Research map / Architectures and efficiency / Quantization
Enhanced Bayesian Compression via Deep Reinforcement Learning
Check 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 |
|---|---|---|---|
| Use EBC compression instead of the original full-precision model | LeNet-300-100 · MNIST | top-1 test error (%): 1.6 -> 1.8 | Hurt |
| Use EBC compression instead of the original full-precision model | LeNet-5 · MNIST | top-1 test error (%): 0.9 -> 1.0 | Hurt |
| Use EBC compression instead of the original full-precision model | VGG-16 · CIFAR-10 | top-1 test error (%): 8.4 -> 8.8 | Hurt |
| Use EBC compression instead of the original full-precision model | ResNet-18 · CIFAR-10 | top-1 test error (%): 6.8 -> 7.2 | Hurt |
| Use EBC compression instead of the original full-precision model | ResNet-18 · ImageNet validation | top-1 validation error (%): 31.6 -> 31.8 | Hurt |
| Use EBC compression instead of the original full-precision model | ResNet-18 · ImageNet validation | top-5 validation error (%): 11.3 -> 11.4 | Hurt |
| Use EBC rather than TTQ | ResNet-18 · ImageNet validation | top-1 validation error (%): 34.1 -> 31.8 | Helped |
Closest papers by meaning
- Bayesian Compression for Deep Learning (NeurIPS 2017 · similarity 0.89)
- Variational Bayesian Quantization (ICML 2020 · similarity 0.86)
- Online Learned Continual Compression with Adaptive Quantization Modules (ICML 2020 · similarity 0.86)
- N2N learning: Network to Network Compression via Policy Gradient Reinforcement Learning (ICLR 2018 · similarity 0.86)
- Retraining-Free Model Quantization via One-Shot Weight-Coupling Learning (CVPR 2024 · similarity 0.84)
- ReLU to the Rescue: Improve Your On-Policy Actor-Critic with Positive Advantages (ICML 2024 · similarity 0.84)
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.