Awesome

GP-VAE: Deep Probabilistic Time Series Imputation

Code for paper

Overview

Our approach utilizes Variational Autoencoders with Gaussian Process prior for time series imputation.

• The inference model takes time series with missingness and predicts variational parameters for multivariate Gaussian variational distribution.

• The Gaussian Process prior encourages latent representations to capture the temporal correlations in data.

• The generative model takes the sample from posterior approximation and reconstructs the original time series with imputed missing values.

Dependencies

• Python >= 3.6
• TensorFlow = 1.15
• Some more packages: see requirements.txt

Run

1. Clone or download this repo. cd yourself to it's root directory.

2. Grab or build a working python enviromnent. Anaconda works fine.

3. Install dependencies, using pip install -r requirements.txt

4. Download data: bash data/load_{hmnist, sprites, physionet}.sh.

5. Run command CUDA_VISIBLE_DEVICES=* python train.py --model_type {vae, hi-vae, gp-vae} --data_type {hmnist, sprites, physionet} --exp_name <your_name> ...

   To see all available flags run: python train.py --help

Reproducibility

We provide a set of hyperparameters used in our final runs. Some flags have common values for all datasets by default. For reproducibility of reported results run:

• HMNIST: python train.py --model_type gp-vae --data_type hmnist --exp_name reproduce_hmnist --seed $RANDOM --testing --banded_covar --latent_dim 256 --encoder_sizes=256,256 --decoder_sizes=256,256,256 --window_size 3 --sigma 1 --length_scale 2 --beta 0.8 --num_epochs 20
• SPRITES: python train.py --model_type gp-vae --data_type sprites --exp_name reproduce_sprites --seed $RANDOM --testing --banded_covar --latent_dim 256 --encoder_sizes=32,256,256 --decoder_sizes=256,256,256 --window_size 3 --sigma 1 --length_scale 2 --beta 0.1 --num_epochs 20
• Physionet: python train.py --model_type gp-vae --data_type physionet --exp_name reproduce_physionet --seed $RANDOM --testing --banded_covar --latent_dim 35 --encoder_sizes=128,128 --decoder_sizes=256,256 --window_size 24 --sigma 1.005 --length_scale 7 --beta 0.2 --num_epochs 40