As the learning take place, connections are modified considering a defined error. There are several ways to solve the same problem with different weight connections.
In complement to Erik Cuevas' post, one such a solution in a dynamic (evolutive) context may be "Soft Competitive Learning" as suggested by Bernd Fritzke. In particular, "Growing Cell Structures" are derived from concepts like SOM, Neural Gas and Elastic Nets. Follow:
B. Fritzke, " Growing Cell Structures A Self-Organizing Network for Unsupervised and Supervised Learning ", 1994 - http://www.wi.hs-wismar.de/~cleve/vorl/projects/dm/ss13/SOTA/quellen/1-s2.0-0893608094900914-main.pdf
B. Fritzke, "Growing Grid - a self-organizing network with constant neighborhood range and adaptation strength", 1995 - https://booksc.org/book/6920674/b62794
B. Fritzke, " A Growing Neural Gas Network Learns Topologies ", 1995 - https://papers.nips.cc/paper/893-a-growing-neural-gas-network-learns-topologies.pdf
Doherty et al., " Hierarchical Growing Neural Gas ", 2005 - https://uhra.herts.ac.uk/bitstream/handle/2299/3968/901763.pdf?sequence=1&isAllowed=y
https://www.demogng.de/JavaPaper/node23.html
and more recently:
Lopez-Rubio et al., "Growing Hierarchical Probabilistic Self-Organizing Graphs", 2011 -
Before using ANN, it need a learning which is a process that produces an output that is as close as possible to the desired output by adjusting network parameters , there are two learning methods: Supervised and unsupervised.
Learning rules of ANN includes weights modifiable depending on the input it receives, its output value, and the associated teacher response.
when you start walking you will get problems, So start working first, if start facing difficulties, start learning to solving the challenges one by one . Here before supervising you have to learn yourself absolutely doing that work.