Crossover: 2014

Chapter 260 High-level game: the battle for the right to speak
People who meet these two conditions at the same time are often completely rare.

It is precisely because of this kind of professional labeling that there are very few people.

Often expert-level annotation is only capable of handling small-scale data.

For large-scale data labeling and ultra-large-scale data labeling tasks, it is often impossible to rely on expert labeling.

For large-scale data labeling and ultra-large-scale data labeling, we can only seek crowdsourcing.

The crowdsourcing model is a service model that integrates scattered individuals (including part-time jobs) and small annotation teams into the platform to complete a complete project.

The main advantages of this model are low cost and relatively flexible.

Although the long-term work of machine learning in the previous life is to be able to replace expert annotation with crowdsourced data or simply unlabeled data.

But I really want to ask which one is more popular, expert annotation or crowdsourced data.

That is naturally the former.

In this case, it is easy to get a higher premium by exaggerating the proportion of expert annotations in the labeled data.

Sounds like a profiteer.

But that's just what it sounds like.

If there is really a big difference between the labeling data that Lin Hui praised so much and the current level of expert labeling.

It's useless for the forest to blow the sky.

After all, you need to be hard to strike iron.

Since Lin Hui dared to claim that the proportion of expert annotations in the annotated data is very high, he has confidence.

how to say?

In the next few years, even some data marked by non-experts will be standardized and normalized in the industry.

In many cases, even crowdsourced data may not be worse than the level marked by experts who have not yet fully standardized and planned in the industry.

Even if Lin Hui used the non-expert labeled data in the next few years to fool people.

If you want to come to this time and space, someone will buy it.

No way, technical strength means that you can do whatever you want.

Usually, money can do whatever it wants.

But while doing whatever you want, others are likely to treat you as a grandson.

You can do whatever you want with technology.

But when you do whatever you want, others have to treat you like a grandfather.

Take a look at some manufacturers in the later generations who are celebrating their birthdays and grabbing the first release (no specific reference, don’t take a seat according to the number)

Probably, we can get a glimpse of the saying "grandfather status due to technology".

2021 time and space came to 2014 time and space.

Although there is only a seven-year time difference between these two time and space.

But this means that Lin Hui has at least seven years of information gap in most fields.

However, due to the short legs in some aspects of this era, Lin Hui has an information gap of far more than seven years in some fields.

In this time and space, many technologies that were originally not so advanced in previous lives are existences that are ahead of the times.

In this way, Lin Hui can really do whatever he wants.

Even in many areas of this time and space, Lin Hui is really a grandfather-level existence.

Others set foot in the same field that is to be a grandson.

If you become a grandson, will there be anyone?

Of course, I don't know how many people are lining up to be grandchildren.

In some fields, even Lin Hui's potential status is not as exaggerated as "grandfather level".

But Lin Hui also has a very obvious prophetic advantage.

This kind of prophetic advantage is not only manifested in the prophet of technical details.

It also focuses on details that ordinary people are not easy to perceive, such as route advantages.

In the case of almost absolute advantages in many fields.

Lin Hui's status is prominent.

Potentially possessed status is even more detached.

Even as long as Lin Hui doesn't commit crimes.

This is almost the era when Lin Hui can do whatever he wants.

But in addition to the hard red line of the law and the most basic personality bottom line.

There will always be an invisible red thread affecting Lin Hui.

Of course, this red line is not moral.

In fact, there is basically no red line in the moral aspect of some people who have achieved fame or small success, only the yellow line.

The general public has quite broad moral constraints on those who have achieved success or achieved something in a small way.

Even before that, Lin Hui was ready to accept moral kidnapping.

But the fact is that Lin Hui has not been charged in this regard.

There are even many people who will try to lure Lin Hui in the direction of the moral abyss.

For example, on Weibo/blog, many people send Lin Hui various private messages to give him a monkey every day.

He even enthusiastically sent various illustrations to Lin Hui (at this time, the basic display of the shielding system), but he was very speechless.

Although Lin Hui has moral standards in his heart.

But potential constraints when it comes to morality are not Lin Ash's potential enemies.

Another red thread that really affects Lin Hui is invisible and intangible.

That is "reasonableness".

In many cases, rational considerations are taken into account.

This made Lin Hui unable to do whatever he wanted.

Even though Lin Hui is ahead in many technologies in an absolute sense, he still cannot transfer them recklessly.

There is an invisible red thread of "rationality".

Lin Hui has many scruples in doing things.

If it is reasonable on the logical level, then Lin Hui has nothing to say.

Lin Hui is a person who respects the internal logic of things and the laws of objective development of things.

But the problem is that many times the so-called "rationality" is not based on logic.

But completely from the "camp" point of view.

Compared with the rational interpretation starting from the logical level.

The more common interpretation of "rationality" is actually the latter.

What are the judging rules from the perspective of camp?
To put it bluntly, it is ideology.

But not all ideologies are keen on rational interpretations.

The main representative of the ideology keen to interpret rationality is naturally the various rules imposed by the West on the world.

This phenomenon is extremely common in the fields of science and technology and academia.

When moving technology, except for some rational considerations are from a purely logical point of view.

Many times, Lin Hui felt that some of the decisions he made for the sake of rationality were just compromises he had to make in consideration of the rules of the Western world.

These Western rules are invisible and intangible.

But most of the time, it influences Lin Hui's choices.

This is undoubtedly very uncomfortable, but if you want to seek benefits under the rules, you have to submit.

But after all, in many cases, the first comer is king, and the latecomer suffers.

The fact is that even when Lin Hui traveled, people often had to endure the same situation.

Simply, go back to seven years ago, when there was a lot of superior information.

Lin Hui is expected to accelerate the change of this situation.

Various technical transfers carried out by the forest ash station or prepared to be carried out.

The ultimate goal of Lin Hui is to be able to say goodbye to some rules shaped by the current world and the West.

When Lin Hui grows up to the point where he can stand without relying on the rules of the West, even the West has to rely on the rules formulated by Lin Hui to stand up.

That Lin Hui can completely say goodbye to the current Western rules.

This indictment of a reading of "rationality" not based on logic is more than an indictment of certain institutions prone to injustice.

This accusation of "rationality" boils down to:

——The "right to speak" dispute.

Having the right to speak is equivalent to being able to explain the rationality by yourself.

In this way, having an absolute right to speak is equivalent to having authority.

And with absolute authority, sometimes it can even ignore the factual objective laws.

Including but not limited to logical reasonableness.

The scene is roughly: "Whatever is reasonable or unreasonable, lz is rational, and if lz says it is reasonable, it is reasonable."

When it comes to authority, what is the use of such capriciousness?
Of course it is useful, to be able to have such a voice and authority.

It will make the speed of forest ash handling straight to full.

Moreover, having absolute authority can also directly maximize the benefits of the information brought by Lin Hui from his previous life.

After all, absolute right to speak means absolute monopoly.

After gaining absolute authority, even if Lin Hui never thought about pursuing a monopoly in the market, he could still achieve the same effect as a monopoly.

Take the "marked data" that Lin Hui just thought of.

Lin Hui has a lot of labeled data in his past life information.

Annotated data involving machine learning can be exchanged for money.

Annotated data in natural language processing can also be exchanged for money.

But is labeled data the fastest way to exchange money?
of course not.

When it comes to natural language processing, it is narrow knowledge that really pays the most money.

Narrow knowledge is also a source of knowledge in natural language processing.

There are three main categories of knowledge sources applied to NLP projects:
—Narrow knowledge, algorithms and data.

As the old saying goes, there is a golden house in the book.

Although the real golden house is not always the case.

But knowledge can be exchanged for money, this is well known.

Since knowledge can be exchanged for money.

Then the source of knowledge can naturally be exchanged for money.

Even the means of knowledge source is likely to make money faster than knowledge itself.

Algorithms and data, which are the source of knowledge for natural language processing, can be exchanged for money.

Narrow knowledge, which is also the source of natural language processing knowledge, can naturally be exchanged for money.

It is basically known to the world that algorithms can be exchanged for money.

Even if you didn't know that this money can be exchanged before, if you pay attention to Lin Hui's busy focus some time ago, you can understand it.

Previously, what involved Lin Hui most of the time was basically the matter of the generative summary algorithm.

If there is no profit, it will not be too early, and if there is not enough profit to drive Lin Hui, naturally he will not rest on his laurels.

It is basically not difficult to understand that algorithms can be exchanged for money.

Algorithms often directly affect the efficiency of some algorithm-driven products.

And efficiency is real money.

Algorithms that can directly affect efficiency can easily be exchanged for generous rewards.

Understanding the algorithm can exchange money.

In fact, it is not difficult to understand why data can be exchanged for money.

After all, data is the cornerstone of many machine learning algorithms.

The emergence of machine learning algorithms often relies on labeled data.

And for a long period of time, machine learning algorithms do not only rely on labeled data.

And it relies on a large amount of labeled data.

When the amount of labeled data is small, it is often not enough to train a machine learning algorithm with excellent performance.

From this perspective, it is not difficult to understand why data can be exchanged for money.

In many cases, data can even be completely understood as a kind of tacit knowledge.

The process of data labeling is actually the process of structuring and labeling the scattered and discrete data.

Apart from algorithms and data, what is the so-called narrow knowledge?

Knowledge in the narrow sense generally refers to explicit knowledge defined by humans through rules or dictionaries.

Knowledge in the narrow sense mainly includes three types:

—that is, language knowledge, common sense knowledge and world knowledge.

Among them, language knowledge refers to the definition or description of the morphology, syntax or semantics of the language.

Its main feature is the definition of a collection of synonyms.Each synset consists of words with the same meaning.

Common sense knowledge refers to the basic knowledge that people acquire based on common experience.

World knowledge includes entities, entity attributes, relationships between entities, etc.

Maybe some people don't understand?
Why can such knowledge be exchanged for money?
Aren't these things all obvious?

Although this knowledge is essentially explicit knowledge that people can understand.

But explicit knowledge that is obvious to people.

Not equal to the machine is also obvious.

For this kind of knowledge, it is often processed through regularization or lexiconization so that this knowledge can be easily understood by machines.

Knowledge that is easily understood by machines is called narrow knowledge, also known as expert knowledge.

Although the main training model now is to seek the algorithm or even the data itself.

But knowledge in the narrow sense is quite marketable.

Anyway, judging from the previous communication with Eve Carly.

It’s already 2014, and Silicon Valley still has to cooperate with Harvard, Oxford and other universities to develop some machine learning.

The reason why these people rely on Harvard and Oxford is that they expect these universities to carry out data labeling.

The main thing should be counting on the blessing of these colleges and universities in the narrow sense of knowledge.

It is easy for these people to do so.

After all, when it first involved model data in natural language processing, people trained with narrow knowledge instead of relying on data and algorithms.

Even in Lin Hui's impression, even in his previous life, before the rapid rise of the Internet, the only way for people to train natural language processing models was to use narrow knowledge.

And Lin Hui has quite a lot of knowledge in the narrow sense in his hands, and his level should be much higher than what is used in the western world now.

Judging from the fact that Silicon Valley spends tens of millions of dollars every year in order to obtain knowledge in a narrow sense.

If some narrow sense knowledge in Lin Hui's hands can be monetized, it is more convenient than using labeled data to monetize.

But this is only theoretically easy to realize.

Lin Hui does not have the absolute right to speak and the accompanying authority.

How does Lin Hui tell potential audiences that the narrow knowledge materials he possesses are superior to the general materials used in Silicon Valley today?

In fact, Lin Hui, the potential buyer, is very clear about it.

You must know that even in the next few years, there will not be too many buyers who are interested in a large amount of narrow knowledge and are not short of money.

Lin Hui estimates that the buyers who may be interested in large-scale narrow knowledge bases in this time and space are nothing more than super giants such as Microsoft and Google.

But even if he knew about these potential buyers, it was impossible for Lin Hui to take the initiative to find them.

(End of this chapter)