This repository contains the experiment implementations for "Monotonic Transformation Invariant Multi-task Learning". This repository contains two subdirectories:
mtl/— Code used for the Illustrative Example and NYUv2 Multitask Learning Experiment.mtrl/— Code used for the Metaworld MT10 Multitask Reinforcement Learning Experiment.
Each of these subdirectories includes a separate README that covers how to set up and run the experiments.