RAT-SPN

Code for UAI'19: Random Sum-Product Networks: A Simple and Effective Approach to Probabilistic Deep Learning

V0.2

• RAT-SPN model
• Experiments for generative learning of RAT-SPNs using EM
• Experiments for discriminative learning of RAT-SPNs using Adam

Setup

git clone https://github.com/cambridge-mlg/RAT-SPN

cd RAT-SPN

./install_tensorflow_venv.sh

source ratspn_venv/bin/activate

python download_preprocess_data.py

Quick Run for Generative Experiments

This will simply train a single RAT-SPN (no crossvalidation).

python quick_run_rat_spn_generative.py

python quick_eval_rat_spn_generative.py

Quick Run for Discriminative Training on MNIST

This will simply train a single RAT-SPN for each depth.

python quick_run_rat_spn_mnist.py

python quick_eval_rat_spn_discriminative.py

Full Training

See the run_.py and eval_.py files

Acknowledgments

This project received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie Grant Agreement No. 797223 (HYBSPN).