Software

An initial list of SBI-related software packages

SBI (python) - general purpose SBI framework
SBI Benchmarking (python) - for benchmarking
MadMiner - aimed at particle physics
swyft (python) the reference implementation of Truncated Marginal Neural Ratio Estimation (TMNRE), an efficient, simulation-based inference technique for complex data and expensive simulators
carl an early NN-based SBI package
lampe (python) is a flexible, amortized sbi library. The philosophy of LAMPE is to avoid obfuscation and expose all components so it is most useful for researchers.
pydelfi (python) with neural density estimators and adaptive acquisition of simulations.

Latest Posts

A comparative study of simulation-based inference methods for epidemic models with identifiability considerations

G Jang, KS Candan, G Chowell - PLOS Computational Biology, 2026 - journals.plos.org

… Simulation-based inference provides a promising alternative by approximating posterior distributions through forward simulations rather than an explicit likelihood …

Benchmark of Likelihood-Free Inference Methods based on Neural and Optimal Transport Approaches

S Aka, M Kratz, P Naveau - arXiv preprint arXiv:2605.30516, 2026 - arxiv.org

… Simulation-based inference (SBI) has become an increasingly important framework for parameter estimation in models for which simulation is feasible, including cases …

Local Manifold Identification with Latent Linear Models and OT Flows

S Khoo, S Liu, M Beaumont - ICML 2026 Workshop on Foundations of … - openreview.net

… model’s intractable likelihood, we adopt a simulationbased inference approach. By estimating the Fisher score from simulated observation-parameter pairs, we perform …

Conservative neural posterior estimation via distributionally robust training

W Laplante, Y Hikida, C Dellaporta, FX Briol… - arXiv preprint arXiv …, 2026 - arxiv.org

… Simulation-based inference with neural posterior estimation (NPE) often yields overconfident and unreliable posteriors under limited simulation budgets. To address this, …