SCOT: Set Cover Optimal Teaching

Set Cover Optimal Teaching for Sequential Decision Making with Inverse Reinforcement Learning

In this repository we recreate the SCOT machine teaching algorithm for custom grid environments, robust to stochastic environments with arbitrary numbers of 'colored' squares (features).

Usage

Install dependencies with Python3

pip3 install -r requirements.txt

To run initial test cases, including those outlined in (Brown and Niekum, 2019), run

python3 main.py

See algorithms folder for all our reinforcement learning algorithms.

References

• Machine Teaching for Inverse Reinforcement Learning: Algorithms and Applications