Neural Networks and Artificial Intelligence in PHP?

Asked

Viewed 4,323 times

2

I’m interested in creating smart software in PHP, but I don’t know where to start, for there is little material about it.

What neural networks and artificial intelligence have in common or how they work together?

Where to start developing systems with AI in PHP? Some example (code) to take as a starting point?

  • Neural Networks is a sub-topic of Artificial Intelligence, there are several subjects on the subject, books (in by Portuguese and in English), tutorials on the internet and etc, but it is very vast.... Although I find the subject interesting I believe that this question is very wide.... I welcomed the question because I like the subject, and would find it interesting if someone were willing to write a text here rs, but I find it difficult...

  • I would start looking for books on the subject, and then frameworks in php, for AI you should keep in mind tree and tree search methods, find the best solution to a problem using some "smart" criteria, neural networks is the opposite, the algorithm "learns" after a careful teaching through hundreds, if not thousands of iterations the best way to a solution

  • One more thing... have you thought about JS? To develop a neural network, I THINK (with a huge think) that PHP is very robust... a server-side Javascript, using Node.js could be more efficient and intuitive, there are very good tools for creating an RNA (see http://synaptic.juancazala.com/#/)but with PHP is not impossible, but would not be MY first choice.

  • Hello. As you have already commented, the subject is very broad, so a response that covers well what you want would be excessively long. Try to study the subject a little more, and come back here with more objective questions that you will always be welcome. : ) Anyway, in practice you can implement an RNE in any language, but it is necessary to check whether this is interesting or not (for example, someone suggested using JS, which may not be ideal because the most computationally expensive is the training of the network, not its execution).

  • About what these concepts have in common, read more here, in the other questions tag [tag:artificial intelligence]. Finally, on where to start, I suggest you study the technology before focusing on implementation. If you understand how Rnes work, you can implement something in PHP. Just then, look for something already ready. If PHP is not really a requirement, I would suggest using Python with the fantastic library scikit-Learn (that has muuuiiiitttaaa thing ready).

  • I really know that the question is a little broad, but I ran the risk of being negative or closed, because it is better to know directly from people who already know about it than to read articles on the internet from people who really do not know about it. About the choice of PHP was because it is that language "xodó", but I’ve worked with Python, I’ll take a look. I’ll try to continue the search for material, although it’s a bit difficult.

Show 1 more comment

1 answer

3

Sorry about the formatting, I’m a little rushed, but follow a few tips.

"Firstly, the paradigm of neurocomputing is motivated by cognitive abilities of biological neural networks, inspired by knowledge of neuroscience. The foundations of RNA are the model of a neuron, network topology and learning paradigms. The Mcculloch and Pitts model serves as a basic model of a neuron artificial.

Artificial Neuraus Networks are very useful in specialist systems, i.e., replicate and repeat procedures." ¹ (RAUBER, T. W.)

For the beginning of studies I want to try to understand a little bit about how the human brain works and how cognitive analysis works. It would be interesting also mathematical and physical knowledge because many equations are complex.

One of the preconceived languages for AI is Prolog that can help with predicative programming. According to Wikipedia:

"Prolog is a programming language that fits the paradigm of Programming in Mathematical Logic."

NOTE: Remember that Neural Networking is an architecture specialist! Applying it in an environment with multiple functionality can give work and is indicated the use of logic Fuzzy.

"RNA has numerous algorithms for pattern recognition: Kohonen, Perceptron, Adaline, Backpropagation and many others, each one with its specificity. The main advantage of using the Backpropagation is that it works with multi-layers and solves "nonlinearly separable" problems and some algorithms do not resolve.

In short, a "non-linearly separable" problem is one where we can separate 2 distinct classes on the two-dimensional Cartesian axis just tracing a straight line." ² (DEV Media)

A widely used form of the neural network is Backpropagation Backpropagation - DEV Media

Follow a link from an introductory presentation found on the internet, excuse me if you already know the same. RNA presentation

¹-RAUBER, T. W. Artificial neural networks. Federal University of Espírito Santo, 2005.

² - http://www.devmedia.com.br/redes-neurais-artificiais-algoritmo-backpropagation/28559 Accessed on 03/06/2016

Browser other questions tagged

You are not signed in. Login or sign up in order to post.