Keynote Speakers:
Topic: The Neural Mechanisms of
Value-Based Decision-Making
Over the course of the last 300 years, three principle groups of scholars have
attempted to understand how we make choices. Operating exclusively at the level
of observable behavior, economists have developed clear mathematical models
that place constraints on the representations of value decision-makers of
various types could, in principle, employ. Psychologists, working at the level
of mental states, have developed algorithmic models of the choice process.
Working at a more mechanistic level, neurobiologists have described the
information processing features of the primate brain and traced the flow of
information through that circuitry as decisions are made. Within the last 10
years those three traditions have begun to converge, and overlapping constraints
from each field have been used to identify many of the core feature
of the mechanism of choice. Data from both monkeys and humans now indicates
that frontal and basal ganglia circuits learn and store the values of an
animals past actions, a process that seems to be critically dependent on the
neurotransmitter dopamine. These stored values influence choice through a fronto-parietal network that takes as an input the values
of the options before a chooser and generates as an output a motor plan. A more
detailed outline of this system and some of the key experimental data serve to
identify it will be presented.
Topic: Moving Viewpoint: what makes human subjects different from computer agents?
Humans observe themselves and others from various
viewpoints. This fact has been mentioned in various fields. The results of the
Sally-Ann Test (Wimmer and Perner
1983) show that most children sufficiently old can make inferences from another
person's viewpoint. Adam Smith (1759) emphasized that humans take the impartial
spectator's viewpoint to judge what they themselves and others should do. The
Self-projection theory (Bucker and Carroll 2007) claims that humans move their
viewpoint freely in the time-space to make decisions. Jumping out of the system
(Hoffstadter 1979), or observing the system to which
one belongs from the outsider's viewpoint, plays an important role in human
thinking.
Such movement of human viewpoint, which is more often
unconscious than not, characterizes human thinking. It gives human systems such
dynamic stability/predictability and instability/unpredictability that could
not be seen in simulations with computer agents. I shall discuss it, referring
to my current behavioural and neuroeconomic
experiments of intertemporal preference. In relation
to this I may also mention economics education and the scientists' viewpoint
from which human behaviour is understood.
Topic: Agent
Dynamics in Complex Multilevel Systems of Systems of Systems
Social systems have many levels. Individual people form
dynamic subsystems of interacting agents at microlevels.
Nations and groups of nations form dynamic systems of interacting agents at macrolevels. Social groups, companies, organisations,
federations and institutions form dynamic systems of interacting agents at mesolevels. There are agent interactions between levels,
with emergent bottom-up and top-down dynamics. The social universe is formed
from systems of systems of systems with ever-changing boundaries. Networks are
used extensively to investigate agent interactions. Hypernetworks
provide a generalisation of networks that may provide
a coherent way of representing systems and their dynamics at many interacting
levels. This talk will sketch the theory of hypernetworks
and illustrate its application with examples from diverse areas including
multi-robot systems, education, city planning, and climate change.