CausalCity: Complex Simulations with Agency for Causal Discovery and Reasoning

Anonymous NeurIPS Submission

Simulation Download

The compiled binary of the simulation is available here (size - 800MB): DOWNLOAD


The python API for AirSim provides the binding for launching the scenarios and logging data. Our simulation uses a custom version of the python bindings that can be downloaded here. Download and unzip this file and then install the airsim library as below.


Install airsim python bindings:

>>> python install

Launch the simulation and a python terminal:

>>> import airsim

Set up car client:

>>> c = airsim.client.CarClient()

Run a scenario:

>>> c.simRunConsoleCommand('ce RunScenario Scenario_1')

Get RGB image and segmentation mask:

>>> segResponse = c.simGetImages([airsim.ImageRequest("0", airsim.ImageType.Scene, False, False), airsim.ImageRequest("0", airsim.ImageType.Segmentation, False, False)])

A more detailed script used to generate the scenarios used in our submission can be found here.