Subgrid BoostCNN: Efficient Boosting of Convolutional Networks via Gradient-Guided Feature Selection

2025-07-30

AI Summary

To address the challenges of large parameter counts, high training costs, and heavy reliance on manual architecture design in CNNs, this paper proposes a novel CNN training framework integrating dynamic feature selection with boosting-style learning. The method introduces a sub-grid selection mechanism coupled with gradient-guided importance sampling to adaptively focus on salient feature regions; it further embeds weight updates directly into a least-squares loss optimization process, enabling feature-importance-driven adaptive parameter adjustment. Crucially, the approach requires no architectural modifications and significantly reduces dependence on manual hyperparameter tuning and deep layer stacking. Evaluated on multiple fine-grained image classification benchmarks, the method achieves an average accuracy improvement of 3.2–5.7% over standard CNNs while reducing training iterations by approximately 40%. It thus jointly enhances predictive accuracy, computational efficiency, and model interpretability.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Dimensionality Reduction/Feature Selection
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Convolutional Neural Networks (CNNs) have achieved remarkable success across a wide range of machine learning tasks by leveraging hierarchical feature learning through deep architectures. However, the large number of layers and millions of parameters often make CNNs computationally expensive to train, requiring extensive time and manual tuning to discover optimal architectures. In this paper, we introduce a novel framework for boosting CNN performance that integrates dynamic feature selection with the principles of BoostCNN. Our approach incorporates two key strategies: subgrid selection and importance sampling, to guide training toward informative regions of the feature space. We further develop a family of algorithms that embed boosting weights directly into the network training process using a least squares loss formulation. This integration not only alleviates the burden of manual architecture design but also enhances accuracy and efficiency. Experimental results across several fine-grained classification benchmarks demonstrate that our boosted CNN variants consistently outperform conventional CNNs in both predictive performance and training speed.

Problem

• Reducing computational cost of training CNNs
• Automating optimal CNN architecture discovery
• Improving CNN accuracy and training efficiency

Innovation

• Dynamic feature selection with BoostCNN principles
• Subgrid selection and importance sampling strategies
• Boosting weights integrated via least squares loss