Equivariant-Aware Structured Pruning for Efficient Edge Deployment: A Comprehensive Framework with Adaptive Fine-Tuning

2025-11-21

AI Summary

To address the tension between resource constraints on edge devices and the need for geometric robustness, this paper proposes an efficient deployment framework for group-equivariant neural networks. Methodologically, it tightly integrates C4-group equivariant convolution with equivariance-aware structured pruning, introduces a novel neuron-level pruning strategy, and incorporates adaptive fine-tuning to preserve transformation equivariance. Additionally, knowledge distillation, dynamic INT8 quantization, and learning-rate scheduling are employed to optimize deployment efficiency. Experiments on EuroSAT, CIFAR-10, and Rotated MNIST demonstrate that the framework achieves 29.3% parameter compression while substantially recovering accuracy. The resulting models are lightweight (<1M parameters), retain rotational equivariance, and exhibit enhanced geometric robustness. This work establishes a verifiable, equivariance-preserving compression paradigm tailored for edge vision tasks.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Computer Vision: Adversarial Attacks & Robustness
• Intelligent Robots: Learning & Optimization for ROB

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

This paper presents a novel framework combining group equivariant convolutional neural networks (G-CNNs) with equivariant-aware structured pruning to produce compact, transformation-invariant models for resource-constrained environments. Equivariance to rotations is achieved through the C4 cyclic group via the e2cnn library,enabling consistent performance under geometric transformations while reducing computational overhead. Our approach introduces structured pruning that preserves equivariant properties by analyzing e2cnn layer structure and applying neuron-level pruning to fully connected components. To mitigate accuracy degradation, we implement adaptive fine-tuning that automatically triggers when accuracy drop exceeds 2%, using early stopping and learning rate scheduling for efficient recovery. The framework includes dynamic INT8 quantization and a comprehensive pipeline encompassing training, knowledge distillation, structured pruning, fine-tuning, and quantization. We evaluate our method on satellite imagery (EuroSAT) and standard benchmarks (CIFAR-10, Rotated MNIST) demonstrating effectiveness across diverse domains. Experimental results show 29.3% parameter reduction with significant accuracy recovery, demonstrating that structured pruning of equivariant networks achieves substantial compression while maintaining geometric robustness. Our pipeline provides a reproducible framework for optimizing equivariant models, bridging the gap between group-theoretic network design and practical deployment constraints, with particular relevance to satellite imagery analysis and geometric vision tasks.

Problem

• Develops compact transformation-invariant models for resource-constrained edge deployment
• Preserves geometric equivariance during structured pruning to maintain transformation robustness
• Mitigates accuracy degradation through adaptive fine-tuning and quantization techniques

Innovation

• Equivariant-aware structured pruning preserves transformation properties
• Adaptive fine-tuning triggers automatically for accuracy recovery
• Dynamic INT8 quantization and comprehensive pipeline enhance efficiency