Research map / Probabilistic and scientific modeling / Molecules and proteins
Scaling up Continuous-Time Markov Chains Helps Resolve Underspecification
Alkis Gotovos, Rebekka Burkholz, John Quackenbush, Stefanie Jegelka
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 the proposed approximate likelihood method instead of the previous exact approach | CTMC learning · Learning runtime, n=15 | runtime: 43\,\text{s} -> 4\,\text{s} | Improved |
| Use the proposed approximate likelihood method instead of the previous exact approach | CTMC learning · Learning runtime, n=20 | runtime: 121\,\text{m} -> 8\,\text{s} | Improved |
| Add independent items to the modeled ground set | CTMC recovery on synthetic data · Two synthetic datasets | KL divergence | Improved |
| Use the proposed MCMC proposal rather than a uniform proposal | Gradient approximation · Synthetic data, larger ground sets | gradient approximation error | Improved |
| Include more than 70 additional items | CTMC learning on TCGA glioblastoma data · PDGFRA amplification and point mutation | robustness of inferred time order | Improved |
| Scale the ground set to n=50 | CTMC learning on TCGA glioblastoma data · IDH1 mutation and TP53 mutation | robustness of inferred time order | Improved |
| Increase the modeled ground set rather than relying on a small subset | CTMC learning on TCGA glioblastoma data · MDM2 amplification and CDK4 amplification | clarity of inferred time order | No clear effect |
| Use the proposed MCMC proposal rather than a uniform proposal | Gradient approximation · Synthetic data, n=10 | gradient approximation error | No clear effect |
| Introduce attractive interactions between pairs of additional items (ablation) | Posterior time estimation · Synthetic settings with interacting added items | posterior time variance | Hurt |
| Introduce repulsive interactions between pairs of additional items (ablation) | Posterior time estimation · Synthetic settings with interacting added items | posterior time variance | No clear effect |
Closest papers by meaning
- Learning Multiple Markov Chains via Adaptive Allocation (NeurIPS 2019 · similarity 0.85)
- A survey and benchmark of high-dimensional Bayesian optimization of discrete sequences (NeurIPS 2024 · similarity 0.84)
- Fine-Tuning Discrete Diffusion Models via Reward Optimization with Applications to DNA and Protein Design (ICLR 2025 · similarity 0.83)
- Foundation Inference Models for Markov Jump Processes (NeurIPS 2024 · similarity 0.83)
- Simple Guidance Mechanisms for Discrete Diffusion Models (ICLR 2025 · similarity 0.83)
- Discrete Markov Probabilistic Models: An Improved Discrete Score-Based Framework with sharp convergence bounds under minimal assumptions (ICML 2025 · 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.
