# Zebra: Memory Bandwidth Reduction for CNN Accelerators With Zero Block Regularization of Activation Maps

arXiv: https://arxiv.org/abs/2205.00779 · HTML: https://mlsanity.com/paper/2205.00779/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 9 of its 13 references that are in the map.
- First posted on arXiv 2 May 2022.

## Builds on (most cited first, 9 of 9)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift](https://mlsanity.com/paper/1502.03167/) (2015)
- [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://mlsanity.com/paper/1704.04861/) (2017)
- [Learning both Weights and Connections for Efficient Neural Networks](https://mlsanity.com/paper/1506.02626/) (2015)
- [Learning Efficient Convolutional Networks through Network Slimming](https://mlsanity.com/paper/1708.06519/) (2017)
- [Dynamic Channel Pruning: Feature Boosting and Suppression](https://mlsanity.com/paper/1810.05331/) (2018)
- [Channel Gating Neural Networks](https://mlsanity.com/paper/1805.12549/) (2018)
- [Dynamic Runtime Feature Map Pruning](https://mlsanity.com/paper/1812.09922/) (2018)

## Built on by (most cited first, 3 of 3)

- [Drastically Reducing the Number of Trainable Parameters in Deep CNNs by Inter-layer Kernel-sharing](https://mlsanity.com/paper/2210.14151/) (2022)
- [Dynamic Gradient Sparse Update for Edge Training](https://mlsanity.com/paper/2503.17959/) (2025)
- [Balancing Uncertainty and Diversity of Samples: Leveraging Diversity of Least, High Confidence Samples for Effective Active Learning](https://mlsanity.com/paper/2605.22169/) (2026)

## Abstract (arXiv)

The large amount of memory bandwidth between local buffer and external DRAM has become the speedup bottleneck of CNN hardware accelerators, especially for activation maps. To reduce memory bandwidth, we propose to learn pruning unimportant blocks dynamically with zero block regularization of activation maps (Zebra). This strategy has low computational overhead and could easily integrate with other pruning methods for better performance. The experimental results show that the proposed method can reduce 70% of memory bandwidth for Resnet-18 on Tiny-Imagenet within 1% accuracy drops and 2% accuracy gain with the combination of Network Slimming.