Research map / Architectures and efficiency / Neural architecture search
DARTS: Differentiable Architecture Search
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 the second-order architecture-gradient approximation instead of the first-order approximation (ablation) | DARTS · CIFAR-10 test | test error (%): 3.00 $\pm$ 0.14 -> 2.76 $\pm$ 0.09 | Helped |
| use the second-order architecture-gradient approximation instead of the first-order approximation (ablation) | DARTS · PTB test | perplexity: 57.6 -> 55.7 | Helped |
| use DARTS search instead of random search | architecture search · CIFAR-10 test | test error (%): 3.29 $\pm$ 0.15 -> 2.76 $\pm$ 0.09 | Helped |
| use DARTS search instead of random search | architecture search · PTB test | perplexity: 59.4 -> 55.7 | Helped |
| jointly optimize architecture and weights over training plus validation data using coordinate descent instead of bilevel optimization (ablation) | DARTS convolutional architecture search · CIFAR-10 test | test error (%): 2.76 $\pm$ 0.09 -> 4.16 $\pm$ 0.16 | Hurt |
| optimize architecture and weights simultaneously over training plus validation data using SGD instead of bilevel optimization (ablation) | DARTS convolutional architecture search · CIFAR-10 test | test error (%): 2.76 $\pm$ 0.09 -> 3.56 $\pm$ 0.10 | Hurt |
| increase search network depth from 8 to 20 cells while reducing initial channels (ablation) | DARTS convolutional search · CIFAR-10 test | test error (%): 2.76 $\pm$ 0.09 -> 2.88 $\pm$ 0.09 | Hurt |
| transfer the CIFAR-10-discovered cell to ImageNet mobile classification | DARTS · ImageNet mobile test | top-1 error (%): 26.7 vs 24.3 | Mixed |
| transfer the PTB-discovered cell to WikiText-2 | DARTS · WikiText-2 test | perplexity: 70.4 -> 69.6 | Helped |
Closest papers by meaning
- Understanding and Robustifying Differentiable Architecture Search (ICLR 2020 · similarity 0.91)
- iDARTS: Differentiable Architecture Search with Stochastic Implicit Gradients (ICML 2021 · similarity 0.91)
- BaLeNAS: Differentiable Architecture Search via the Bayesian Learning Rule (CVPR 2022 · similarity 0.90)
- Stabilizing Differentiable Architecture Search via Perturbation-based Regularization (ICML 2020 · similarity 0.90)
- DrNAS: Dirichlet Neural Architecture Search (ICLR 2021 · similarity 0.89)
- EG-NAS: Neural Architecture Search with Fast Evolutionary Exploration (AAAI 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.