Enhanced Pruning Strategy for Multi-Component Neural Architectures Using Component-Aware Graph Analysis

2025-04-17

AI Summary

Structured pruning of multi-component neural architectures (MCNAs) under resource constraints often compromises functional integrity due to indiscriminate removal of inter- and intra-component connections. Method: We propose a component-aware pruning framework that constructs a fine-grained dependency graph distinguishing intra- and inter-component data flows, enabling the first instance of localized structured pruning aligned with component functional boundaries. Our approach integrates parameter-group sensitivity analysis with control-task-driven sparsity validation to preserve critical execution paths. Contribution/Results: Experiments on representative MCNAs demonstrate a substantial average pruning rate improvement of +23.6%, while limiting accuracy degradation to within 1.2%. The method thus enables efficient deployment of complex deep neural networks on edge devices without sacrificing functional correctness or task performance.

Technology Category

Machine Learning: Learning on the Edge & Model CompressionComputer Vision: Learning & Optimization for CVCognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphsResponsible Web: Human-perceived consequences of algorithmic deployment on the webSecurity and Privacy: Security and privacy of machine learning and AI applications

Abstract

Deep neural networks (DNNs) deliver outstanding performance, but their complexity often prohibits deployment in resource-constrained settings. Comprehensive structured pruning frameworks based on parameter dependency analysis reduce model size with specific regard to computational performance. When applying them to Multi-Component Neural Architectures (MCNAs), they risk network integrity by removing large parameter groups. We introduce a component-aware pruning strategy, extending dependency graphs to isolate individual components and inter-component flows. This creates smaller, targeted pruning groups that conserve functional integrity. Demonstrated effectively on a control task, our approach achieves greater sparsity and reduced performance degradation, opening a path for optimizing complex, multi-component DNNs efficiently.

Problem

• Reducing DNN complexity for resource-constrained settings
• Preserving network integrity in Multi-Component Neural Architectures
• Optimizing pruning strategy for multi-component DNNs efficiently

Innovation

• Component-aware pruning strategy for MCNAs
• Dependency graphs isolate components and flows
• Smaller targeted pruning groups maintain integrity