nbi 0.3.1

Neural Bayesian Inference

nbi: neural bayesian inference

Do you have challanging inference problems that are difficult to solve with standard MCMC or Nested Sampling methods? Are you looking to fit the same forward model to thousands or millions of observed datasets? nbi may be your solution.

nbi is an engine for Neural Posterior Estimation (NPE) focused on out-of-the-box application to common astronomical data, such as light curves and spectra. Compared to related packages, nbi requires minimal customization and can easily substitute for It also implements a custom NPE algorithm that integrates importance sampling, which allows for efficient, asymptotically exact results.

Installation

To install this package, we recommend that you create a dedicated conda environment with Python 3.7 or higher

conda create -n nbi python=3.10 && conda activate nbi

Then pip install this package

pip install nbi

Quick Start

The examples/ directory contains complete examples that demonstrates the functionality of nbi. A bare-bone example below illustrates the basic API, which follows the scikit-learn style:

import nbi

# specify hyperparameters
flow = {
    "n_dims": 1,
    "flow_hidden": 32,
    "num_blocks": 4
}
featurizer = {
    "type": "resnet-gru",
    "dim_in": 1,
    "depth": 3
}
engine = nbi.NBI(
    flow,
    featurizer,
    simulator,
    noise,
    priors
)
engine.fit(
    n_sims=1000,
    n_rounds=1,
    n_epochs=100
)
y_pred, weights = engine.predict(x_obs, x_err, n_samples=2000)

References

nbi: the Astronomer's Package for Neural Posterior Estimation (Zhang et al. 2023)

• Accepted to the "Machine Learning for Astrophysics" workshop at the International Conference for Machine Learning (ICML). Link to the paper will be updated here.

Masked Autoregressive Flow for Density Estimation (Papamakarios et al. 2017)
https://arxiv.org/abs/1705.07057

Featurizers: ResNet (He et al. 2015; https://arxiv.org/abs/1512.03385), Gated Recurrent Units (GRU; Cho et al. 2014; https://arxiv.org/abs/1406.1078), ResNet-GRU (Zhang et al. 2021; https://iopscience.iop.org/article/10.3847/1538-3881/abf42e)

Acknowledgments

The nbi package is expanded from code originally written for ''Real-time Likelihood-free Inference of Roman Binary Microlensing Events with Amortized Neural Posterior Estimation''' (Zhang et al. 2021). The Masked Autoregressive Flow in this package is partly adapted from the implementation in https://github.com/kamenbliznashki/normalizing_flows. Work on this project was supported by the National Science Foundation award #2206744 ("CDS&E: Accelerating Astrophysical Insight at Scale with Likelihood-Free Inference").