Multitask Representation Learning in Linear MDPs
The dataset is used for multitask representation learning in linear MDPs. It contains 80 different tasks, each with a different destination position, fire configuration, and action deviation probability p.
BibTex:
