My technology affects the whole world

Zheng Shijian and others were so excited that they plunged into deep learning, imagining that they could also make the same product as Ma Tian, ​​but deep learning is not that simple.

Deep learning essentially builds a machine learning architecture model with multiple hidden layers, and obtains a large amount of more representative feature information through large-scale training data. In this way, samples can be classified and predicted, and the accuracy of classification and prediction can be improved.

The first layer is one or even several feature algorithms. Without sufficient mathematical knowledge, the architecture cannot be built, let alone subsequent model training.

Zheng Shijian and others worked several overnight classes before they finally understood the deep learning method mentioned in Kilfry's paper.

Later, I started to apply it according to the paper and train the model. When I saw the final result, the accuracy was only 16%, which made me stupid.

Could it be that Ma Tian lied to us? Tang Jiefei raised his own doubts.

Others nodded in agreement.

But Zheng Shijian objected: Impossible. I watched their employees pass the facial recognition gate made by Ma Tian, ​​and the accuracy rate is indeed frightening. Anyway, I didn't see any recognition errors.

Also, there is no need for Ma Tian to lie to us, because he can say nothing at all.

Everyone was silent. Zheng Shi saw that what he said made sense, and Ma Tian could have said nothing. As for deliberately pointing out mistakes to waste their time, this is not necessary for Ma Tian.

Could it be that Ma Tian has studied a set of deep learning algorithms by himself? One person put forward his own opinion. He analyzed, Ma Tian can publish mathematics articles on Nature, and people in the industry also recognize that Ma Tian is good at mathematics, which shows that Ma Tian’s mathematics is indeed extraordinary, and it is not impossible for him to write his own algorithms for deep learning!”

That makes sense!

I agree with this!

Others nodded, agreeing with this point of view.

Zheng Shijian was silent, because he imagined a possibility: Did Ma Tian predict that they would not be able to do it, so he directly told him about the path of deep learning?

Let's find a way to contact domestic mathematicians related to deep learning and see if we can ask them to take a look! Zheng Shijian finally decided. Now that their results are so poor, they can only ask those mathematicians for advice.

Although he thinks the effect is not great, after all, he also has a PhD in mathematics and computer science from a well-known foreign university.

A few days later, Zheng Shijian and his team members asked several domestic mathematicians studying deep learning, and the answers they received were all shaking their heads.

Some mathematicians even questioned them: You said that Ma Tian made a face recognition algorithm with a recognition rate of 99%? Are you lying to me?

Seeing that Zheng Shi was not talking nonsense, he waved his hand and asked his subordinates to show him the video posted by Ma Tian's employee on the Internet. The math family member was dumbfounded on the spot.

I can't do it! After watching the video, the mathematician directly admitted that he couldn't do it. Mathematicians have nothing to lie about. If they can't do it, they can't do it. If they can't do it, they can't do it!

After several days of asking for help across the country, no one responded. Zheng Shi saw that he was also smoking in the office and sighing.

Although he had been mentally prepared for it, he still couldn't help but feel decadent when the outcome was what he expected.

Tang Jiefei suddenly said: How about asking Ma Tian to help us?

Zheng Shi didn't say anything, but another colleague interrupted and asked: Brother Fei, have you never understood Ma Tiantian?

I really haven't studied it carefully! Tang Jiefei nodded. He is just a programmer who studies with his eyes closed. He has no time to care about things on the Internet.

Brother Fei, Ma Tian is not only a mathematician, he is also the founder of Little Blue Car and is worth billions! And he is also famous for being crazy and is called a horse madman!

Tang Jiefei was silent when he heard this. He was stupid and actually asked Ma Tianlai to help.

But you can't just sit there and do nothing. Just keep doing it. I don't believe it. 19-year-old Ma Tian can research it, but we can't? Lin Zhanwei from the team threw the information on the table. Not convinced. Apparently he was hit and his mentality was a bit unbalanced.

Do you think this is hard work? Just do it if you can do it? If you can't do math, you can't! Some people disagreed with Lin Zhanwei's words, but agreed with Tang Jiefei that it would be best for Ma Tian to come to help, or to ask Ma Tian for advice. That’s fine, it can save a lot of effort.

Didn't I say that I couldn't hire Ma Tian? Besides, why should Ma Tian help us? Can we pay the consulting fee to impress this billionaire?

What if you don't try?

…

Okay, stop arguing!

Seeing the team members quarreling, Zheng Shijian had to call out organizational discipline: Study for another week. If there is no direction after one week, I will go find Ma Tian!

Then a week later, Ma Tian met Zheng Shijian and Tang Jiefei in the office.

Mr. Ma, I'm here to ask for your help today! Zheng Shijian said with a blushing face.

Because Ma Tian was too young, he was a little embarrassed to have him, who was nearly 35 years old and also a boss, come down to ask for help.

Ma Tian looked at Zheng Shi in front of him and saw the two people, and roughly guessed their intentions.

Tell me what's going on first! Ma Tian said in a calm tone.

Mr. Ma, the main issue is about face recognition. We have also understood and tried the deep learning you mentioned last time, but it did not achieve the same effect as you! Zheng Shijian completely put down his face and asked for advice directly, So this time, we mainly want you to see what went wrong with us?

Ma Tian smiled. Sure enough, things were as he expected. Even if Zheng Shijian was told about the path of deep learning, they would not be able to figure it out in a short time.

Can you first tell me how many layers your deep learning architecture has? Ma Tian asked.

We currently have five layers as written in the paper, and we optimize and process the images layer by layer to obtain feature value transfer. Zheng Shijian answered truthfully.

Level 5? It's strange that there are so few with a high success rate! If I guess correctly, you didn't process the original data that couldn't be separated, right?

Don’t use traditional neural network methods for deep learning. You can flexibly change the writing algorithm to increase the level, as long as the final time complexity is ideal...

Ma Tianshu spoke out loudly about the problems of Zheng Shijian and others.

But Zheng Shijian and others didn't even understand the level of neural network algorithms, and they couldn't understand what Ma Tian said just like a scumbag listening to a teacher's lecture.

At this moment, Zheng Shijian and Tang Jiefei, who didn't understand, felt like they were really bad students as they listened to Ma Tian talk about a bunch of mathematical formulas and theorems.

All they needed was a notebook and a pen to take notes, but as a teacher, Ma Tian didn't have a blackboard, so they couldn't even take notes.