Research map / Architectures and efficiency / Pruning and efficient tuning
Data-Independent Neural Pruning via Coresets
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| prune neurons with the data-independent coreset method | LeNet-300-100 · MNIST | error (%): 2.16 -> 2.03 | Helped |
| prune neurons with the data-independent coreset method | LeNet-300-100 · MNIST | compression: 90% | Helped |
| prune dense-layer neurons with the coreset method | VGG-16 · CIFAR-10 | error (%): 8.95 -> 8.16 | Helped |
| prune dense-layer neurons with the coreset method | VGG-16 · CIFAR-10 | compression: 75% | Helped |
| select neurons with the coreset algorithm instead of uniform sampling | LeNet-200-105 · MNIST | approximation error | Helped |
| select neurons with the coreset algorithm instead of uniform sampling (ablation) | LeNet-300-100 · MNIST | classification accuracy | Helped |
| add fine-tuning after neuron selection (ablation) | LeNet-300-100 with coreset selection · MNIST | classification accuracy | Helped |
| add fine-tuning after uniform neuron sampling (ablation) | LeNet-300-100 with uniform sampling · MNIST | classification accuracy | Helped |
| use uniform sampling rather than coreset selection (ablation) | LeNet-300-100 · MNIST | fine-tuning time: close to 2 epochs; about half of that time | Hurt |
Closest papers by meaning
- Pruning Neural Networks via Coresets and Convex Geometry: Towards No Assumptions (NeurIPS 2022 · similarity 0.95)
- “Learning-Compression” Algorithms for Neural Net Pruning (CVPR 2018 · similarity 0.91)
- Data-Dependent Coresets for Compressing Neural Networks with Applications to Generalization Bounds (ICLR 2019 · similarity 0.91)
- Fast as CHITA: Neural Network Pruning with Combinatorial Optimization (ICML 2023 · similarity 0.89)
- EPSD: Early Pruning with Self-Distillation for Efficient Model Compression (AAAI 2024 · similarity 0.88)
- Towards Meta-Pruning via Optimal Transport (ICLR 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.