1 paper
Even Klemsdal, Sverre Herland, Abdulmajid Murad
To increase autonomy in reinforcement learning, agents need to learn useful behaviours without reliance on manually designed reward functions. To that end, skill discovery methods…