Research map / Architectures and efficiency / Neural architecture search
M-NAS: Meta Neural 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 |
|---|---|---|---|
| add task-specific meta-weight modulation (ablation) | M-NAS · Toy regression harmonic | MSE: 1.973 ± 0.186 -> 1.232 ± 0.121 | Helped |
| add task-specific meta-weight modulation (ablation) | M-NAS · Toy regression averaged | MSE: 1.772 ± 0.356 -> 0.352 ± 0.032 | Helped |
| add task-specific meta-weight modulation (ablation) | M-NAS · Multi-datasets 5-way 1-shot | accuracy: 43.52% -> 48.57% | Helped |
| add task-specific meta-weight modulation (ablation) | M-NAS · Multi-datasets 5-way 5-shot | accuracy: 57.67 % -> 61.89 % | Helped |
| add task-specific meta-weight modulation (ablation) | M-NAS · MiniImagenet 5-way 1-shot | accuracy: 50.43 ± 1.73% -> 51.37 ± 1.41% | Helped |
| task-aware architectures instead of one shared architecture (ablation) | M-NAS-Shared · Toy regression averaged | MSE: 0.360 ± 0.037 -> 0.352 ± 0.032 | Helped |
| task-aware architectures instead of one shared architecture (ablation) | M-NAS-Shared · Multi-datasets 5-way 1-shot | accuracy: 47.88% -> 48.57% | Helped |
| task-aware architectures instead of one shared architecture (ablation) | M-NAS-Shared · Multi-datasets 5-way 5-shot | accuracy: 61.37 % -> 61.89 % | Helped |
| task-aware architectures instead of one shared architecture (ablation) | M-NAS-Shared · MiniImagenet 5-way 1-shot | accuracy: 50.78 ± 1.66% -> 51.37 ± 1.41% | Helped |
| searched shared architecture instead of hand-crafted 5 FCs, with modulation (ablation) | MAML-style meta-learner · Toy regression averaged | MSE: 0.502 ± 0.041 -> 0.360 ± 0.037 | Helped |
Closest papers by meaning
- Towards Fast Adaptation of Neural Architectures with Meta Learning (ICLR 2020 · similarity 0.93)
- Rapid Model Architecture Adaption for Meta-Learning (NeurIPS 2022 · similarity 0.92)
- Meta Architecture Search (NeurIPS 2019 · similarity 0.92)
- Towards Neural Architecture Search through Hierarchical Generative Modeling (ICML 2024 · similarity 0.89)
- Insights from the Use of Previously Unseen Neural Architecture Search Datasets (CVPR 2024 · similarity 0.89)
- einspace: Searching for Neural Architectures from Fundamental Operations (NeurIPS 2024 · similarity 0.89)
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.