Databases Reference
In-Depth Information
FIGURE 4.5: Growth rate of neurons in an HGN composition as a func-
tion of the number of different pattern elements and the pattern size.
©IEEE. Reprinted, with permission, from Amin, A.H.M.; Khan, A.I.; “A
divide-and-distribute approach to single-cycle learning HGN network for
pattern recognition,” 11th International Conference on Control Automa-
tion Robotics & Vision (ICARCV), 2010, pp.2118-2123, 7-10 Dec. 2010 doi:
10.1109/ICARCV.2010.5707852.
4.3.1 Distributed Neurons of HGN Network
According to Nasution [52], an HGN network can be decomposed into a
number of sub-compositions, according to the number of hosts available in
the physical network. Figure 4.6 shows a one-dimensional HGN composition
for a pattern of size 13, distributed onto four different hosts. Each neuron in
the composition is treated as a memory block on a host that is communicated
through an allocated terminal known as a port. In a computer system, a port
is used to establish communication channels between processes.
Each neuron in this network model is supplied with an additional param-
eter known as the port number. The port number identifies each neuron and
is used in inter-neuron communications. The communication between hosts
is achieved using physical communication, such as the Ethernet (using IP
address). Limitations of this approach include the following:
1. Additional parameter and indices. Each neuron in the composition needs
to acquire a unique port number, column index, row index, and ID. The
