Altruism and Selfishness in Believable Game Agents: Deep Reinforcement Learning in Modified Dictator Games
نویسندگان
چکیده
منابع مشابه
Non-reciprocal altruism in dictator games
We carry out a double blind dictator game experiment where the anonymous recipients are randomly drawn from the Swedish general population, and any donations are mailed to the recipients. About a third of the subjects donate some money. 2000 Elsevier Science S.A. All rights reserved.
متن کاملTaking, Giving, and Impure Altruism in Dictator Games
We show that, if giving is equivalent to not taking, impure altruism could account for List's (2007) finding that the payoff to recipients in a dictator game decreases when the dictator has the option to take. We examine behavior in dictator games with different taking options but equivalent final payoff possibilities. We find that the recipients tend to earn more as the amount the dictator mus...
متن کاملCooperation in Subset Team Games: Altruism and Selfishness
This paper extends the theory of subset team games, a generalization of cooperative game theory requiring a payoff function that is defined for all subsets of players. This subset utility is used to define both altruistic and selfish contributions of a player to the team. We investigate properties of these games, and analyze the implications of altruism and selfishness for general situations, f...
متن کاملAltruism and selfishness.
Many situations in human life present choices between (a) narrowly preferred particular alternatives and (b) narrowly less preferred (or aversive) particular alternatives that nevertheless form part of highly preferred abstract behavioral patterns. Such alternatives characterize problems of self-control. For example, at any given moment, a person may accept alcoholic drinks yet also prefer bein...
متن کاملDeep Reinforcement Learning for General Game Playing (Theory and Reinforcement)
We created a machine learning algorithm that, given the rules of a two-player adversarial game, outputs a function Q such that Q(s, a) approximates the expected utility of action a in state s. Our Q-function is a neural network trained using reinforcement learning across data generated by selfplay. We then create an agent for the game using the policy determined by the learned Q-function. Our a...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: IEEE Transactions on Games
سال: 2020
ISSN: 2475-1502,2475-1510
DOI: 10.1109/tg.2020.2989636