Research map / Architectures and efficiency / Neural architecture search
EMT-NAS:Transferring Architectural Knowledge Between Tasks From Different Datasets
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 |
|---|---|---|---|
| search architectures jointly across tasks using architectural crossover (ablation) | EMT-NAS on CIFAR-10 and CIFAR-100 · CIFAR-10 | accuracy (%): 96.23±0.20 -> 96.49±0.16 | Helped |
| search architectures jointly across tasks using architectural crossover (ablation) | EMT-NAS on CIFAR-10 and CIFAR-100 · CIFAR-100 | accuracy (%): 79.70±0.44 -> 80.68±0.43 | Helped |
| add fitness re-evaluation to multi-task search (ablation) | EMT-NAS · CIFAR-10 | accuracy (%): 96.49±0.16 -> 96.73±0.15 | Helped |
| add fitness re-evaluation to multi-task search (ablation) | EMT-NAS · CIFAR-100 | accuracy (%): 80.68±0.43 -> 81.86±0.10 | Helped |
| add fitness re-evaluation to single-task search (ablation) | Single-Tasking NAS · CIFAR-10 | accuracy (%): 96.23±0.20 -> 96.29±0.25 | No clear effect |
| add fitness re-evaluation to single-task search (ablation) | Single-Tasking NAS · CIFAR-100 | accuracy (%): 79.70±0.44 -> 80.48±0.28 | Helped |
| increase normal-cell stacking from two to three (ablation) | EMT-NAS · CIFAR-10 | accuracy (%): 96.73±0.15 -> 97.04±0.04 | Helped |
| increase normal-cell stacking from two to three (ablation) | EMT-NAS · CIFAR-100 | accuracy (%): 81.86±0.10 -> 82.60±0.38 | Helped |
| transfer architecture knowledge across related datasets during search | EMT-NAS · CIFAR-100 | accuracy (%): 79.70±0.44 -> 81.86±0.10 | Helped |
| transfer architecture knowledge across related datasets during search | EMT-NAS · CIFAR-10 | accuracy (%): 96.24±0.20 -> 96.73±0.15 | Helped |
| set random mating probability to 0.3 rather than 0 (ablation) | EMT-NAS · CIFAR-100 | test accuracy (%): 80.31±0.40 -> 81.86±0.10 | Helped |
| vary cross-task mating probability across the tested values (ablation) | EMT-NAS · CIFAR-100 | test accuracy (%): 80.31±0.40; 80.46±0.71; 81.11±0.54; 81.86±0.10; 81.26±0.27; 81.15±0.31 | Mixed |
Closest papers by meaning
- TransNAS-Bench-101: Improving Transferability and Generalizability of Cross-Task Neural Architecture Search (CVPR 2021 · similarity 0.85)
- Towards Fast Adaptation of Neural Architectures with Meta Learning (ICLR 2020 · similarity 0.84)
- EG-NAS: Neural Architecture Search with Fast Evolutionary Exploration (AAAI 2024 · similarity 0.84)
- Runtime Analysis of Evolutionary NAS for Multiclass Classification (ICML 2025 · similarity 0.83)
- Design Principle Transfer in Neural Architecture Search via Large Language Models (AAAI 2025 · similarity 0.83)
- Insights from the Use of Previously Unseen Neural Architecture Search Datasets (CVPR 2024 · similarity 0.82)
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.