UNSW CoFA Engineering FASS
iCinema
iCinema Projects
iCinema Infrastructure
Related Projects
Publications
Conferences & Seminars
Current Exhibitions
Papers
Postgraduate Research
Annual Reports
Vacancies
 
Scenario² (page 3 of 3)

Experimental Design

Scenario² tests how groups of agents acquire, interpret and translate gestural information into intentional behaviour through their interaction with human participants. It tests the limits of machine and human cognitive interaction. The objective is to test machine agent group perceptions and the constraints defining these perceptions. Anticipated is evidence of autonomous group activity by the agents, who collectively arrange themselves in responsive ways that can be recovered meaningfully.

The Scenario² experiment utilizes transpositional performance schemas, established by Beckett in Quadrat 1 + 2, 1981, an experimental television and theatre work that explores the problem of how group autonomy is defined by means of contending spatial interrelationships. Beckett's research is employed as it provides an appropriate aesthetic definition of group autonomy as other-intentional, that is, predicated on shared positional actions. In Quadrat, for example, characters mutually define each other by means of their respective territorial manoeuvres as they move backwards and forwards across the boundaries of a quadrant. Quadrat is drawn on as a way of aesthetically conceptualizing the relationship between spatialization and group consciousness. For example, in one sample experiment, participants are confronted with several machine agents who are trying to cluster themselves into a group. One of the agents takes on the role of herd master, circling his colleagues as he barks orders. Despairing over his failure, he moves towards the human participants in confusion. The participants in turn may attempt to assist by forming into a group. The agents react unpredictably, possibly mimicking, murmuring and then fleeing the scene.

The experiment generates a cascading series of gestural and clustering behaviours, testing and evaluating the network of meaningful deliberations and ascriptions of machine and human agents as they attempt to make sense of each others behaviours.


Funded by:
Australian Research Council Discovery Grant: Co-evolutionary narrative as machine autonomy in the relationship between artificial agents and human participants in the interactive cinema
Australian Research Council Queen Elizabeth II Fellowship Grant: Co-evolutionary narrative as machine autonomy in the relationship between artificial agents and human participants in the interactive cinema



page: