1 paper
Darius Muglich, Luisa Zintgraf, Christian Schroeder de Witt +2
Self-play is a common paradigm for constructing solutions in Markov games that can yield optimal policies in collaborative settings. However, these policies often adopt highly-spec…