Research map / Label-efficient and robust learning / Recognition and long-tail
GreedyNASv2: Greedier Search With a Greedy Path Filter
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 greedy path filtering and operation merging | GreedyNASv2 · ImageNet, MB-SE+MixConv+Shuffle (large) | ACC (%): 76.5 -> 77.5 | Helped |
| use greedy path filtering and operation merging | GreedyNASv2 · ImageNet, MB-SE+MixConv+Shuffle (large) | ACC on supernet (%): 35.1 -> 43.8 | Helped |
| use PU learning instead of PN learning | path filter on NAS-Bench-Macro · NAS-Bench-Macro, 1% training data | Precision: 79.52 -> 97.21 | Helped |
| use PU learning instead of PN learning | path filter on NAS-Bench-Macro · NAS-Bench-Macro, 1% training data | Recall: 85.82 -> 97.08 | Helped |
| add path-level shrinkage (ablation) | GreedyNASv2 · ImageNet, MB-SE+MixConv+Shuffle (large) | ACC in retraining (%): 76.8 -> 77.5 | Helped |
| add path-level shrinkage (ablation) | GreedyNASv2 · ImageNet, MB-SE+MixConv+Shuffle (large) | ACC on supernet (%): 42.1 -> 43.8 | Helped |
| add operation-level shrinkage (ablation) | GreedyNASv2 · ImageNet, MB-SE+MixConv+Shuffle (large) | ACC in retraining (%): 77.2 -> 77.5 | Helped |
| add operation-level shrinkage (ablation) | GreedyNASv2 · ImageNet, MB-SE+MixConv+Shuffle (large) | ACC on supernet (%): 39.6 -> 43.8 | Helped |
| lower the operation-merging threshold to 0.6 or 0.7 (ablation) | GreedyNASv2 operation shrinkage · ImageNet, MB-SE+MixConv+Shuffle | retraining accuracy | Hurt |
| use path filter during evolutionary search (ablation) | GreedyNASv2 · ImageNet, MB-SE+MixConv+Shuffle | searched-path accuracy | Helped |
| use learned similar-operation embeddings to identify operation pairs | GreedyNASv2 operation shrinkage · MB-SE, similar operation pairs | Spearman's rank correlation (%): 83.62 -> 99.26 | Helped |
Closest papers by meaning
- GreedyNAS: Towards Fast One-Shot NAS With Greedy Supernet (CVPR 2020 · similarity 0.94)
- PA&DA: Jointly Sampling Path and Data for Consistent NAS (CVPR 2023 · similarity 0.89)
- Generalizing Few-Shot NAS with Gradient Matching (ICLR 2022 · similarity 0.88)
- Efficient Few-Shot Neural Architecture Search by Counting the Number of Nonlinear Functions (AAAI 2025 · similarity 0.86)
- Subnet-Aware Dynamic Supernet Training for Neural Architecture Search (CVPR 2025 · similarity 0.86)
- HEP-NAS: Towards Efficient Few-shot Neural Architecture Search via Hierarchical Edge Partitioning (AAAI 2025 · 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.