Journal
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
Volume 33, Issue 12, Pages 7237-7250Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TNNLS.2021.3084527
Keywords
Neurons; Sensitivity; Biological neural networks; Computer architecture; Network topology; Topology; Training; Compression; deep networks; pruning; regularization; sparse networks
Categories
Funding
- Sisvel Technology, Turin, Italy
- European Union's Horizon 2020 Research and Innovation Programme [825111]
Ask authors/readers for more resources
The method utilizes neuron sensitivity as a regularization term to learn sparse topologies and effectively compress neural networks for resource-constrained devices. Experimental results show that the method achieves competitive compression ratios compared to state-of-the-art references.
Deep neural networks include millions of learnable parameters, making their deployment over resource-constrained devices problematic. Sensitivity-based regularization of neurons (SeReNe) is a method for learning sparse topologies with a structure, exploiting neural sensitivity as a regularizer. We define the sensitivity of a neuron as the variation of the network output with respect to the variation of the activity of the neuron. The lower the sensitivity of a neuron, the less the network output is perturbed if the neuron output changes. By including the neuron sensitivity in the cost function as a regularization term, we are able to prune neurons with low sensitivity. As entire neurons are pruned rather than single parameters, practical network footprint reduction becomes possible. Our experimental results on multiple network architectures and datasets yield competitive compression ratios with respect to state-of-the-art references.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available