SeReNe: Sensitivity based Regularization of Neurons for Structured Sparsity in Neural Networks

TL;DR

SeReNe introduces a sensitivity-based regularization technique that prunes neurons with low sensitivity to create structured sparsity in neural networks, enabling efficient deployment on resource-limited devices.

Contribution

The paper presents a novel regularization method leveraging neuron sensitivity to prune entire neurons, improving structured sparsity and deployment efficiency.

Findings

Achieves competitive compression ratios on multiple architectures.

Prunes neurons with minimal impact on network output.

Effective across various datasets and network types.

Abstract

Deep neural networks include millions of learnable parameters, making their deployment over resource-constrained devices problematic. SeReNe (Sensitivity-based Regularization of Neurons) 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 areable to prune neurons with low sensitivity. As entire neurons are pruned rather then 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.