Multi-task learning enhanced frequency estimation method via the gated split-complex neural network

Xiaoqi Hou

doi:10.1016/j.measurement.2026.120902

Multi-task learning enhanced frequency estimation method via the gated split-complex neural network

2026-02-20

0

PRE

AI

DOI：10.1016/j.measurement.2026.120902

Original

Original request for help

Save

Abstract

En 中文

• Propose cGMTNet for deep learning-based frequency estimation with better explainability. • Multi-task learning improves estimation accuracy and network robustness to interference. • Skip connections fuse multi-scale features, while GLU dynamically preserves key information. • The cGMTNet achieves high accuracy and efficiency in estimation on synthetic and real data.

Keywords:

frequency estimation

multi-task learning

gated split-complex neural network

explainability

robustness

Multi-task learning enhanced frequency estimation method via the gated split-complex neural network

Abstract

Journal

Organization