Research map / Architectures and efficiency / Neural architecture search
Neural Architecture Generator Optimization
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 |
|---|---|---|---|
| Use hierarchical HNAG rather than flat RNAG with BOHB | network generator optimization · CIFAR10 | test accuracy: 94.3(0.13) -> 96.6(0.15) | Helped |
| Use hierarchical HNAG rather than flat RNAG with BOHB | network generator optimization · CIFAR100 | test accuracy: 73.0(0.50) -> 79.3(0.31) | Helped |
| Use hierarchical HNAG rather than flat RNAG with BOHB | network generator optimization · SPORT8 | test accuracy: 93.6(0.76) -> 94.9(0.52) | Helped |
| Use hierarchical HNAG rather than flat RNAG with BOHB | network generator optimization · MIT67 | test accuracy: 68.3(0.78) -> 74.2(0.67) | Helped |
| Use hierarchical HNAG rather than flat RNAG with BOHB | network generator optimization · FLOWERS102 | test accuracy: 95.7(0.38) -> 97.9(0.18) | Helped |
| Use multi-objective BO to optimize for validation accuracy and sample memory | HNAG · CIFAR10 | memory consumption (MB): 57.9(0.52) -> 12.8(0.00) | Helped |
| Expand the search to include merge-option and operation probabilities (ablation) | HNAG-BOHB · CIFAR10 | validation accuracy: 95.6 -> 94.7 | Hurt |
| Expand the search to include merge-option and operation probabilities (ablation) | HNAG-BOHB · CIFAR100 | validation accuracy: 77.2 -> 74.5 | Hurt |
| Expand the search to include merge-option and operation probabilities (ablation) | HNAG-BOHB · SPORT8 | validation accuracy: 95.3 -> 94.0 | Hurt |
| Expand the search to include merge-option and operation probabilities (ablation) | HNAG-BOHB · MIT67 | validation accuracy: 71.8 -> 68.8 | Hurt |
| Expand the search to include merge-option and operation probabilities (ablation) | HNAG-BOHB · FLOWERS102 | validation accuracy: 93.3 -> 91.0 | Hurt |
| Optimize stage ratios and channel ratios (ablation) | HNAG-BOHB · CIFAR10 | validation accuracy: 95.6 -> 95.7 | No clear effect |
Closest papers by meaning
- Interpretable Neural Architecture Search via Bayesian Optimisation with Weisfeiler-Lehman Kernels (ICLR 2021 · similarity 0.93)
- Towards Neural Architecture Search through Hierarchical Generative Modeling (ICML 2024 · similarity 0.92)
- Evaluating The Search Phase of Neural Architecture Search (ICLR 2020 · similarity 0.92)
- EG-NAS: Neural Architecture Search with Fast Evolutionary Exploration (AAAI 2024 · similarity 0.91)
- Insights from the Use of Previously Unseen Neural Architecture Search Datasets (CVPR 2024 · similarity 0.91)
- einspace: Searching for Neural Architectures from Fundamental Operations (NeurIPS 2024 · similarity 0.91)
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.