Research map / Architectures and efficiency / Neural architecture search
TA-GATES: An Encoding Scheme for Neural Network Architectures
Xuefei Ning, Zixuan Zhou, Junbo Zhao, Tianchen Zhao, Yiping Deng, Changcheng Tang, Shuang Liang, Huazhong Yang, Yu Wang
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 TA-GATES encoder instead of GATES | GATES · NB101, 1% training | Kendall's Tau: 0.6321 -> 0.6686 | Improved |
| use TA-GATES encoder instead of GATES | GATES · NB201, 0.1% training | Kendall's Tau: 0.4309 -> 0.5382 | Improved |
| use TA-GATES encoder instead of GATES | GATES · NB201, 0.5% training | Kendall's Tau: 0.6702 -> 0.6707 | No clear effect |
| use TA-GATES encoder instead of GATES | GATES · NDS ENAS, 100% training | Kendall's Tau: 0.6467 -> 0.6683 | Improved |
| add symmetry breaking (Add) to TA-GATES (ablation) | TA-GATES (None) · NB101, 5% training | Kendall's Tau: 0.7581 -> 0.7744 | Improved |
| add symmetry breaking (Add) to TA-GATES (ablation) | TA-GATES (None) · NB301, 1% training | Kendall's Tau: 0.6182 -> 0.6351 | Improved |
| add symmetry breaking (Add) to TA-GATES (ablation) | TA-GATES (None) · NB201, 0.5% training | Kendall's Tau: 0.6838 -> 0.6707 | Hurt |
| use Random symmetry breaking instead of none (ablation) | TA-GATES (None) · NB101, 1% training | Kendall's Tau: 0.6510 -> 0.6425 | Hurt |
| use Concat symmetry breaking instead of none (ablation) | TA-GATES (None) · NB101, 1% training | Kendall's Tau: 0.6510 -> 0.6585 | Improved |
| use Add instead of Concat symmetry breaking (ablation) | TA-GATES · NB101, 1% training | Kendall's Tau: 0.6585 -> 0.6686 | Improved |
| TA-GATES with half+final signals vs Multi-GATES (half accuracy) | Multi-GATES · NB101, 5% training, half accuracy | Kendall's Tau: 0.2912 -> 0.4615 | Improved |
| TA-GATES with half+final signals vs Multi-GATES (final accuracy) | Multi-GATES · NB101, 5% training, final accuracy | Kendall's Tau: 0.3341 -> 0.5850 | Improved |
Later papers that compared with this method
1 later accepted paper reports TAGATES as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Encodings for Prediction-based Neural Architecture Search (ICML 2024 · better on 3 of 4 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- A Generic Graph-Based Neural Architecture Encoding Scheme With Multifaceted Information (TPAMI 2022 · similarity 0.89)
- A Generic Graph-based Neural Architecture Encoding Scheme for Predictor-based NAS (ECCV 2020 · similarity 0.88)
- Neural Architecture Optimization (NeurIPS 2018 · similarity 0.88)
- Encodings for Prediction-based Neural Architecture Search (ICML 2024 · similarity 0.86)
- Building Optimal Neural Architectures using Interpretable Knowledge (CVPR 2024 · similarity 0.85)
- NN-Former: Rethinking Graph Structure in Neural Architecture Representation (CVPR 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.
