Research map / Probabilistic and scientific modeling / Variational and generative sampling
L-C2ST: Local Diagnostics for Posterior Approximations in Simulation-Based Inference
Julia Linhart, Alexandre Gramfort, Pedro L. C. Rodrigues
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 |
|---|---|---|---|
| increase N_train from 10^2 to 10^5 | l-C2ST · SLCP N_cal=5000 | run-time (seconds): 5.92 -> 1.81 | Improved |
| increase N_train from 10^2 to 10^5 | l-C2ST-NF · SLCP N_cal=10000 | run-time (seconds): 43.99 -> 7.62 | Improved |
| increase N_cal from 5000 to 10000 | l-C2ST · SLCP N_train=10^2 | run-time (seconds): 5.92 -> 27.62 | Hurt |
| use l-C2ST instead of local-HPD | local-HPD · SLCP N_cal=10000, N_train=10^5 | run-time (seconds): 530.45 -> 3.65 | Improved |
| use l-C2ST-NF instead of l-C2ST | l-C2ST · SLCP and Two Moons | true positive rate | Improved |
| use l-C2ST-NF instead of local-HPD | local-HPD · Two Moons N_train=1000 | calibration samples to reach max TPR: 2000 | Improved |
| use l-C2ST instead of local-HPD | local-HPD · SLCP N_train=1000 | maximum TPR | Improved |
| increase observation dimensionality (Two Moons/SLCP to Bernoulli GLM) | l-C2ST · sbibm tasks | calibration samples to reach TPR=1: 2000 -> 5000 | Hurt |
| use regression-based oracle C2ST vs local-HPD | local-HPD · Two Moons | calibration samples to reach TPR=1: 5000 | No clear effect |
Closest papers by meaning
- Addressing Misspecification in Simulation-based Inference through Data-driven Calibration (ICML 2025 · similarity 0.86)
- CoLT: The conditional localization test for assessing the accuracy of neural posterior estimates (NeurIPS 2025 · similarity 0.86)
- Effortless, Simulation-Efficient Bayesian Inference using Tabular Foundation Models (NeurIPS 2025 · similarity 0.86)
- Consistency Models for Scalable and Fast Simulation-Based Inference (NeurIPS 2024 · similarity 0.85)
- Learning Likelihood-Free Reference Priors (ICML 2025 · similarity 0.85)
- Variational Inference with Coverage Guarantees in Simulation-Based Inference (ICML 2024 · similarity 0.84)
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.
