FCOS: A Two-Stage Recoverable Model Pruning Framework for Automatic Modulation Recognition

TL;DR

This paper introduces FCOS, a two-stage pruning framework that combines channel-level pruning and layer collapse diagnosis to significantly compress deep learning models for automatic modulation recognition while maintaining high accuracy.

Contribution

The paper proposes a novel hierarchical pruning framework, FCOS, that effectively reduces model size and computational complexity for AMR tasks through combined channel and layer pruning techniques.

Findings

Achieves 95.51% FLOPs reduction and 95.31% parameter reduction.

Maintains near-original accuracy with only 0.46% accuracy drop.

Outperforms existing pruning methods on multiple AMR benchmarks.

Abstract

With the rapid development of wireless communications and the growing complexity of digital modulation schemes, traditional manual modulation recognition methods struggle to extract reliable signal features and meet real-time requirements in modern scenarios. Recently, deep learning based Automatic Modulation Recognition (AMR) approaches have greatly improved classification accuracy. However, their large model sizes and high computational demands hinder deployment on resource-constrained devices. Model pruning provides a general approach to reduce model complexity, but existing weight, channel, and layer pruning techniques each present a trade-off between compression rate, hardware acceleration, and accuracy preservation. To this end, in this paper, we introduce FCOS, a novel Fine-to-COarse two-Stage pruning framework that combines channel-level pruning with layer-level collapse diagnosis to achieve extreme compression, high performance and efficient inference. In the first stage of FCOS, hierarchical clustering and parameter fusion are applied to channel weights to achieve channel-level pruning. Then a Layer Collapse Diagnosis (LaCD) module uses linear probing to identify layer collapse and removes the collapsed layers due to high channel compression ratio. Experiments on multiple AMR benchmarks demonstrate that FCOS outperforms existing channel and layer pruning methods. Specifically, FCOS achieves 95.51% FLOPs reduction and 95.31% parameter reduction while still maintaining performance close to the original ResNet56, with only a 0.46% drop in accuracy on Sig2019-12. Code is available at https://github.com/yaolu-zjut/FCOS.