@inproceedings{9679781909194b739a7f8839bfbf6b91,
title = "A definition of happiness for reinforcement learning agents",
abstract = "What is happiness for reinforcement learning agents? We seek a formal definition satisfying a list of desiderata. Our proposed definition of happiness is the temporal difference error, i.e. the difference between the value of the obtained reward and observation and the agent{\textquoteright}s expectation of this value. This definition satisfies most of our desiderata and is compatible with empirical research on humans. We state several implications and discuss examples.",
keywords = "Machine ethics, Optimism, Pleasure, Reward prediction error, Temporal difference error, Well-being",
author = "Mayank Daswani and Jan Leike",
note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2015.; 8th International Conference on Artificial General Intelligence, AGI 2015 ; Conference date: 22-07-2015 Through 25-07-2015",
year = "2015",
doi = "10.1007/978-3-319-21365-1_24",
language = "English",
isbn = "9783319213644",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "231--240",
editor = "Alexey Potapov and Jordi Bieger and Ben Goertzel",
booktitle = "Artificial General Intelligence - 8th International Conference, AGI 2015, Proceedings",
address = "Germany",
}