deepbayes 0.1.0

Package for doing approximate Bayesian inference on deep neural networks [based in TF 2.0+]

deepbayes

deepbayes is a package that aims to make training Bayesian neural networks (BNNs) simple.

Implemented on top of TensorFlow 2.0 and working with tf.keras as a primary interface, one can simply declare a keras (or TF) model and pass it to the desired deepbayes optimizer and it will compile the model into an object that is ready to perform approximate inference.

deepbayes is actively under development, and location and names of packages may change (such changes will be noted at the bottom of this repo). jax support coming soon :)

Install:

pip install deepbayes

Supported Optimizers

BayesByBackprop

Implemented based on description in https://arxiv.org/abs/1505.05424

Hamiltonian (Hybrid) Monte Carlo

Implemented based on description in https://arxiv.org/abs/1206.1901

NoisyAdam (aka Vadam)

Implemented based on description in https://arxiv.org/abs/1712.02390

Variational Online Gauss Newton (aka VOGN)

Implemented based on description in https://arxiv.org/abs/1806.04854 and https://arxiv.org/abs/2002.10060

Stochastic Weight Averaging - Gaussian (aka SWAG)

Implemented bsed on description in https://arxiv.org/pdf/1902.02476.pdf

Stochastic Gradient Descent

Not sure what to reference you to here... a google will suffice

Supported Analysis

In addition to producing posterior distributions for BNNs, deepbayes also supports some basic analysis including

• Adversarial attacks for BNNs (FGSM, PGD, CW)
• Uncertainty quantification
• Statistical guarentees of robustness - https://arxiv.org/pdf/1903.01980.pdf
• Probabalistic safety estimates - https://arxiv.org/pdf/2004.10281.pdf

Dependancies:

tensorflow, tensorflow-probability, numpy, tqdm, statsmodels,

Future Support

Below we have a tentative to-do list of inference methods to impliment (7 to be exact) and other properties we want to tool to have.

• HTML documentation
• Non-Gaussian Posterior (requires extension of posterior representation as well)
• Stochastic Variational Inference (SVI)
• Probabilistic Backprop (PBP)
• Sequential Monte Carlo optimizer (SMC)
• Monte Carlo Dropout optimizer (MCD)
• Stochastic Gradient Langevin Dynamics optimizer (SGLD)
• Stochastic Gradient Markov Chain Monte Carlo (SGMCMC)
• Riemann manifold HMC (RMHMC)
• JAX autodif library instead of TF (deepbayes.jax.optimizers)
• Binary Bayesian neural networks