NARCA: NARS Controlled Agent

An agent capable of playing various games in Gym environments, using NARS for planning

Installation

You need to have OpenNARS-For-Applications installed somewhere. It is expected that you will export a NARS_HOME environment variable to specify where it is built.

Dependencies are listed in the env.yml file, that can be used to create a Conda environment. lock.yml is a lock-file that you can use to make a reproducible environment, Conda might fail to create it though. For standard virtual environments you just need to use pip. You can also use Poetry, via poetry install, poetry shell will open a shell with the virtual environment activated.

Griddly requires to install Vulkan SDK, refer to its README.

Install in editable mode to your environment: pip install -e .

Running

There are multiple experiments under the experiments directory. Each one can be launched with python <path to the script>. You can observe the rendered environment and run TensorBoard to see performance metrics. TensorBoard run data gets saved to runs. There is also Neptune.ai integration that you can use, you need to set NEPTUNE_PROJECT and NEPTUNE_API_TOKEN.

Each Griddly environment has its own subdirectory, each has at least the main experiment - this is supposed to hold the currently established simplest best performing agent.

Notable experiments

Besides the usual main experiments there are some interesting variations as follows:

• experiments/zelda/procgen: uses a level generator that places a key and a door randomly around the level.
• experiments/drunk_dwarf/curriculum_learning: tries to find out whether there is a measurable gain in learning efficiency with a curriculum learning setup.