Research map / Continual learning / Online and replay-based
Overcoming Multi-Model Forgetting in One-Shot NAS With Diversity Maximization
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 |
|---|---|---|---|
| Add NSAS loss to RandomNAS supernet training | RandomNAS · CIFAR-10 | test error (%): 2.85±0.08 -> 2.64 | Helped |
| Add NSAS loss to GDAS supernet training | GDAS · CIFAR-10 | test error (%): 2.93 -> 2.73 | Helped |
| Add NSAS loss to RandomNAS supernet training | RandomNAS · CIFAR-100 | test error (%): 17.63* -> 17.56 | Helped |
| Add NSAS loss to GDAS supernet training | GDAS · CIFAR-100 | test error (%): 18.38 -> 18.02 | Helped |
| Add NSAS loss to RandomNAS supernet training | RandomNAS · PTB | validation perplexity: 57.8 -> 59.22 | Hurt |
| Add NSAS loss to RandomNAS supernet training | RandomNAS · PTB | test perplexity: 55.5 -> 56.84 | Hurt |
| Add NSAS loss to GDAS supernet training | GDAS · PTB | validation perplexity: 59.8 -> 59.74 | Helped |
| Add NSAS loss to GDAS supernet training | GDAS · PTB | test perplexity: 57.5 -> 57.24 | Helped |
| Use NSAS loss instead of the normal cross-entropy loss | RandomNAS · Architecture ranking by inherited versus retrained validation accuracy | Kendall Tau: τ = 0.0909 -> τ = 0.4242 | Helped |
| Use NSAS loss instead of the normal cross-entropy loss | GDAS · Architecture ranking by inherited versus retrained validation accuracy | Kendall Tau: τ = −0.1818 -> τ = 0.3030 | Helped |
Closest papers by meaning
- SUMNAS: Supernet with Unbiased Meta-Features for Neural Architecture Search (ICLR 2022 · similarity 0.92)
- Efficient Few-Shot Neural Architecture Search by Counting the Number of Nonlinear Functions (AAAI 2025 · similarity 0.92)
- HEP-NAS: Towards Efficient Few-shot Neural Architecture Search via Hierarchical Edge Partitioning (AAAI 2025 · similarity 0.90)
- Subnet-Aware Dynamic Supernet Training for Neural Architecture Search (CVPR 2025 · similarity 0.90)
- Boosting Order-Preserving and Transferability for Neural Architecture Search: a Joint Architecture Refined Search and Fine-tuning Approach (CVPR 2024 · similarity 0.89)
- Revisiting Neural Networks for Few-Shot Learning: A Zero-Cost NAS Perspective (ICML 2025 · 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.