Research map / Architectures and efficiency / Mixture of experts
MoE-Enhanced Explainable Deep Manifold Transformation for Complex Data Embedding and Visualization
Zelin Zang, Yuhao Wang, Jinlin Wu, Hong Liu, Yue Shen, Zhen Lei, Stan Ziqing Li
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 |
|---|---|---|---|
| add MoE network (ablation) | DMT-ME · MNIST | SVC accuracy: 90.1 -> 97.8 | Improved |
| add MoE network (ablation) | DMT-ME · EPI | SVC accuracy: 79.2 -> 87.0 | Improved |
| add hyperbolic mapper (HMapper) (ablation) | DMT-ME · MNIST | SVC accuracy: 96.8 -> 97.8 | Improved |
| add hyperbolic mapper (HMapper) (ablation) | DMT-ME · EPI | SVC accuracy: 84.9 -> 87.0 | Improved |
| add Gumbel matcher exclusive loss routing (ablation) | DMT-ME · EPI | Trustworthiness: 67.9 -> 70.3 | Improved |
| add Gumbel matcher exclusive loss routing (ablation) | DMT-ME · MNIST | SVC accuracy: 97.5 -> 97.8 | Improved |
| add data augmentation (augmented neighborhood) (ablation) | DMT-ME · MNIST | SVC accuracy: 34.4 -> 97.8 | Improved |
| use L_SMM instead of InfoNCE objective (ablation) | DMT-ME · MNIST | SVC accuracy: 82.5 -> 97.8 | Improved |
| use L_SMM instead of InfoNCE objective (ablation) | DMT-ME · EPI | SVC accuracy: 35.6 -> 87.0 | Improved |
| use SMM loss instead of tSNE loss (gradient stability) (ablation) | DMT-ME · HCL | SVC accuracy: 15.1 -> 86.0 (CV 0.406 -> 0.144) | Improved |
| increase number of experts from 1 to 20 (ablation) | DMT-ME · 4-dataset average | SVC accuracy: 70.6 -> 74.7 | Improved |
| increase number of experts to 10 from 8 (ablation) | DMT-ME · 4-dataset average | SVC accuracy: 72.8 -> 70.9 | Mixed |
Closest papers by meaning
- Understanding How Dimension Reduction Tools Work: An Empirical Approach to Deciphering t-SNE, UMAP, TriMap, and PaCMAP for Data Visualization (JMLR 2021 · similarity 0.85)
- SpaceMAP: Visualizing High-Dimensional Data by Space Expansion (ICML 2022 · similarity 0.85)
- Finsler Multi-Dimensional Scaling: Manifold Learning for Asymmetric Dimensionality Reduction and Embedding (CVPR 2025 · similarity 0.85)
- HyperTools: a Python Toolbox for Gaining Geometric Insights into High-Dimensional Data (JMLR 2018 · similarity 0.84)
- Partition First, Embed Later: Laplacian-Based Feature Partitioning for Refined Embedding and Visualization of High-Dimensional Data (ICML 2025 · similarity 0.84)
- Harnessing Data Asymmetry in Manifold Learning (NeurIPS 2026 · similarity 0.83)
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.
