Crossover: 2014

Chapter 196 Appreciation from the boss

He Tianchang also gained a lot from this exchange with Lin Hui.

Of course, at He Tianchang's level, besides the most cutting-edge knowledge, it is already difficult to acquire any new knowledge.

And even if Lin Hui has some academic advantages, it is impossible to show off regardless of the occasion.

Therefore, He Tianchang's main gains are:
Through the communication with Lin Hui, some previous judgments were confirmed.

The first is a series of patents involved in Nanfeng APP.

At that time, among the bunch of information about Lin Hui collected from Lao Fu, what surprised He Tianchang the most was the bunch of patent applications.

When you see Linhui's complicated patent applications at home and abroad.

He Tianchang felt as if he saw a large net that was carefully woven.

As for why we weave such a net, the high probability is to build patent barriers.

When he came to this judgment, He Tianchang admired Lin Hui very much.

After all, it is really rare to be able to have this calmness at the age of 18.

During the phone call, after He Tianchang raised this matter, he got verification from Lin Hui.

After confirming his own judgment.

He Tianchang couldn't help admiring Lin Hui even more.

Of course, He Tianchang knew that the reason why Lin Hui admitted this without hesitation was because the barrier Lin Hui built had already been realized.

He Tianchang has been paying attention to the situation of Lin Hui's patent application.

Therefore, He Tianchang had noticed the fact that Lin Hui's patent barriers were successfully organized into a network.

He Tianchang felt that if Lin Hui's plan hadn't been realized, even if He Tianchang raised this question to the extent that the two had just met for the first time.

Lin Hui may not readily admit this matter.

If it were 20 years ago, He Tianchang would probably be disgusted by this cautious and calculating personality.

But now, He Tianchang appreciates this character instead.

_For a time research seemed to be too eager to progress.

Many technologies are recklessly brought out before the completion of the construction of technical barriers.

Many technologies are obviously of great commercial value, but they are often exploited by opponents due to insufficient preparation.

In the end, it was obviously gold inlaid with jade, but it could only be sold for the price of cabbage.

Can not help but make people sad.

In the communication with Lin Hui, He Tianchang asked some questions about generative text summarization algorithms.

Of course, what He Tianchang was curious about was not the specific technology itself.

What he was curious about was Lin Hui's scientific research process.

How did a person like Lin Hui, who has not studied systematically in academic aspects, handle the world's leading algorithm by himself?

He Tianchang asked Lin Hui some routine questions.

For example, how did Lin Hui get the corpus for language model training?
In fact, this is not the first time Lin Hui has heard this question.

Eve Carly had asked Lin Hui this question in a letter before.

But even if Lin Hui had never heard this question asked before, he didn't have to worry about being asked.

When it comes to generative text summarization algorithm patents, the most likely to confuse the outside world is the problem of corpus.

If you don't think about how to explain the corpus used to train the language model in advance, you will rush to come up with new results in text summarization, and you will be easily questioned in one way or another.

Lin Hui had noticed this problem a long time ago.

Forewarned is forearmed, without prejudging the waste.

For this question, Lin Hui prepared at least three alternative answers.

Lin Hui told Professor He Tianchang the speech he had prepared earlier.

He Tianchang felt a sense of enlightenment.

He couldn't help feeling that the waves behind the Yangtze River pushed the waves ahead.

However, in the several methods mentioned by Lin Hui.

What is strange about He Tianchang is the method of automatically building a text corpus with the help of the Internet:

When using this method to construct a corpus, the user only needs to provide the required text category system.

Then collect a large number of websites from the Internet, extract and analyze the content hierarchy of the website and the webpage content information corresponding to each keyword.

The texts required by users are screened out from each website as candidate corpus.

The formed corpus is then denoised.

In fact, this method He Tianchang remembers reading some foreign university journals that seem to have recorded research in this area.

But that failed abroad, because the collected corpus was too noisy and had too many stop words, making it unusable.

Why did Lin Hui propose this method.

Mo Feilin must have a unique understanding of the algorithm used for denoising.

In fact, He Tianchang is not very good at NLP.

But it's not a big problem, some old friends in He Tianchang's country are very good at this.

He Tianchang silently wrote down the matter.

Lin Hui gained a lot from the academic exchange with He Tianchang.

First of all, he learned about the research situation in the frontier direction of natural language processing in China through He Tianchang Linhui.

And what is the current state of research in the direction of NLP in China?

A simple summary is "a blank sheet of paper"

Of course, it is not to say that there are no domestic researches on natural language processing, but the research progress of these people is roughly similar to that of international research.

In other words, it lags behind the research progress of the previous life in 2014 as a whole.

Under such circumstances, if Lin Hui wanted to be an academic porter, it seemed that he would be in a no-man's land.

The two directions of natural language processing and neural network learning are almost a blank sheet of paper in front of Lin Hui.

Waiting for Lin Hui to write a magnificent chapter on it.

But even so, when it comes to specific implementation, Lin Hui will not do everything.

It only needs Lin Hui to make some key progress in time.

There is no need to be too greedy for progress when it comes to some trivial progress.

After all, academic achievements are not achieved overnight.

Anything that involves something as simple as the theory behind generative summarization algorithms is super troublesome.

Although troublesome, the results are worth looking forward to.

when the call is about to end.Lin Hui thanked Professor He Tianchang several times for helping him apply for some support.

He Tianchang insisted that even if there are supportive policies in the future, Lin Hui deserves them.

That's what he said, and He Tianchang also thought so in his heart.

Although He Tianchang's research direction does not involve natural language processing.

But this does not mean that He Tianchang is ignorant of the research direction involving natural language processing.

Stones from other mountains can be used to attack jade. In scientific research, there are often sayings that by analogy.

In many cases, properly learning from the research ideas of other industries can bring inspiration to one's own research direction.

Therefore, even if the main direction of attack is not the direction of natural language processing.

But this does not affect He Tianchang's attention to other research directions.

At the very least, He Tianchang is still very concerned about the progress made in some computers and computer derivatives.

In addition, the direction of natural language processing is more friendly, and does not require a high threshold.

Therefore, He Tianchang has dabbled in natural language processing.

(End of this chapter)