeling.

Why It Matters

Provides a stable anchor for cross-feature dependencies, addressing the instability of re-discovering relationships at each layer in standard attention mechanisms.

Evidence

assigns each feature a learnable feature identity, a persistent embedding that captures intrinsic semantic properties throughout the network

Metadata & Links