@inproceedings{aae6fe95db0143beb692b47d53710fc4,
title = "Shielding atari games with bounded prescience",
abstract = "We present the first explicit-state method for analysing and ensuring the safety of DRL agents for Atari games. Our method only requires access to the emulator. We give a suite of 42 properties that characterise {"}safe behaviour{"} for 31 games. We evaluate the safety of the best available DRL agents which, as our experiments show, violate most of our properties. We propose a countermeasure that implements shielding using bounded explicit-state exploration. Our method improved their overall safety, producing the safest DRL agents for Atari games currently available.",
keywords = "Safe AI, Deep Reinforcement Learning, Atari Games",
author = "M. Giacobbe and Mohammadhosein Hasanbeig and Daniel Kroening and Hjalmar Wijk",
year = "2021",
month = may,
day = "3",
language = "English",
series = "AAMAS Conference proceedings",
publisher = "International Foundation for Autonomous Agents and Multiagent Systems",
pages = "1507–1509",
booktitle = "AAMAS '21",
address = "United Kingdom",
note = "AAMAS '21 : 20th International Conference on Autonomous Agents and Multiagent Systems ; Conference date: 03-05-2021 Through 07-05-2021",
}