Structure-Aware Automatic Channel Pruning by Searching with Graph Embedding

2025-06-13

AI Summary

Existing channel pruning methods predominantly rely on local heuristics or weight-norm-based criteria, neglecting global structural dependencies among channels across layers—leading to suboptimal pruning decisions and substantial accuracy degradation. To address this, we propose the first end-to-end automated pruning framework grounded in graph embedding: we model the CNN topology as a weighted directed graph and employ a Graph Convolutional Network (GCN) to learn globally informed channel importance representations. Building upon this, we introduce a combinatorial-space constrained search mechanism that jointly optimizes layer-wise pruning ratios. Our approach is both structurally aware and fully automated. Extensive experiments on CIFAR-10 and ImageNet demonstrate that our method achieves significantly higher accuracy than state-of-the-art approaches for pruned ResNet and VGG-16 models, while simultaneously improving compression efficiency and generalization capability.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Channel pruning is a powerful technique to reduce the computational overhead of deep neural networks, enabling efficient deployment on resource-constrained devices. However, existing pruning methods often rely on local heuristics or weight-based criteria that fail to capture global structural dependencies within the network, leading to suboptimal pruning decisions and degraded model performance. To address these limitations, we propose a novel structure-aware automatic channel pruning (SACP) framework that utilizes graph convolutional networks (GCNs) to model the network topology and learn the global importance of each channel. By encoding structural relationships within the network, our approach implements topology-aware pruning and this pruning is fully automated, reducing the need for human intervention. We restrict the pruning rate combinations to a specific space, where the number of combinations can be dynamically adjusted, and use a search-based approach to determine the optimal pruning rate combinations. Extensive experiments on benchmark datasets (CIFAR-10, ImageNet) with various models (ResNet, VGG16) demonstrate that SACP outperforms state-of-the-art pruning methods on compression efficiency and competitive on accuracy retention.

Problem

• Automates channel pruning with global structure awareness
• Optimizes pruning decisions using graph convolutional networks
• Enhances compression efficiency and accuracy retention

Innovation

• Uses GCNs to model network topology
• Automates topology-aware channel pruning
• Dynamically adjusts pruning rate combinations