Sept. 27, 2018, 10:15 a.m.
[COG Seminar] Gleb Polevoy: "The Game of Reciprocation Habits" and "Reciprocation Effort Games"
Two presentations by Gleb Polevoy.
Title (1): The Game of Reciprocation Habits
Authors (1): Gleb Polevoy, Mathijs de Weerdt, Catholijn Jonker
Title (2): Reciprocation Effort Games
Authors (2): Gleb Polevoy, Mathijs de Weerdt
Speaker: Gleb Polevoy
Time and place: Thursday, 27-th September 2018, 10:15 am, room 310.
Article (1): https://doi.org/10.3233/978-1-61499-672-9-417
Article (2): https://doi.org/10.1007/978-3-319-76892-2_4
People often act on reciprocal habits, almost automatically responding
to others' actions. A robot who interacts with humans may also reciprocate, in order to
come across natural and to be predictable. We aim to facilitate a decision support system that
advises on utility-efficient habits in these ubiquitous interactions.
To this end, given a model for reciprocation behavior
with parameters that represent habits,
we define a game that describes which habit one should adopt
to increase the utility of the process. The used model
specifies an agent's
action as a weighted combination of the others' previous actions (reacting) and either
i) her innate kindness, or ii) her own previous action (inertia).
We analyze reciprocation attitude change only for a pairwise interaction,
and the coefficient change for any number of agents.
For the case of two agents,
to analyze what happens when everyone reciprocates rationally,
we define a game where an agent may choose her habit, which is either her
reciprocation attitude (i or ii), or
both her reciprocation attitude and weight.
For a general connected network, when all agents have attitude ii),
we define a game where an agent chooses her weights.
We characterize the Nash equilibria of these games and consider their efficiency.
We find that the less kind agents should adjust to the kinder agents
to improve both their own utility as well as the social welfare.
This constitutes advice on improving cooperation and explains
real life phenomena in human interaction, such
as the societal benefits from adopting the behavior of the kindest person, or
becoming more polite as one grows up.
Consider people dividing
their time and effort between friends, interest clubs, and reading seminars.
These are all reciprocal interactions,
and the reciprocal processes determine the utilities
of the agents from these interactions. To advise on efficient effort division,
we determine the existence and efficiency of the Nash equilibria of the game of
allocating effort to such projects. When no minimum effort
is required to receive reciprocation, an equilibrium always exists, and if
acting is either easy to everyone, or hard to everyone, then every equilibrium
is socially optimal. If a minimal effort is needed to participate,
we prove that not contributing at all is an equilibrium, and for two
agents, also a socially optimal equilibrium can be found.
Next, we extend the model, assuming that the need to react
requires more than the agents can contribute to acting, rendering
the reciprocation imperfect.
We prove that even then, each interaction converges
and the corresponding game has an equilibrium.