Hierarchy of web development and neural networks?

Question

Neural networks

Hey!
I do not think that I correctly formed the question, so I will try to explain what I want to understand.
I'm currently working as a Frontend Jun. In the future, I want to develop cool web services using neural networks, machine learning and everything connected with it. Therefore, I want to understand something and listen to the opinion of other experienced people.
So, Web development in its classical form consists of:
- frontend;
-backend;
I suspect that, for example, neural networks will stand between front-end and back-end when viewed in the context of web development. Or is it generally a separate issue and neural networks cannot be thrust between anything? Which branch (frontend, backend) do you think neural networks will be closer to?

Answer 1

Backend
There is no movement towards the front At
all

Answer 2

In the classical understanding of a web service, the backend is the entire logic of the process, the frontend is primarily the rendering of the user interface.
Let's take a look at the example of a service for finding a celebrity that looks the most similar to you:
On the front-end side: a form for filling in your photo and rendering the results.
On the backend side: getting your photo and feeding it to a neural network trained on a large dataset in advance. At the output of the neural network - several options, ordered by proximity to your photo, they are given to the frontend for rendering.
Why is it impossible to give the neural network to the client side? First of all, because it can weigh a lot, secondly, because of the complexity of calculations, and thirdly, because of the risk of losing intellectual property, which is a trained neural network.
These arguments, however, are not unambiguous restrictions, you can always make a neural network in the frontend, but it will work at the expense of the client's resources.

Answer 3

This is a separate topic, the end result of which can be put on the web