#### DMCA

## Ad hoc coordination in multiagent systems with applications to human-machine interaction (2013)

Venue: | In International Conference on Autonomous Agents and Multi-agent Systems (AAMAS |

Citations: | 1 - 0 self |

### Citations

786 | Games with Incomplete Information Played by Bayesian Players
- Harsanyi
- 1968
(Show Context)
Citation Context ...ormation relevant to its decision making. What relates this problem to ad hoc coordination is the fact that no player knows the private information, and hence behaviour, of any other player. Harsanyi =-=[9]-=- introduced Bayesian games in which the private information of a player is abstractly represented by its type, admitting a solution in the form of the Bayesian Nash equilibrium. However, while the not... |

281 | Rational Learning Leads to Nash Equilibrium,” Center for
- Kalai, Lehrer
- 1991
(Show Context)
Citation Context ...tting a solution in the form of the Bayesian Nash equilibrium. However, while the notion of private information is useful to describe ad hoc coordination, the learning processes studied therein (e.g. =-=[7,10]-=-) are not directly applicable, since the focus has traditionally been on equilibrium considerations but not on efficiency. This thesis is aimed at developing novel models and solutions for the ad hoc ... |

53 | Learning to play Bayesian games,”
- Dekel, Fudenberg, et al.
- 2004
(Show Context)
Citation Context ...tting a solution in the form of the Bayesian Nash equilibrium. However, while the notion of private information is useful to describe ad hoc coordination, the learning processes studied therein (e.g. =-=[7,10]-=-) are not directly applicable, since the focus has traditionally been on equilibrium considerations but not on efficiency. This thesis is aimed at developing novel models and solutions for the ad hoc ... |

48 | Ad hoc autonomous agent teams: Collaboration without pre-coordination
- Stone, Kaminka, et al.
- 2010
(Show Context)
Citation Context ...prior description of the human’s behaviour (no prior coordination). Ad hoc coordination is a problem which has been recognised as important only recently, hence prior work is relatively limited, e.g. =-=[6,8,11]-=-. However, the assumptions made by the solutions proposed therein imply that they only address certain aspects of the larger problem. For example, in [6,8] it is assumed that all agents follow complex... |

23 | To teach or not to teach?: decision making under uncertainty in ad hoc teams
- Stone, Kraus
- 2010
(Show Context)
Citation Context ...ertain aspects of the larger problem. For example, in [6,8] it is assumed that all agents follow complex pre-specified plans which define roles and synchronised action sequences for each role, and in =-=[1,12,13]-=- it is assumed that the other agents’ behaviours are a priori known and fixed (i.e. they do not learn), and that all agents, including the ad hoc agent, have common payoffs. Furthermore, the problem d... |

21 |
Coordination and adaptation in impromptu teams.
- Bowling, McCracken
- 2005
(Show Context)
Citation Context ...prior description of the human’s behaviour (no prior coordination). Ad hoc coordination is a problem which has been recognised as important only recently, hence prior work is relatively limited, e.g. =-=[6,8,11]-=-. However, the assumptions made by the solutions proposed therein imply that they only address certain aspects of the larger problem. For example, in [6,8] it is assumed that all agents follow complex... |

18 | Leading Ad Hoc Agents in Joint Action Settings with Multiple Teammates.
- Agmon, Stone
- 2012
(Show Context)
Citation Context ...ertain aspects of the larger problem. For example, in [6,8] it is assumed that all agents follow complex pre-specified plans which define roles and synchronised action sequences for each role, and in =-=[1,12,13]-=- it is assumed that the other agents’ behaviours are a priori known and fixed (i.e. they do not learn), and that all agents, including the ad hoc agent, have common payoffs. Furthermore, the problem d... |

11 | Leading a best-response teammate in an ad hoc team.
- Stone, Kaminka, et al.
- 2010
(Show Context)
Citation Context ...ertain aspects of the larger problem. For example, in [6,8] it is assumed that all agents follow complex pre-specified plans which define roles and synchronised action sequences for each role, and in =-=[1,12,13]-=- it is assumed that the other agents’ behaviours are a priori known and fixed (i.e. they do not learn), and that all agents, including the ad hoc agent, have common payoffs. Furthermore, the problem d... |

10 | Dynamically formed human-robot teams performing coordinated tasks.
- Dias, Harris, et al.
- 2006
(Show Context)
Citation Context ...prior description of the human’s behaviour (no prior coordination). Ad hoc coordination is a problem which has been recognised as important only recently, hence prior work is relatively limited, e.g. =-=[6,8,11]-=-. However, the assumptions made by the solutions proposed therein imply that they only address certain aspects of the larger problem. For example, in [6,8] it is assumed that all agents follow complex... |

5 | Comparative evaluation of MAL algorithms in a diverse set of ad hoc team problems.
- Albrecht, Ramamoorthy
- 2012
(Show Context)
Citation Context ...pared five multiagent learning algorithms in a comprehensive set of ad hoc coordination problems, using criteria such as convergence rate, social welfare and fairness, 1415 and equilibrium attainment =-=[3]-=-. The compared algorithms, while representing major approaches in the field (e.g. opponent modelling, policy hill-climbing, and regret-minimisation) were originally developed in a heterogeneous settin... |

4 | A game-theoretic model and best-response learning method for ad hoc coordination in multiagent systems (extended abstract
- Albrecht, Ramamoorthy
- 2013
(Show Context)
Citation Context ...n problem. I conceptualise this problem formally using a gametheoretic model, called the stochastic Bayesian game, in which the behaviour of a player is determined by its private information, or type =-=[5]-=-. Based on this model, I derive a solution, called Harsanyi-Bellman Ad Hoc Coordination (HBA), which utilises a set of user-defined types in a planning procedure to find optimal actions in the sense o... |

1 |
Analysis and correction of the data communication model of a continuous-time distributed application with emphasis on state consistency and fault tolerance (German
- Albrecht
- 2010
(Show Context)
Citation Context ...em A long-term goal of this thesis is to study ad hoc coordination problems in complex multiagent systems which involve humans. To this end, I plan to use the internet-based realtime game Saga Online =-=[2]-=-. Therein, the players choose to be “Warriors” or “Magicians” and have to defeat each other using a variety of learnable skills. This game has several useful advantages, such as intuitive user interfa... |

1 |
Exploiting causality in the monitoring task of
- Albrecht, Ramamoorthy
- 2013
(Show Context)
Citation Context ... of the variables that directly affect it change. This insight led to the formulation of a new algorithm, called Passivity-based Monitoring, which exploits passivity to accelerate the monitoring task =-=[4]-=-. The idea is to maintain beliefs over individual aspects of the system, and to perform selective updates over these beliefs by exploiting passivity. I show empirically, in both synthetic systems and ... |