Enhanced Bayesian Compression via Deep Reinforcement Learning
Abstract
In this paper, we propose an Enhanced Bayesian Compression method to flexibly compress the deep networks via reinforcement learning. Unlike the existing Bayesian compression method which cannot explicitly enforce quantization weights during training, our method learns flexible codebooks in each layer for an optimal network quantization. To dynamically adjust the state of codebooks, we employ an Actor-Critic network to collaborate with the original deep network. Different from most existing network quantization methods, our EBC does not require re-training procedures after the quantization. Experimental results show that our method obtains low-bit precision with acceptable accuracy drop on MNIST, CIFAR and ImageNet.
Cite
Text
Yuan et al. "Enhanced Bayesian Compression via Deep Reinforcement Learning." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019. doi:10.1109/CVPR.2019.00711Markdown
[Yuan et al. "Enhanced Bayesian Compression via Deep Reinforcement Learning." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019.](https://mlanthology.org/cvpr/2019/yuan2019cvpr-enhanced/) doi:10.1109/CVPR.2019.00711BibTeX
@inproceedings{yuan2019cvpr-enhanced,
title = {{Enhanced Bayesian Compression via Deep Reinforcement Learning}},
author = {Yuan, Xin and Ren, Liangliang and Lu, Jiwen and Zhou, Jie},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
year = {2019},
doi = {10.1109/CVPR.2019.00711},
url = {https://mlanthology.org/cvpr/2019/yuan2019cvpr-enhanced/}
}