Research map / Probabilistic and scientific modeling / Variational and generative sampling
Analysis of high-dimensional Continuous Time Markov Chains using the Local Bouncy Particle Sampler
Tingting Zhao, Alexandre Bouchard-Côté
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 LBPS-HMC (trajectory length 0.2) instead of HMC alone | Bayesian GLM chain GTR sampler · Protein kinase real data, 20x20 rate matrix | minimum ESS across 190 parameters: 95 -> 773 | Improved |
| use LBPS-HMC (trajectory length 0.2) instead of HMC alone | Bayesian GLM chain GTR sampler · Protein kinase real data, 20x20 rate matrix | median ESS across 190 parameters: 740 -> 1864 | Improved |
| use LBPS-HMC instead of HMC as rate matrix dimension grows from 5x5 | Bayesian GLM chain GTR sampler · Synthetic rate matrices 5x5 to 30x30 | minimum ESS per second | Mixed |
| increase LBPS trajectory length (ablation) | LBPS-HMC · Synthetic 20x20 rate matrix | maximum ESS per second | Improved |
| vary LBPS trajectory length (0.025 to 0.25) (ablation) | LBPS-HMC · Synthetic 20x20 rate matrix | minimum ESS per second | No clear effect |
| increase LBPS trajectory length from 0.025 to 0.25 (ablation) | LBPS-HMC · Synthetic 20x20 rate matrix | walltime (seconds): 240677 -> 403932 | Hurt |
| combine LBPS for bivariate weights with HMC for univariate weights | Bayesian GLM chain GTR sampler · Synthetic 8x8 rate matrix | ARD between posterior samples: median ARD 2.70% | No clear effect |
Closest papers by meaning
- Stochastic Bouncy Particle Sampler (ICML 2017 · similarity 0.88)
- A Continuous Time Framework for Discrete Denoising Models (NeurIPS 2022 · similarity 0.84)
- Chain of Log-Concave Markov Chains (ICLR 2024 · similarity 0.83)
- Piecewise deterministic generative models (NeurIPS 2024 · similarity 0.82)
- Discrete Markov Probabilistic Models: An Improved Discrete Score-Based Framework with sharp convergence bounds under minimal assumptions (ICML 2025 · similarity 0.82)
- Multi-Marginal Couplings for Metropolis--Hastings (NeurIPS 2026 · similarity 0.82)
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.
