To Prune or not to Prune

By K. Barry, C. Mohammed. 1998

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Controlled Sparsity via Constrained Optimization or: How I Learned to Stop Tuning Penalties and Love Constraints (2022, background): “Zhu and Gupta [47] start from a partially or fully pre-trained model and consider a sparsification scheme in which the network density is gradually reduced, while fine-tuning the model to compensate for any potential loss in performance due to pruning.”
Non-vacuous Generalization Bounds at the ImageNet Scale: a PAC-Bayesian Compression Approach (2018, methodology): “Zhu & Gupta (2017) study pruning of MobileNet in the context of energy-efﬁcient inference in resource-constrained environments.”