Yes, that was Igor; but, hopefully, hasn't changed
much since!
Yes, that was Igor; but, hopefully, hasn't changed
much since!
Presents a basis for neuro (C library for manipulating with neural networks): neuro.h (rename to 'neuro.h' to work properly) and neuro.c (rename to 'neuro.c' to work properly). If you would like to use the neuro library, you can mail to me for sample applications (neural models) that would enable you easyly to use it. Since neural library has been developed parallel with my work it has been evaluated by updating some approaches together with writing new functions and descriptors. I was lazy and short with time to tidy up all that changes immediately as they have appeared and therefore the library contains some garbage (doesn't influence other stuff but just increases the size). Although the header (neuro.h) has been cleaned (not as much as I'd like to), I'm afraid the library will require some more pruning before the first release. The header file is intended as an information on the library as it consists used data structures (at the beginning of the document), types of descriptors defined in the library, available neuron models and learning algorithms. A neural network is defined by defining its genetic structure and internal connections of neurons that are neural networks for themselves. Neural networks are NOT simulated using matrices, but are rather presented using LIST/THREE structures. The main idea is that the same functions can be used (although some should be slightly changed) for implementing neural networks on multiprocessor systems (or even on a group of computers connected through a network) with different level of parallelism available, from 1 node - 1 processor to 1 neural system - 1 processor/computer. The library is fully based on the fuzzy one and a neural network is described as a set of fuzzy sets connected through inference engines (used functions) and with "fuzzy rules" stored in a way of connection weights.
.
Then she moves into her office (6/96) 
.
After closing office she moves into the cot (6/96) 
.
And that's not all so far.
Mia is grown up, or at least a bit older, and she remains very busy.
She has her little friends she takes a care of for a whole day. Carrying them through the house (1-3/98) 
. Helping
them to have a rest on the stairways (1-3/98) 
.
And cradling them to fall asleep (1-3/98) 
.
While
they are sleeping she can dress her new shoes (1-3/98) 
.
and get out in the park (Western Springs, Auckland, NZ) for a walk
(1-3/98) 
. Of course, she can't miss her
sunglasses!
Now is the third millennium. Mia moved to Australia and graduated the pre-school (1/2001).
.
So, when she is not at school, she can enjoy the playground in Wollongong Safe Harbour (2/2002).
.
At the begin of year 2000, I moved to Nortel Networks' R&D TC
Wollongong, Australia, and have been working on CDMA development since.
I'm a part-time PhD student at Computer
Science Department, University
of Auckland, Auckland, New Zealand. My thesis is from fuzzy set theory
and neural networks domain with application on vision. Supervisors are
Alan Creak and Clark
Thomborson.