The game concludes with the protagonist fully accepting his pathetic existence. Kisaragi pats him on the head like a dog, mocking him one last time as she walks away to spend his money, leaving him broken, exhausted, and desperately waiting for her next summons.
Both policies are parametrised by deep neural networks (see Fig. 1). The SP receives the same raw observation as the TC but produces a compact waypoint using a goal‑conditioned encoder . rj01322897
| Baseline | Description | |----------|-------------| | | End‑to‑end DDPG on raw sensor inputs ( The game concludes with the protagonist fully accepting