SPvR: Structured Pruning via Ranking

Atif Hassan, Jiaul H. Paik, Swanand Khare
Proceedings of the Forty-first Conference on Uncertainty in Artificial Intelligence, PMLR 286:1663-1676, 2025.

Abstract

Deep neural networks have achieved state-of-the-art performance in multiple domains but are increasingly resource-intensive, limiting their deployment on constrained devices. We introduce Structured Pruning via Ranking (SPvR), a novel structured pruning approach to address this challenge for classification tasks. SPvR prunes pre-trained networks in terms of function composition and network width while adhering to a user-specified parameter budget. Our method leverages local grouping and global ranking modules to generate smaller yet effective networks tailored to a given dataset and model. Finally, we train the pruned networks from scratch, instead of fine-tuning. Our evaluations demonstrate that SPvR significantly surpasses existing state-of-the-art pruning methods on benchmark datasets, using standard architectures. Even with a $90$% reduction in size, SPvR’s sub-networks experience a minimal drop in test accuracy $(<1$%$)$ while on ImageNet1K, we outperform all baselines by achieving $<1$% Top-5 accuracy drop when pruning $70$% of ResNet50 parameters. Additionally, when compared to MobileNetV3, an SPvR pruned network improves the Top-1 accuracy by $3.3$% with $20$% less parameters. Furthermore, we empirically show that SPvR achieves reduced inference latency, underscoring its practical benefits for deploying neural networks on resource-constrained devices.

Cite this Paper

BibTeX
@InProceedings{pmlr-v286-hassan25b, title = {SPvR: Structured Pruning via Ranking}, author = {Hassan, Atif and Paik, Jiaul H. and Khare, Swanand}, booktitle = {Proceedings of the Forty-first Conference on Uncertainty in Artificial Intelligence}, pages = {1663--1676}, year = {2025}, editor = {Chiappa, Silvia and Magliacane, Sara}, volume = {286}, series = {Proceedings of Machine Learning Research}, month = {21--25 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v286/main/assets/hassan25b/hassan25b.pdf}, url = {https://proceedings.mlr.press/v286/hassan25b.html}, abstract = {Deep neural networks have achieved state-of-the-art performance in multiple domains but are increasingly resource-intensive, limiting their deployment on constrained devices. We introduce Structured Pruning via Ranking (SPvR), a novel structured pruning approach to address this challenge for classification tasks. SPvR prunes pre-trained networks in terms of function composition and network width while adhering to a user-specified parameter budget. Our method leverages local grouping and global ranking modules to generate smaller yet effective networks tailored to a given dataset and model. Finally, we train the pruned networks from scratch, instead of fine-tuning. Our evaluations demonstrate that SPvR significantly surpasses existing state-of-the-art pruning methods on benchmark datasets, using standard architectures. Even with a $90$% reduction in size, SPvR’s sub-networks experience a minimal drop in test accuracy $(<1$%$)$ while on ImageNet1K, we outperform all baselines by achieving $<1$% Top-5 accuracy drop when pruning $70$% of ResNet50 parameters. Additionally, when compared to MobileNetV3, an SPvR pruned network improves the Top-1 accuracy by $3.3$% with $20$% less parameters. Furthermore, we empirically show that SPvR achieves reduced inference latency, underscoring its practical benefits for deploying neural networks on resource-constrained devices.} }
Endnote
%0 Conference Paper %T SPvR: Structured Pruning via Ranking %A Atif Hassan %A Jiaul H. Paik %A Swanand Khare %B Proceedings of the Forty-first Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2025 %E Silvia Chiappa %E Sara Magliacane %F pmlr-v286-hassan25b %I PMLR %P 1663--1676 %U https://proceedings.mlr.press/v286/hassan25b.html %V 286 %X Deep neural networks have achieved state-of-the-art performance in multiple domains but are increasingly resource-intensive, limiting their deployment on constrained devices. We introduce Structured Pruning via Ranking (SPvR), a novel structured pruning approach to address this challenge for classification tasks. SPvR prunes pre-trained networks in terms of function composition and network width while adhering to a user-specified parameter budget. Our method leverages local grouping and global ranking modules to generate smaller yet effective networks tailored to a given dataset and model. Finally, we train the pruned networks from scratch, instead of fine-tuning. Our evaluations demonstrate that SPvR significantly surpasses existing state-of-the-art pruning methods on benchmark datasets, using standard architectures. Even with a $90$% reduction in size, SPvR’s sub-networks experience a minimal drop in test accuracy $(<1$%$)$ while on ImageNet1K, we outperform all baselines by achieving $<1$% Top-5 accuracy drop when pruning $70$% of ResNet50 parameters. Additionally, when compared to MobileNetV3, an SPvR pruned network improves the Top-1 accuracy by $3.3$% with $20$% less parameters. Furthermore, we empirically show that SPvR achieves reduced inference latency, underscoring its practical benefits for deploying neural networks on resource-constrained devices.
APA
Hassan, A., Paik, J.H. & Khare, S.. (2025). SPvR: Structured Pruning via Ranking. Proceedings of the Forty-first Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 286:1663-1676 Available from https://proceedings.mlr.press/v286/hassan25b.html.

Related Material