Research map / Probabilistic and scientific modeling / Molecules and proteins
Diffusion Earth Mover’s Distance and Distribution Embeddings
Alexander Y Tong, Guillaume Huguet, Amine Natik, Kincaid Macdonald, Manik Kuchroo, Ronald Coifman, Guy Wolf, Smita Krishnaswamy
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 Diffusion EMD instead of ClusterTree | Distribution embeddings on Spherical MNIST · Spherical MNIST | Accuracy: 91.91 -> 95.94 | Improved |
| Use Diffusion EMD instead of ClusterTree | Distribution embeddings on Spherical MNIST · Spherical MNIST | P@10: 0.393 -> 0.611 | Improved |
| Use Diffusion EMD instead of ClusterTree | Distribution embeddings on Spherical MNIST · Spherical MNIST | Spearman ρ: 0.484 -> 0.673 | Improved |
| Use Diffusion EMD instead of QuadTree | Distribution embeddings on Spherical MNIST · Spherical MNIST | Accuracy: 79.56 -> 95.94 | Improved |
| Use Diffusion EMD instead of QuadTree | Distribution embeddings on Spherical MNIST · Spherical MNIST | P@10: 0.294 -> 0.611 | Improved |
| Use Diffusion EMD instead of QuadTree | Distribution embeddings on Spherical MNIST · Spherical MNIST | Spearman ρ: 0.335 -> 0.673 | Improved |
| Use Diffusion EMD instead of QuadTree | Distribution embeddings on Spherical MNIST · Spherical MNIST | Time: 16m -> 34m | Hurt |
| Reduce the interpolative-decomposition rank threshold (ablation) | Diffusion EMD · Swiss roll | Accuracy | Hurt |
| Subsample centers in the L1 representation using interpolative decomposition (ablation) | Diffusion EMD · Swiss roll | Representation size | Improved |
| Vary the maximum scale K after a sufficient minimum (ablation) | Diffusion EMD with Chebyshev approximation · Swiss roll | P@10 | No clear effect |
| Increase the Chebyshev polynomial order past a sufficient point (ablation) | Diffusion EMD with Chebyshev approximation · Swiss roll | P@10 | No clear effect |
| Vary the number of scales past a sufficient point (ablation) | Diffusion EMD with Chebyshev approximation · Swiss roll | P@10 | No clear effect |
Closest papers by meaning
- Probability Density Geodesics in Image Diffusion Latent Space (CVPR 2025 · similarity 0.85)
- Learning Distances from Data with Normalizing Flows and Score Matching (ICML 2025 · similarity 0.84)
- Toward a Unified Geometry Understanding : Riemannian Diffusion Framework for Graph Generation and Prediction (NeurIPS 2025 · similarity 0.83)
- Manifold Diffusion Fields (ICLR 2024 · similarity 0.83)
- Follow the Energy, Find the Path: Riemannian Metrics from Energy-Based Models (NeurIPS 2025 · similarity 0.83)
- Wasserstein Wormhole: Scalable Optimal Transport Distance with Transformer (ICML 2024 · 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.
