An Information-Theoretic Justification for Model Pruning
This work provides a theoretical justification for model pruning in neural network compression, which is incremental as it builds on existing compression literature.
The paper tackles the neural network compression problem by applying rate-distortion theory to analyze the trade-off between compression ratio and model performance, showing that pruning is essential for good compression and proposing a novel pruning strategy that outperforms baselines on CIFAR-10 and ImageNet datasets.
We study the neural network (NN) compression problem, viewing the tension between the compression ratio and NN performance through the lens of rate-distortion theory. We choose a distortion metric that reflects the effect of NN compression on the model output and derive the tradeoff between rate (compression) and distortion. In addition to characterizing theoretical limits of NN compression, this formulation shows that \emph{pruning}, implicitly or explicitly, must be a part of a good compression algorithm. This observation bridges a gap between parts of the literature pertaining to NN and data compression, respectively, providing insight into the empirical success of model pruning. Finally, we propose a novel pruning strategy derived from our information-theoretic formulation and show that it outperforms the relevant baselines on CIFAR-10 and ImageNet datasets.