Shulou(Shulou.com)06/02 Report--

Produced by big data Digest

Compiled by Li Lei, Jian Jieqiong, Yunzhou

If you have opened a browser in the past few years, you must have seen the word "neural network" hundreds of times.

In this short article, I will give you a preliminary introduction to the background information in this field and the neural network itself. The next five minutes may not make you a world-class expert in this field quickly, but it is easy for you to go through a meaningful entry stage. In addition, you will learn some popular terms (especially for further reading according to the list at the end of the article), which can be your after-dinner conversation.

What is machine learning?

To understand neural networks, we first need to understand machine learning. In order to understand machine learning, we have to talk about human learning, or "classical programming".

In classic programming, I, the developer, need to understand all aspects of the problem I'm trying to solve and know exactly how to find a solution.

For example, suppose I want my program to know the difference between a square and a circle. One way to solve this problem is to write a program that can detect edges and corners. If my program detects four corners, then the shape is a square, and if it can't detect any corners, then the shape is a circle.

So what is machine learning? Generally speaking, machine learning = learning from examples.

In machine learning, when faced with the problem of distinguishing circles from squares, we will design a learning system that takes many examples of shapes and their categories (squares or circles) as input. In this way, we hope that the machine can learn the features that can be used to distinguish shapes.

Then, once the machine has learned all the features, we can give it an image that has never appeared before and hope it can classify it correctly.

What is a neuron?

In the context of neural networks, neuron is a wonderful term, and "smart" people do not want to use the simple name function. Of course, in the context of mathematics and computer science, function is also a wonderful name that receives input, implements processing logic, and outputs results.

More importantly, a neuron can be thought of as a learning unit.

Therefore, we need to understand what is a learning unit in the context of machine learning. Then we can understand the most basic structure of the neural network, that is, neurons.

To illustrate this, suppose I try to understand the relationship between the number of words in a blog post and the number of words people actually read from the post. Please note-we are doing this in the field of machine learning, so we have to learn from examples.

So I've collected a lot of examples of word count in my blog posts, expressed as x, and how many words people actually read in these posts, expressed as y, and I assume that there is a certain relationship between them, expressed as f.

The magic of this is that I just need to tell the machine (program) what the relationship I expect to see (such as a straight line) looks like, and the machine will understand the actual graphics it needs to draw.

So what do I get here?

Next time, I want to write a blog post with x words, and the machine can use its discovered relationship f to tell me that I can expect people to actually read the number of words, y.

So, the neural network is.

Well, if a neuron is a function, then a neural network is a function network! This means that we have many (many) such functions and learning units whose inputs and outputs are intertwined, that is, they are input and output to each other.

As the designer of this network, I need to answer the following questions:

How do I model input and output? (for example, if the input is some text, can I model it in letters? What about numbers and vectors? ... . )

What are the functions of each neuron? Are they linear? Or exponential? .)

What is the architecture of the network? (that is, the output of which function is the input of which function? )

What popular terms can I use to describe my network?

Once I have answered these questions, I can "show" many examples of correct input and output to my network, so that when I "show" it a new sample input that it has never seen before, it will know the correct output.

You can also visit the super-cool website of the Neural Network Playground to better understand what the process means.

Neural network playground:

Https://playground.tensorflow.org/

Neural Network-A never-ending story

Because the field is actually expanding, the amount of new content appearing every minute is impossible for anyone to track. (I wonder if one day we will be able to build an AI that can track human progress in the field of AI)

The first thing to know when entering this field is that no one knows everything. So you don't have to worry about being inferior to others, just stay curious. :)

Therefore, some excellent resources are recommended:

Gal Yona, one of the great bloggers in this field. Her posts range from core technical explanations to semi-philosophical comments.

Explanation of core technology:

Https://towardsdatascience.com/do-gans-really-model-the-true-data-distribution-or-are-they-just-cleverly-fooling-us-d08df69f25eb

To semi-philosophical comments:

Https://towardsdatascience.com/the-tale-of-1001-black-boxes-62d12b5886aa

Siraj Raval is a youtuber with lots of videos, from theoretical explanations to hands-on tutorials!

Theoretical explanation:

Https://www.youtube.com/watch?v=xRJCOz3AfYY

Hands-on tutorial:

Https://www.youtube.com/watch?v=pY9EwZ02sXU

Christopher Olah, a passionate and insightful researcher, has an awesome blog that includes posts ranging from basic concepts to in-depth learning.

From the basic concepts:

Http://colah.github.io/posts/2015-09-Visual-Information/

Deep learning:

Https://distill.pub/2017/feature-visualization/

Moving towards data Science is the largest medium-sized publication in the field, and no matter how many minutes or hours you have, you can go to their home page and explore everything from utilities to depth algorithms.

Moving towards data science:

Https://towardsdatascience.com/

Related reports:

Https://medium.freecodecamp.org/neural-networks-for-dummies-a-quick-intro-to-this-fascinating-field-795b1705104a

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.