Some cool machine learning stuff. Provide code, explanations or whatever I might find interesting.
Author: Juan Maroñas Molano ([email protected]) [PRHLT Research Center, Universidad Politécnica de Valencia]
- Python version 3.7
- Requirements:
pip install -r requirements.txt
- Might find useful to run
./install.sh
In folder models you can find different models, follow instructions there:
-
models/Bayesian_NN/
Clean up and refactorize code for:
- Add GP and TGP
- Stochastic gradient MCMC ( Hamiltonian Monte Carlo ) BNN ref, for classification
- Point Estimate NN (should not go here but useful for comparison), for classification
- Probabilistic data augmentation using MCMC
- Probabilistic data augmentation with Mean Field VI (aka VAE)
- Probabilistic data augmentation with flows.
- Add regression example to the Bayesian NN models. The ones comparing MFVI and HMC