An iterative exploitation-exploration loop lets simple saliency scores such as weight magnitude beat more complex sparse-training and pruning baselines on ImageNet and other benchmarks.
Pruning for Better Domain Generalizability
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
In this paper, we investigate whether we could use pruning as a reliable method to boost the generalization ability of the model. We found that existing pruning method like L2 can already offer small improvement on the target domain performance. We further propose a novel pruning scoring method, called DSS, designed not to maintain source accuracy as typical pruning work, but to directly enhance the robustness of the model. We conduct empirical experiments to validate our method and demonstrate that it can be even combined with state-of-the-art generalization work like MIRO(Cha et al., 2022) to further boost the performance. On MNIST to MNIST-M, we could improve the baseline performance by over 5 points by introducing 60% channel sparsity into the model. On DomainBed benchmark and state-of-the-art MIRO, we can further boost its performance by 1 point only by introducing 10% sparsity into the model. Code can be found at: https://github.com/AlexSunNik/Pruning-for-Better-Domain-Generalizability
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Advancing Weight and Channel Sparsification with Enhanced Saliency
An iterative exploitation-exploration loop lets simple saliency scores such as weight magnitude beat more complex sparse-training and pruning baselines on ImageNet and other benchmarks.