Pseudo-visibility: A Game Mechanic Involving Willful Ignorance

FLAIRS-35 (2022)
Download Edit this record How to cite View on PhilPapers
Abstract
We present a game mechanic called pseudo-visibility for games inhabited by non-player characters (NPCs) driven by reinforcement learning (RL). NPCs are incentivized to pretend they cannot see pseudo-visible players: the training environment simulates an NPC to determine how the NPC would act if the pseudo-visible player were invisible, and penalizes the NPC for acting differently. NPCs are thereby trained to selectively ignore pseudo-visible players, except when they judge that the reaction penalty is an acceptable tradeoff (e.g., a guard might accept the penalty in order to protect a treasure because losing the treasure would hurt even more). We describe an RL agent transformation which allows RL agents that would not otherwise do so to perform some limited self-reflection to learn the training environments in question.
PhilPapers/Archive ID
ALEPAG
Upload history
Archival date: 2022-06-30
View other versions
Added to PP index
2022-06-30

Total views
54 ( #62,807 of 71,190 )

Recent downloads (6 months)
54 ( #14,752 of 71,190 )

How can I increase my downloads?

Downloads since first upload
This graph includes both downloads from PhilArchive and clicks on external links on PhilPapers.