Research map / Probabilistic and scientific modeling / Variational and generative sampling
SpHMC: Spectral Hamiltonian Monte Carlo
Haoyi Xiong, Kafeng Wang, Jiang Bian, Zhanxing Zhu, Cheng-Zhong Xu, Zhishan Guo, Jun Huan
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 |
|---|---|---|---|
| Disable Metropolis-Hastings correction (ablation) | SpHMC · EMNIST, l2-SVM | test error (%): 15.38 -> 22.71 | Hurt |
| Disable Metropolis-Hastings correction (ablation) | SpHMC · MNIST, SVM | test error (%): 3.59 -> 3.61 | No clear effect |
| Disable Metropolis-Hastings correction (ablation) | SpHMC · MNIST, l2-Perceptron | test error (%): 12.54 -> 18.84 | Hurt |
| Augment with SpHMC samples instead of original data only | SVM · MNIST | test error (%): 5.96 -> 3.59 | Improved |
| Augment with SpHMC samples instead of original data only | l2-SVM · Fashion MNIST | test error (%): 19.88 -> 15.19 | Improved |
| Augment with SpHMC samples instead of original data only | SVM · EMNIST | test error (%): 21.21 -> 13.14 | Improved |
| Use SpHMC instead of Kernel HMC for augmentation | l1-Log Reg · MNIST | test error (%): 34.14 -> 5.49 | Improved |
| Use SpHMC instead of Kernel HMC for augmentation | SVM · EMNIST | test error (%): 25.18 -> 13.14 | Improved |
| Use SpHMC instead of conditional GAN for augmentation | l2-Log Reg · EMNIST | test error (%): 33.21 -> 14.41 | Improved |
| Use SpHMC instead of GAN for augmentation | SVM · MNIST | test error (%): 5.99 -> 3.59 | Improved |
| Use SpHMC samples instead of plain compression (LASSO spectral representation) | SVM · MNIST | test error (%): 3.61 -> 3.59 | Improved |
| Use SpHMC instead of compression | l2-Perceptron · Fashion MNIST | test error (%): 25.09 -> 19.82 | Improved |
Closest papers by meaning
- Stochastic Gradient Hamiltonian Monte Carlo Methods with Recursive Variance Reduction (NeurIPS 2019 · similarity 0.91)
- On the Convergence of Hamiltonian Monte Carlo with Stochastic Gradients (ICML 2021 · similarity 0.89)
- Stochastic Variance-Reduced Hamilton Monte Carlo Methods (ICML 2018 · similarity 0.89)
- A Hybrid Stochastic Gradient Hamiltonian Monte Carlo Method (AAAI 2021 · similarity 0.89)
- Nonasymptotic analysis of Stochastic Gradient Hamiltonian Monte Carlo under local conditions for nonconvex optimization (JMLR 2024 · similarity 0.89)
- Metropolis Adjusted Microcanonical Hamiltonian Monte Carlo (NeurIPS 2025 · similarity 0.86)
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.
