Crossover: 2014

Chapter 155 Are You Eager to Open That Door (Part [-])
Mina Kali is fairly familiar with China.

But it was the first time for Eve Carly to come to this ancient oriental country.

The experience along the way is a novel adventure for her.

Although her younger sister is in the imperial capital, she almost never envisages coming to China.

The reason for this trip to China is because of Ewald Cherry's suggestion.

Of course, this is only one aspect of the reason.

On the other hand, it was because Lin Hui was in the Northern Territory, which was the most important reason why she came to China.

She can't wait to meet the super genius who built the generative summarization algorithm.

After flying to Didu International Airport, after going through a series of necessary entry procedures.

Eve Carly didn't stay in the imperial capital for almost a moment. After meeting with Mina Carly, the two took the opportunity to go straight to the Northern Territory.

According to the previously negotiated and arranged process between Mizuki and MIT.

During the three days from the 23rd to the 25th, of the six academic symposia, three of the six academic symposiums originally required Eve Carly to attend.

But Eve Carly's mind is full of forest ash, and she has no interest in participating in academic conferences.

So all these academic conferences were shunned by Eve Carrie under the pretext of being ill.

Lin Hui, is such a stranger thousands of miles away worth making Eve Carly so excited?

Of course it is worth it. It is difficult for researchers who are not in related fields to appreciate what Lin Hui's previously proposed generative summary algorithm means.

As Ewald Cherry said earlier:

"The appearance of LIN HUI is like a beam of dazzling light suddenly appearing in the dark and distant wilderness.

In the segmented field of natural language processing text summarization, it is not too much to praise him. "

This statement is no exaggeration.

As the leader of a top research group in the field of text summarization.

Eve Carly knew very well why they failed to come up with a generative summary algorithm before LIN HUI.

Not because they don't try hard.

It is because the existing research on natural language processing involving text summarization has largely reached a dead end.

In this case, if other algorithm teams follow their research ideas.

Even if they surpass them, it is impossible to surpass them too much.

However, the generative text summarization algorithm proposed by Lin Hui easily crushed their previous research results.

What this means is self-evident.

For researchers in natural language processing, the greatest significance of Lin Hui's proposed generative summarization algorithm is not in the value of the algorithm itself.

It is because Lin Hui has most likely opened a new door in the field of natural language processing.

Because of this, Lin Hui's research results on generative text summarization algorithms are highly valued by the Massachusetts Institute of Technology.

It was also because of this that she would be so happy only after she got Lin Hui's consent to the email she sent to Lin Hui asking for a meeting.

She seemed to hear Lin Hui calling to her: "The door to a whole new world is there, are you eager to open that door?"

Although this feeling was in Eve Carly's imagination, Eve Carly would not give up even if there was even the slightest possibility.

What does it mean to open a new door in the field of science?

It means that this is something that is very likely to be recorded in the history of scientific development!

A scientific researcher at the level of Eve Carly may have a mediocre feeling about the "profit" of money.

But not everyone can say "no" calmly to the "name" that can be recorded in the history of scientific development.

Anyway, Eve Carly felt that she could not resist such an opportunity.

Although the heart is full of beautiful vision.

But before meeting with Lin Hui, Eve Carly was not completely excited.

Eve Carly was still vaguely worried.

She was very worried that her meeting with Lin Hui was not as smooth as expected.

Because there was no time to see Lin Hui.

Eve Carly already knew that Lin Hui was a young genius.

The genius who became famous at a young age is of course very admirable.

But not everyone wants to have such a genius by their side.

Eve Carly has seen teenage stars come before.

After all, the most indispensable thing in the field of computer and computer derivatives is genius.

Because I often encounter geniuses.

Eve Carly had a more or less stereotyped image of this young and famous genius in her mind.

Based on these stereotypes.

Although Eve Carly had many ideas about Lin Hui.

But these assumptions are basically inseparable from labels such as young, smart, arrogant, extreme, arrogant, outspoken, and contemptuous.

But even if it is expected that Lin Hui may have many shortcomings.

Deep down in her heart, Eve Carly was ready to tolerate Lin Hui's shortcomings such as arrogance, extremeness, and arrogance.

After all, she came to this eastern country with a learning attitude.

Before meeting Lin Hui, Eve Carly's only expectation for the upcoming communication with Lin Hui was:
——In addition to reaping unreasonable offenses, you can more or less get some valuable academic information.

After she actually saw Lin Hui, she realized that her previous thoughts were somewhat off track.

Lin Hui is indeed very young and smart.

But she was not as arrogant, extreme, and conceited as she had imagined before.

Not only was she not as arrogant and arrogant as she had imagined before.

Lin Huiren is handsome, with a good personality, modest and elegant, and he also takes care of other people's feelings in his speech and demeanor, which makes people feel very comfortable.

In addition to his gentle personality, Lin Hui is also very considerate in dealing with people.

Although the apartment building for this exchange is not very gorgeous, it has a beautiful environment.

The most wonderful thing is that there is an artificial lake not far from here, which is more or less similar to the place where Eve Carly worked before.

And the form of this academic exchange is very similar to Eve Carly's previous team communication.

Several people with the same interests and hobbies get together and start chatting in the form of small talk.

All this gave Eve Carly a rare intimacy in a foreign country.

These thoughtful arrangements made Eve Carly flattered.

Apart from these, what Eve Carly cares most about is Lin Hui's academic attitude.

What surprised Eve Carly the most was Lin Hui's academic attitude.

Although Lin Hui is far ahead of others in terms of research results, Lin Hui has no airs of being a scholar at all academically.

Lin Hui is very good at listening.

This is an extremely rare thing.

Finding an expert who is good at listening among natural language processing experts and scholars seems to be more difficult than finding a giant panda on earth.

Many experts and scholars in natural language processing are from computer practitioners.

In Eve Carly's impression, this kind of personnel has always expressed themselves, and the worst thing they are good at is listening.

Maybe it's not that I'm not good at listening, it's just that I don't like listening.

It seems that listening to other people's thinking lines and research status can easily remind them of the painful experience of debugging in their early years.

But the situation here is very different with Lin Hui, Lin Hui is very good at listening.

At the beginning of the symposium, Eve Carly originally planned to let Lin Hui express his opinion first.

In the end, Lin Hui signaled her to speak first, which made Eve Carly very uncomfortable.

For a while, she didn't know what to say, so she could only explain in detail the email she sent to Lin Hui not long ago.

The reason why Eve Carly introduced this aspect was that she couldn't think of a suitable starting point for the topic.

There is another reason, that is, Eve Carly is very curious about how Lin Hui evaluates the text similarity when constructing the LH text summarization model.

But Eve Carly was too embarrassed to ask this question directly, so she had to make insinuations.

At the beginning of the statement, Eve Carly was still a little nervous, afraid that repeating the repeated content in the email would cause Lin Hui's dissatisfaction.

But Lin Hui didn't seem to mind, and just listened to her statement seriously.

Lin Hui's attitude made Eve Carly less nervous.

While making the statement, Eve Carly noticed a small detail:

On the way back from the airport, when she suggested to Lin Hui to find an interpreter for simultaneous interpretation, Lin Hui almost agreed without thinking.

But in the actual communication, Eve Carly judged from some of Lin Hui's reactions that Lin Hui could actually understand what she expressed directly.

That being the case, why did Lin Hui agree to her request?

Instead of just throwing away the translator and communicating with her?

Perhaps all this is to give equal respect!
This equal respect is not only given to Eve Carly, but mainly to Mina Carly.

Just imagine if Lin Hui could talk to Eve Carly without needing an interpreter at all.

It seems that the most embarrassing is Mina Kali who is traveling with Eve Kali.

It is really not easy for a man to be so careful.

Eve Carly's favor towards Lin Hui increased by a few tenths of a percentage point.

The younger sister seems to have noticed Lin Hui's considerate approach in simultaneous interpretation.

Eve Carrie noticed how many times Mina Carrie had flicked her hair, consciously or not.

Of course, Mina's overtures might just be because of Lin Hui's good looks.

It stands to reason that the appearance of Eastern men is difficult to distinguish in the eyes of Westerners.

But being handsome to a certain extent transcends geographical limitations.

This seems to be the case with Lin Hui. Even if judged by the most stringent aesthetic system, Lin Hui's appearance can score 99 points, out of 10 points.

When seeing Lin Hui for the first time, if Lin Hui didn't take the initiative to reveal his identity, Eve Carly even thought that Lin Hui's identity would be a model.

Of course these are off topic.

After noticing Lin Hui's intentional or unintentional goodwill in the details.

Eve Carly relaxed completely as she made her statement.

He introduced to Lin Hui how people evaluate text similarity in this time and space.

Eve Carly noticed that Lin Hui frowned when she heard that her team had previously used a method based on network knowledge to evaluate text similarity.

Could it be that Lin Hui does not agree with the method of evaluating text similarity based on network knowledge?
Or does Lin Hui think there is any method better than this method?
Eve Carly kept this matter in her heart silently.

After Eve Carly's presentation is complete.

Lin Hui understood what she meant.

However, it did not answer Eve Carly's question directly.

Instead, I asked Eve Carly: "What do you think about the use of vector intervention for semantic text similarity calculation?"

Although this is the first question raised by Lin Hui in this exchange.

But this question caught Eve Carly somewhat off guard.

Eve Carly didn't quite understand why Lin Hui asked this question.

Could it be possible to calculate semantic text similarity without relying on vectors?

But how can this be done?

When the machine recognizes the text, in order for the machine to recognize the natural language, the natural language is often digitized.

However, it is necessary to carry out vectorization to distinguish these values ​​by attributes.

This method has been around for a long time. Eve Carly remembers that in 1977 (this time and space), researchers first proposed the vector space model VSM.

Once this research method is proposed, it is more popular.

Although this method was soon discovered to have a lot of loopholes.

Using the VSM method, when the amount of text is large, the generated text vectors are very sparse, which leads to a waste of space and computing resources;
In addition, VSM ignores the relationship between words in order to achieve the effect of simplifying the model, but in many cases there is a connection between words, so it is unreasonable to simply think that words are independent of each other.

Despite the obvious loopholes, in the following nearly 40 years of history, people still have to introduce vectors for semantic text similarity analysis.

Take Eve Carly's previous team, although they used the method of calculating text similarity based on network knowledge.

But in essence, it just maps the content of the webpage in Wikipedia to a high-dimensional vector,
Then, the semantic text similarity calculation is carried out by the method based on the vector space.

It can be said that it still has not been able to leave the shell of the vector space.

Although 40 years later, the so-called "waste of space and computing resources" encountered in those years can be solved violently through hard heap computing power to some extent.

But this is only to solve the problems encountered in the past.

The complexity of the amount of information faced in text processing now is completely different from that of the past.

At this time, the vectorization is facing a new difficulty - dimension explosion!
The curse of dimensionality (aka the curse of dimensionality) is a term first coined by Richard Bellman when considering optimization problems to describe the analysis and organization of high-dimensional spaces (often hundreds of Thousands of dimensions), various problematic scenarios are encountered due to the increase in volume index.

When an extra dimension is added to a mathematical space, its volume grows exponentially.

Such difficulties are not encountered in low-dimensional spaces.

For example, physical space rarely encounters such problems. After all, physics is usually only modeled in three dimensions.

It's amazing to say the least, although it's hard to encounter the dimension explosion problem physically.

But dimension explosion is common in natural language processing and machine learning.

Any amount of information in this field will easily break through three dimensions.

In fact, in many fields, such as sampling, combinatorics, machine learning, and data mining, the phenomenon of dimension explosion is mentioned.

The common feature of these problems is that when the number of dimensions increases, the volume of the space increases too quickly, so the available data becomes very sparse.

In a high-dimensional space, when all data becomes very sparse and dissimilar from many perspectives, the usual data organization strategies become extremely inefficient.

In fact, Eve Carly and her previous team used network knowledge to measure text similarity.

If all webpages are analyzed directly, it will often lead to the difficulty of sparse calculation of knowledge content.

In fact, this situation is caused by the explosion of dimensions.

Eve Carly is very clear that the current method of using vectors to introduce semantic text similarity will bring about a dimension explosion.

Why did Lin Hui suddenly ask her how to see the introduction of vectors into the calculation of semantic text similarity?

Does Lin Hui really have any way to properly deal with the problem of dimension explosion?
However, the dimension explosion in the direction of machine learning and natural language processing is not so easy to solve.

Or is Lin Hui going to simply bypass the vector to measure the semantic text similarity?

　(●''●)
　　
　
(End of this chapter)