Research map / Architectures and efficiency / Neural architecture search
NAS-Bench-360: Benchmarking Neural Architecture Search on Diverse Tasks
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 DARTS (GAEA) instead of default WRN | WRN · NAS-Bench-360 Spherical | 0-1 error: 85.77 -> 48.23 | Helped |
| use DARTS (GAEA) instead of default WRN | WRN · NAS-Bench-360 NinaPro | 0-1 error: 6.78 -> 17.67 | Hurt |
| use DARTS (GAEA) instead of default WRN | WRN · NAS-Bench-360 FSD50K | 1 - mAP: 0.92 -> 0.94 | Hurt |
| use DARTS (GAEA) instead of default WRN | WRN · NAS-Bench-360 Darcy Flow | relative l2: 0.073 -> 0.026 | Helped |
| use DenseNAS random search instead of its original weight-sharing algorithm | DenseNAS · NAS-Bench-360 CIFAR-100 | 0-1 error: 25.98 -> 25.49 | No clear effect |
| use DenseNAS instead of default WRN | WRN · NAS-Bench-360 Spherical | 0-1 error: 85.77 -> 71.23 | Helped |
| tune WRN hyperparameters with ASHA | WRN · NAS-Bench-360 Spherical | 0-1 error: 85.77 -> 75.46 | Helped |
| tune WRN hyperparameters with ASHA | WRN · NAS-Bench-360 CIFAR-100 | 0-1 error: 23.35 -> 23.39 | No clear effect |
| use TE-NAS zero-cost proxy instead of GAEA | DARTS space · NAS-Bench-360 Darcy Flow | relative l2: 0.026 -> 0.012 | Helped |
| use TE-NAS zero-cost proxy instead of GAEA | DARTS space · NAS-Bench-360 Spherical | 0-1 error: 48.23 -> 56.87 | Hurt |
| use GAEA (ERM) instead of DARTS-V1 in NAS-Bench-201 space | NAS-Bench-201 · NinaPro | 0-1 error: 22.06 -> 7.69 | Helped |
Closest papers by meaning
- NAS-Bench-Suite: NAS Evaluation is (Now) Surprisingly Easy (ICLR 2022 · similarity 0.95)
- NAS-Bench-201: Extending the Scope of Reproducible Neural Architecture Search (ICLR 2020 · similarity 0.94)
- NAS-Bench-101: Towards Reproducible Neural Architecture Search (ICML 2019 · similarity 0.94)
- TransNAS-Bench-101: Improving Transferability and Generalizability of Cross-Task Neural Architecture Search (CVPR 2021 · similarity 0.93)
- einspace: Searching for Neural Architectures from Fundamental Operations (NeurIPS 2024 · similarity 0.91)
- Insights from the Use of Previously Unseen Neural Architecture Search Datasets (CVPR 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.