Research map / Architectures and efficiency / Neural architecture search
$\beta $-DARTS++: Bi-Level Regularization for Proxy-Robust Differentiable Architecture Search
Peng Ye, Tong He, Baopu Li, Tao Chen, Lei Bai, Wanli Ouyang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add Beta-Decay regularization to SDARTS-RS (ablation) | SDARTS-RS · NAS-Bench-201 ImageNet16-120 test | accuracy (%): 26.29 -> 46.34 | Improved |
| add Beta-Decay regularization to RDARTS-L2 (ablation) | RDARTS-L2 · NAS-Bench-201 ImageNet16-120 test | accuracy (%): 18.41 -> 46.34 | Improved |
| add Beta-Decay regularization to SNAS (ablation) | SNAS · NAS-Bench-201 ImageNet16-120 test | accuracy (%): 26.10 -> 46.34 | Improved |
| add flooding (beta-DARTS to beta-DARTS++), 1% proxy data (ablation) | beta-DARTS · NAS-Bench-201 CIFAR-10 test | accuracy (%): 91.11 -> 93.20 | Improved |
| add flooding (beta-DARTS to beta-DARTS++), 10% proxy data (ablation) | beta-DARTS · NAS-Bench-201 CIFAR-10 test | accuracy (%): 89.89 -> 93.34 | Improved |
| add flooding (beta-DARTS to beta-DARTS++), 2% proxy data (ablation) | beta-DARTS · NAS-Bench-201 CIFAR-10 test | accuracy (%): 90.44 -> 93.39 | Improved |
| add flooding (beta-DARTS to beta-DARTS++), 1% proxy data (ablation) | beta-DARTS · NAS-Bench-201 ImageNet16-120 test | accuracy (%): 40.80 -> 45.47 | Improved |
| add flooding (beta-DARTS to beta-DARTS++), 2 proxy channels (ablation) | beta-DARTS · NAS-Bench-201 CIFAR-10 test | accuracy (%): 93.56 -> 93.96 | Improved |
| add flooding (beta-DARTS to beta-DARTS++), 1 proxy channel (ablation) | beta-DARTS · NAS-Bench-201 CIFAR-10 test | accuracy (%): 93.76 -> 93.68 | No clear effect |
| add weight flooding to DARTS, 8 proxy channels (ablation) | DARTS · NAS-Bench-201 CIFAR-10 test | accuracy (%): 54.30 -> 88.28 | Improved |
| add weight flooding to DARTS, 1% proxy data (ablation) | DARTS · NAS-Bench-201 CIFAR-10 test | accuracy (%): 54.30 -> 54.30 | No clear effect |
| search on CIFAR-100 instead of CIFAR-10 (ablation) | beta-DARTS · NAS-Bench-201 CIFAR-10 test | accuracy (%): 94.36 -> 93.99 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DARTS- — DARTS-: Robustly Stepping out of Performance Collapse Without Indicators (ICLR 2021) · better on 7 of 7 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- b-DARTS: Beta-Decay Regularization for Differentiable Architecture Search (CVPR 2022 · similarity 0.97)
- iDARTS: Differentiable Architecture Search with Stochastic Implicit Gradients (ICML 2021 · similarity 0.93)
- Stabilizing Differentiable Architecture Search via Perturbation-based Regularization (ICML 2020 · similarity 0.92)
- $\Lambda$-DARTS: Mitigating Performance Collapse by Harmonizing Operation Selection among Cells (ICLR 2023 · similarity 0.92)
- EG-NAS: Neural Architecture Search with Fast Evolutionary Exploration (AAAI 2024 · similarity 0.91)
- IS-DARTS: Stabilizing DARTS through Precise Measurement on Candidate Importance (AAAI 2024 · similarity 0.90)
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.
