Research map / Graph learning / Graph algorithms and clustering
NASGEM: Neural Architecture Search via Graph Embedding Method
Hsin-Pai Cheng, Tunhou Zhang, Yixing Zhang, Shiyu Li, Feng Liang, Feng Yan, Meng Li, Vikas Chandra, Hai Li, Yiran Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 kernel-guided graph embedding rather than an MLP autoencoder (ablation) | NASGEM predictor · NASGEM-Bench, 10% training data | Kendall's tau: 0.45 -> 0.48 | Improved |
| Use kernel-guided graph embedding rather than an MLP autoencoder (ablation) | NASGEM predictor · NASGEM-Bench, 100% training data | Kendall's tau: 0.47 -> 0.55 | Improved |
| Use kernel-guided graph embedding rather than an MLP autoencoder (ablation) | NASGEM predictor · NASGEM-Bench, 10% training data | Pearson correlation coefficient: 0.57 -> 0.76 | Improved |
| Use kernel-guided graph embedding rather than an MLP autoencoder (ablation) | NASGEM predictor · NASGEM-Bench, 100% training data | Pearson correlation coefficient: 0.71 -> 0.85 | Improved |
| Increase graph node count or reduce embedding dimension (ablation) | NASGEM encoder · Generated graph pairs at varied node counts and embedding dimensions | similarity loss | Hurt |
| Use kernel-guided graph embedding rather than an MLP autoencoder | NASGEM predictor · ImageNet | top-1 test error (%): 24.1 -> 23.3 | Improved |
| Use GEMNet-B rather than MnasNet×1.3 as the backbone | RetinaNet · MS COCO minival | AP: 35.0 -> 35.7 | Improved |
| Use GEMNet-A rather than MnasNet×1.3 as the backbone | Faster RCNN · MS COCO minival | AP: 35.4 -> 35.3 | Hurt |
| Use GEMNet-B rather than MnasNet×1.3 as the backbone | Faster RCNN · MS COCO minival | AP: 35.4 -> 35.7 | Improved |
| Add graph-kernel guidance to the graph embedding (ablation) | NASGEM predictor · NASBench-101 | predicted global accuracy: around 0.2% | Improved |
Closest papers by meaning
- Neural Graph Embedding for Neural Architecture Search (AAAI 2020 · similarity 0.91)
- Understanding and Accelerating Neural Architecture Search With Training-Free and Theory-Grounded Metrics (TPAMI 2023 · similarity 0.90)
- NAS-Bench-Graph: Benchmarking Graph Neural Architecture Search (NeurIPS 2022 · similarity 0.89)
- Neural Architecture Search on ImageNet in Four GPU Hours: A Theoretically Inspired Perspective (ICLR 2021 · similarity 0.89)
- Fitting the Search Space of Weight-sharing NAS with Graph Convolutional Networks (AAAI 2021 · similarity 0.89)
- Encodings for Prediction-based Neural Architecture Search (ICML 2024 · similarity 0.87)
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.
