PruneFuse: Efficient Data Selection via Weight Pruning and Network Fusion

2026-03-27

AI Summary

This work proposes PruneFuse, a novel approach to data selection that addresses the high computational cost and limited scalability of traditional methods. By introducing structured pruning into data selection for the first time, PruneFuse constructs a lightweight proxy network to efficiently identify informative samples. It then employs a neural fusion mechanism to transfer the knowledge acquired by the proxy network back to the original model, thereby optimizing the overall training process. Evaluated across multiple datasets, PruneFuse significantly reduces the computational overhead of data selection, outperforms existing baselines, and accelerates training while maintaining competitive model performance—effectively balancing efficiency and accuracy.

Abstract

Efficient data selection is crucial for enhancing the training efficiency of deep neural networks and minimizing annotation requirements. Traditional methods often face high computational costs, limiting their scalability and practical use. We introduce PruneFuse, a novel strategy that leverages pruned networks for data selection and later fuses them with the original network to optimize training. PruneFuse operates in two stages: First, it applies structured pruning to create a smaller pruned network that, due to its structural coherence with the original network, is well-suited for the data selection task. This small network is then trained and selects the most informative samples from the dataset. Second, the trained pruned network is seamlessly fused with the original network. This integration leverages the insights gained during the training of the pruned network to facilitate the learning process of the fused network while leaving room for the network to discover more robust solutions. Extensive experimentation on various datasets demonstrates that PruneFuse significantly reduces computational costs for data selection, achieves better performance than baselines, and accelerates the overall training process.

Problem

• data selection
• training efficiency
• computational cost
• deep neural networks
• annotation requirements

Innovation

• weight pruning
• network fusion
• data selection
• structured pruning
• efficient training