How to write a neural network for a family tree?

Question

ruby

Hello!
The university received such a task. Make a family tree. Initial data 2 names, output - related affiliation (for example, grandmother-grandson).
You need to do this using neural networks on any PL, I think using Ruby or Python.
Never dealt with NS before. Tell me which way to think, where to look and where to start.
Thanks in advance.

Answer 1

The problem is not that they didn’t deal with neural networks, but that they don’t know how to formulate the task in such a way that it could be solved using machine learning methods. I would recommend some introductory ML course (on coursera or udacity?), with or without neural networks. If you understand how to formulate the task (what is at the input, what is at the output - what needs to be found), then it will become obvious how to use the neural network (or any other classifier). If the task does not need to write code for the neural network, you can take a ready-made implementation, if necessary - well, you can write it; I understand that this is not the problem.

Answer 2

alife.narod.ru/lectures/neural/Neu_ch01.htm - start here, for example. Seems like a good manual.
There are several examples of neural networks on Habré, though not the most successful ones, IMHO.

Answer 3

It seems to me that the task is somewhat complicated by the term "neural network". I think the tournament option is quite suitable for you: https://github.com/balamut108/tournament_generator
I would be grateful if you contribute.