Chapter 136 Tickets Got (Part 2)
In fact, Tong Yongshan did not fully understand any of the three papers that Fang Yu talked about.
After all, a considerable part of it belongs to very professional computer science, which is related to traditional applied mathematics, but not so closely that it requires special study.
Although he didn't understand it, it didn't affect Tong Yongshan's ability to discover its value.
Because scientific thinking methods are universal.
As a top scholar, Tong Yongshan has very strong scientific literacy.
The so-called scientific literacy means that even if what the other party says has long deviated from your professional field, you can still judge whether what the other party says is reliable or not based on the other party's proof ideas, argumentation methods, and choice of evidence.
This sophomore finance student in front of me has no problem with the application and demonstration of the three papers!
After listening to Fang Yu's report, Tong Yongshan had already keenly realized that Fang Yu must have had a purpose in choosing these three documents!
Yes, these three papers appear to be very influential papers on neural networks and deep learning, but Fang Yu just made a conscious effort to verify and draw on the contents of the three papers in his report.
On the surface, Fang Yu's final conclusion is only related to the third paper, but the first two papers played a key basic role in it.
For example, using the pre-training method of deep belief networks to optimize the training of GANs, or using the latent variable model of variational autoencoders to improve the factor model.
This shows that Fang Yu should have at least a preliminary research direction for the application of GANs in strong factor models!
Anyone who has done scientific research knows what is the most important thing for any scientific research project?
Research direction is the most important!
Among a thousand topics, nine hundred may go in the wrong direction, and among the one hundred that go in the right direction, ninety may fail to allocate their physical strength properly and run out of energy halfway.
In the end, there may be only ten people who take the right path and have the ability to reach Rome.
However, there was indeed a very small chance that one of the 900 people who took the wrong path did not reach Rome, but as they walked, they actually arrived in Chang'an.
For example, the main rival of the Fiery Sun Essence Liquid, which has not yet been launched on the market, is the Blue Pill. As we all know, it was originally used to treat cardiovascular and cerebrovascular diseases when it was developed.
But the proportion of such examples in the overall scientific research projects is really too small.
Because of this, when Tong Yongshan realized that Fang Yu might have indeed found a way, giving Fang Yu the initiative for a period of time became a natural choice.
Of course, this is also related to Tong Yongshan’s academic ethics.
If you were to have a cruel and ruthless tutor who only gave you a second work, you would still have to be grateful and drive to his house every day to pick up and drop off his children, and also give him the funds that were credited to his card.
Otherwise, don't even think about graduating.
"Teacher, can I ask what the budget is?" Fang Yu blinked and delivered a soul-stirring blow.
"Uh." Tong Yongshan stuttered for once.
It’s not that the National School of Economics has no funds, but he hasn’t submitted the proposal report for this project yet.
Without a proposal, there is naturally no budget to talk about.
Even if the proposal is approved, it is impossible to use only the national funding.
The money given by the national university is only enough to buy office supplies and provide subsidies and benefits to graduate students.
The bulk of the revenue still has to come from enterprises.
Especially this kind of research related to computer big data. Although Tong Yongshan had never done it before, he also knew that doing research on this kind of thing would be very expensive!
Apart from anything else, deep computing and neural networks are now all done with GPUs, or graphics cards. Buying ten of the latest TitanXs will cost you 100,000 yuan. Add in the motherboard, memory, SSD, and other things, and two or three servers will cost at least 200,000 yuan.
This does not include the cost of purchasing materials, which is another huge expense.
Then, with a system, we need a professional system administrator, right? If not, who will maintain the system with only the few talents in the team?
Hiring a system administrator would cost another 200,000 yuan a year.
Cloud computing services cost at least another 100,000.
Web services, journal subscriptions, dataset purchases, software licenses, development tools and libraries—all of these cost money.
Electricity bills and the like are not included.
After taking all the miscellaneous factors into account, this project cannot even complete the first phase without one million.
Of course, some of the money can be saved. For example, if you hire a junior or senior student to be a system administrator, you might not even need to spend two thousand yuan a month.
But why save?
Tong Yongshan had been in Maiguo for a long time and had no such awareness, so he thought it was quite normal.
"Teacher, if you don't have an accurate budget yet, I have an idea. There's a company I'm quite familiar with that's looking to get involved in this field. I can try to contact them and see if they can get some sponsorship."
Fang Yu's tone was very flat.
When Zhang Chang heard this, he thought, well, it turns out that this person is not only academically strong, but can also bring in funding to join the group.
The faint thought of striving for the leadership of the project disappeared immediately.
"Sponsor?" Even though Tong Yongshan was ignorant of business when he was young, he had become an old man after so many years of experience. When he heard these two words, he felt something was wrong.
I'm old now, so there won't be any female sponsors taking advantage of me, right?
Tong Yongshan remained calm: "What are the conditions?"
Fang Yu thought for a moment and said, "They may have a need for copyright and potential patents, but they probably won't interfere with academic publications."
Copyright?
Tong Yongshan pondered for a moment.
"Of course, if the school's budget approval process is almost complete, just ignore what I said."
"But when I was looking for papers these days, I found that several teams may have already started doing research in this area, such as Ma Baolong from Western University and Derek King from the Krant Institute of Mathematical Sciences at New York University."
"Derek King just published a study on arXiv last month that I think is a pioneering study on this topic. I've sent the document to your email. You can take a look."
Fang Yu pointed to his computer screen.
Tong Yongshan unlocked the screen, put on his reading glasses, opened Fang Yu's email, and became more and more frightened the more he read.
Derek King's article only explains the possible impact of neural networks on nonlinear data processing in strong factor models.
Seemingly unrelated.
But if you just take a few more steps down, you will reach what you want to study!
By then, it will be another academic competition!
This kind of academic competition is very common in the field of natural sciences, but it is not so common in econometrics.
The industry circle is relatively small, and there are not many bigwigs. Each bigwig has a different research direction and is good at different things, so it is actually difficult for them to collide with each other.
But now Tong Yongshan wants to cross the border, and this leap may lead to bad results.
Fang Yu, this kid, hasn't even graduated from college yet, so why does he act like a tycoon?
Tong Yongshan secretly complained that this kid would probably betray his teacher and ancestors and go against human ethics in the future.
But if that day really comes and I retire, it would be a good thing to have a promising disciple.
If this kid really betrays his teacher and ancestors after I retire, I will find an opportunity to remove his calligraphy.
Call him Fang Xiang.
Then let the academic community scold him together.
"Fang Yu, you can talk to this company first and see how much they can invest. If this project is commissioned by the company, it's not impossible, but the school's resources will not be used. Do you understand this?"
Tong Yongshan's words were rather obscure.
Fang Yu nodded. "Okay, teacher. I'll go talk to them tomorrow. However, the school also has the right to sign the document, so the license may still need to be issued by the school."
Tong Yongshan didn't even think about it: "Of course."
Fang Yu smiled slightly.
The ticket to artificial intelligence is now in hand.
(End of this chapter)
Continue read on readnovelmtl.com
