Towards Cost Sensitive Decision Making

Usage

python scripts/run_agent.py --cfg_file=path/to/config --mode=train/test

Code Structure

.
|- scripts
|- src
|   |- agents
|   |   |- fully_observed_ppo: fully observed
|   |   |- concat_action_ppo: concatenate afa and tsk action space
|   |   |- concat_action_mbppo: with generative model
|   |   |- batch_hier_ppo: batch acquisition
|   |   |- batch_hier_mbppo: with generative model
|   |   |- seque_hier_ppo: sequential acquisition
|   |   |- seque_hier_mbppo: with generative model
|   |- environemnts
|   |   |- sepsis: sepsis simulator
|   |   |- episode_length_wrapper: limit episode length
|   |   |- concat_action_wrapper: concat action space
|   |   |- batch_acquire_env: batch acquisition
|   |   |- seque_acquire_env: sequential acquisition
|   |- models
|   |   |- poex_vae_cat_dis: POEx model for environments with categorical observations and discrete actions
|   |- networks
|   |- policies
|   |- utils
|- requirements.txt
|- README.md