Information Technology Reference
In-Depth Information
3.1
Implementation
We implemented the model on the neural networks simulator Leto/Prometheus. Leto/Pr-
ometheus simulates the dynamics of neural networks by an update of the whole network
at each time step. We use groups of neurons with one neuron, and non-modifiable links
between groups. The schema of Fig.4 show this implementation.
The internal states of agents,
S
i
, are relaxation oscillators: the re-entering link of
1
NV
Act
1
th
β
S
1
th
u
0
.
95
1
Agent1
−
1000
V
Act
1
Relax
1
σ
u
2
Recording
σ
NV
Act
2
th
β
S
2
th
0
.
95
Agent2
1
−
1000
Δφ
ini
Relax
2
−
1000
Fig. 4.
Implementation of the two agents. The couples (
S
1
;
Relax
1) and (
S
2
;
Relax
2) are relax-
ation oscillators. The parameters which will be tested are the following:
β
, the threshold which
controls the non-verbal production;
u
1
and
u
2
which control the agents' level of sharing;
Δφ
ini
,
the initial phase-shift between agents.
weight
1
makes the neuron behave as a capacity, and the
Relax
neuron which fires when
a
0
.
95
threshold is reached, inhibits
S
i
and makes it relax (see Fig.5 for an example of
the activation obtained).
V
Act
1
, Agent1's verbal production, is a neuron of constant activity
1
.Thisneuron
feeds the oscillators of both agents, weighted by their level of understanding
u
1
and
u
2
. The values of
u
1
and
u
2
are near
0
.
01
: it enables a well balanced sampling of the
oscillators' activations, the period last around
100
time steps.
In addition to agent understanding
u
1
and
u
2
, three other parameters are modifiable
in this implementation:
- The threshold
β
which controls the triggering of non-verbal signal.
- The sensitivity of agent's internal state to non-verbal signal
σ
which weights
NV
Act i
.
Fig. 5.
Activations of the internal state
S
1
(
t
) for
u
1
=0
.
01
