Reading as a god

Chapter 250 The Disappearing Gift Package
Zhang Shan knows that no amount of energy can be overstated in deep learning.

Paying attention to the field of artificial intelligence is in line with the trend of the times~
We are now in the third wave of artificial intelligence.

From the birth of artificial intelligence at the Dartmouth Conference in 1956, artificial intelligence has developed for 59 years.

During this period, artificial intelligence has experienced ups and downs, experienced several ups and downs, and each upsurge is because the proposal of core technology has aroused people's great interest and attracted a lot of capital investment.

The reason why AI is compelling is, simply put, that artificial intelligence scholars believe that the deep automatic encoding and decoding processes of computers and intelligent networks are similar to human thinking.

It is a process of deep learning from data characterization, abstract cognition to optimal solution.

Due to the deep structure of the human brain, the cognitive process is a complex brain activity process~
Therefore, the learning process of computer and artificial intelligence network simulation from symbol acceptance, symbol decoding, meaning establishment to optimization scheme is also structured;

At the same time, the cognitive process is carried out layer by layer and gradually abstracted. Artificial intelligence is not a product that relies purely on mathematical models.

It is the product of simulating the human brain, human brain neural network and abstract cognition and thinking process.

Exploring oneself has always been a human instinct, and through the study of artificial intelligence, human beings can recognize themselves better!
It should be said that so far, deep learning is the closest intelligent learning method for computers and intelligent networks to the human brain.

In recent years, deep learning has further attempted to directly solve the problem of abstract cognition, and has made breakthrough progress.

In April 2013, "MIT Technology Review" magazine listed deep learning as the first of the top ten breakthrough technologies in 4.

The revolution detonated by deep learning has brought artificial intelligence to a new level. It is not only of great academic significance, but also has strong practicality. The industry has also begun large-scale investment, and a large number of products will benefit from it.

……

The emergence of deep learning stems from more than 30 years of research in computer science, artificial neural networks, and artificial intelligence.

For more than 30 years, Professor Hinton (G.), Department of Computer Science, University of Toronto, Canada, has been engaged in research on machine learning models, neural networks and artificial intelligence, and has made breakthroughs in machine learning models, especially shallow learning models, to achieve A breakthrough has been made in computer abstract cognition.

In 2006, he published the article "Reducing the dimensionality of data with neural networks" in "Science", which discussed the learning model of applying artificial neural network to describe data, and first proposed deep learning (deep learning). learning) concept and computer deep learning model, set off a new upsurge of deep learning in the field of artificial intelligence.

Zhang Shan is fortunate to have read this article!
The two main points of this article are:

First, the multi-hidden layer artificial neural network has excellent feature learning ability, and the learned features have a more essential description of the data, which is conducive to visualization or classification.

Second, the deep neural network can effectively overcome the difficulty of training and optimal solution through "Layer-wise Pre-training". The unsupervised layer-wise initialization method helps to break through the shallow learning model.

Based on the deep belief network (DBN), an unsupervised layer-by-layer training algorithm is proposed, which brings hope to solve the optimization problems related to the deep structure, and then a multi-layer autoencoder deep structure is proposed.

It can be said that Sinton is both a pioneer and a founder in the field of deep learning.

In 2012, Hinton led his students to achieve amazing results on the classification problem on ImageNet, the largest image database at present, and greatly reduced the top five error rates (Top5 error rate) when computers deal with image data problems from 26%. To 15%, greatly improving the accuracy and clarity of artificial intelligence image data processing, which is a level that earlier computers only relied on surface learning and single-layer learning of mathematical models.

The reason why Zhang Shan has never been exposed to deep learning, but has such a thorough understanding of deep learning!

Because in the field of artificial intelligence, deep learning is actually a kind of algorithmic thinking
——Its core is the simulation of deep-level learning of human brain thinking. By simulating the deep-level abstract cognitive process of human brain, it realizes the complex calculation and optimization of data by computer.

The model used in deep learning is a deep neural network model, that is, a neural network that contains multiple hidden layers.Deep learning uses the hidden layer in the model to transform the original input layer by layer into shallow features, middle-level features, high-level features and finally the final task goal through feature combination.Deep learning can complete artificial intelligence tasks that require highly abstract features, such as speech recognition, image recognition and retrieval, natural language understanding, etc.The deep model is an artificial neural network with multiple hidden layers. The multi-layer nonlinear structure makes it have powerful feature expression ability and modeling ability for complex tasks.Training deep models has been a long-standing problem. In recent years, a series of methods represented by hierarchical and layer-by-layer initialization have been proposed, bringing hope to training deep models and achieving success in many application fields.

Since in the final analysis deep learning has to establish a relationship with algorithms, this is not a problem for Zhang Shan at all.

After the system is upgraded to level 100.

Zhang Shan, who has an extraordinary programming talent, is a hundred times more sensitive to algorithms than ordinary people.

Therefore, although the biometric mode of gait recognition is a new technology for Zhang Shan, Zhang Shan can still easily grasp the basic principles of this technology.

After mastering the basic principles and understanding of all its operating mechanisms, Zhang Shan will no longer be ashamed.

Although it will take quite a while to promote this technology to the market, there is no doubt that this mysterious technology gift pack is still very fragrant~
Zhang Shan suddenly remembered that he still had a gift bag unopened
"The host successfully upgraded the system from level 11 to level 12, and gave the host 5 million soft sister coins as a special gift, which is clean and safe to use; in addition, a special mysterious gift package is given to the host~"

The reason is because I didn't take it apart for the first time~
It's because this mysterious gift pack can only be opened on October 10~
However, because a series of upgrades and dormancy disrupted the situation, Zhang Shan only remembered this gift now!

"I'm sorry, host, you have missed the opening time of the gift pack, and the gift pack cannot be opened~"

Zhang Shan:? ?

Is there such a situation~
Thinking of the mysterious gift of possible measures, Zhang Shan is heartbroken~
At this time, I suddenly heard the system prompt:

"The gift package has expired, the host can choose to convert the gift package into other forms of gifts~"

Tsk, since it is no longer possible to open the gift bag, timely stop loss is the kingly way!

(End of this chapter)