Crossover: 2014

Chapter 105 No low-level patents

Although it was difficult for Eve Carly to accept this fact for a while, she quickly calmed down.

Whether it is a zero-sum game or a non-zero-sum game.

Although their team lost in the competition with LIN HUI, she was convinced of the loss.

The algorithm of LIN HUI has a crushing lead both in terms of actual performance efficiency and algorithm theory, rather than leading by 01:30 points.

If it is only leading by 01:30, it may be a fluke, but a crushing lead means an unquestionable gap in strength.

Eve Carly has a deep understanding of this, and the generative summary algorithm proposed by LIN HUI has even created some unprecedented research in subdivided fields.

A new subdivision that is easier to know is sentiment analysis for natural language processing.

This direction is a brand new topic for Eve Carly, who has only done research on extractive text summarization before.

But this is far from all. The generative summary algorithm proposed by LIN HUI should also involve many pioneering things.

However, due to the reason of peeking at the leopard, Eve Carly is not yet clear how much groundbreaking research is involved in the algorithm of LIN HUI.

Although it is not clear how many new segments are involved.

However, Eve Carly relies on the intuition of scientific researchers, and it is conservatively estimated that the algorithm of LIN HUI will involve five or six pioneering research in subdivided fields.

However, it was precisely because of this that Eve Carly couldn't understand.

Obviously, the algorithm of LIN HUI is so much ahead, why should we seek the acquisition of many "low-level" patents?
Eve Carly looked at the patents that Lin Huihui was seeking to acquire on the patent website.

It seems to be very "low-level".

When it comes to the distinction between low-level and high-level, there is no difference between low-level and high-level patents themselves.

However, in a research system, due to the different division of labor and different levels, there will be low-level and high-level points.

After LIN HUI proposed a generative text summarization algorithm.

The current automatic summarization methods are mainly divided into extractive methods and generative methods.

After recent research on the technical route of generative summarization algorithm, Eve Carly has been able to easily understand the difference between generative text summarization and traditional extractive text summarization.

The so-called extractive summarization is to extract key text units from the original document to form a summary.

The generative summary is based on the understanding of the input original text to form a summary. The generative summary model tries to understand the content of the text, and can generate words that are not in the original text, which is closer to the essence of the summary and has the potential to generate high-quality summaries.

Although there is a slight difference between the two in specific summarization, both are essentially automatic text summarization.

Since both are automatic text summarization, the technical framework of both can be summarized as follows:

Content representation → weight calculation → content selection → content organization.

ⅠContent representation is the process of dividing the original text into text units, mainly preprocessing work such as word segmentation, words, and sentences;

The main purpose of content representation is to process raw text into a form that can be easily analyzed by algorithms through preprocessing.

II The weight calculation is to calculate the corresponding weight score for the text unit (that is, the original text after preprocessing).

The purpose of this step is to complete the preliminary analysis of the preprocessed original text through this series of calculations.

ⅢContent selection is to select the corresponding subset of text units from the weighted text units (that is, the text analyzed by the weight of step Ⅱ) to form the abstract candidate set. According to the required abstract length, linear programming, submodular function, heuristic formula algorithm, etc. to select text units;

ⅣContent organization refers to sorting the content of the candidate set to form the final abstract, which can be output in order according to the word count requirements. Some researchers also propose to use methods based on semantic information, templates, and neural network learning to generate abstracts that meet the requirements.

(ps: ... In terms of popular understanding, II is a bit similar to finding key paragraphs when reading and summarizing articles;

Ⅲ is similar to the process of further finding key sentences and keywords on the basis of Ⅱ;
IV is similar to the process of forming the final reading summary with appropriate language after determining the key sentences and keywords)

Judging from the corresponding descriptions of these levels of the technical framework, it can be seen that all three are very important, whether it is weight calculation, content selection, or content organization.

If you can't figure out the weight calculation and content selection, you can't figure out where to summarize the text when summarizing.

After all, not all places in an article are the key points. Just like when we read and summarize, we usually focus on the end of the first paragraph and the beginning and end of each paragraph. It can be said that weight calculation and content selection play a role It is to judge where the corresponding text feature points of the text to be processed are mainly concentrated.

If you can't figure out the content organization, you can't get a reasonable and smooth text summary even if you can find the place with the most dense text features.

Content representation is less important than these three.

Because of the above-mentioned division of labor, if things related to text summarization should be divided into layers.

Then the patents related to weight calculation, content selection, and content organization can be said to be advanced patents in the text summarization system.

Patents related to content representation are low-level patents.

And "A New Method of Text Judgment, Screening and Comparison" is essentially used to identify original texts.

According to the introduction just now, this undoubtedly belongs to the level of content representation.

Just such a "low-level" patent,
Eve Carly really doesn't understand why LIN HUI seeks to purchase such a patent.

Is it true that as Ewald Cherry said, what LIN HUI cares about is not the algorithm patent itself of "A New Method of Text Judgment, Screening and Comparison".

What LIN HUI cares about is her who got the patent?

how is this possible?If you are just interested in her, you can contact her directly.

After all, it is very easy to find the contact information of scientific researchers.

Eve Carly guessed several possibilities, but did not guess a reasonable explanation.

……

Although a little hard to understand.

But in the end, Eve Carly sold the patent "A New Method for Text Judgment, Screening and Comparison" to LIN HUI.

After all, the question that the seller has to think about is whether the buyer's bid is appropriate, not why the buyer buys.

In addition, Eve Carly noticed that one of the patents that Lin Hui had previously purchased had been successfully transferred at a price of 50 US dollars.

Although the value of this patent was not as high as hers, the seller was Asile Velasquez.

Eve Carly remembered that this person was a senior researcher at Google Research (Google Search, the predecessor of Google AI).

To be honest, the reason why Eve Carly struggled with whether to sell this patent was not because she was worried about making less money, but because she was worried about the negative impact.

But now, no one inside Google is worried about the negative impact caused by the patent transfer, so she has nothing to care about.

As for whether the patent "A New Method for Text Judgment, Screening and Comparison" has potential value, it is no longer important.

Eve Carly was quite looking forward to the new life in the hands of LIN HUI for this patent, which she didn't see any value in any way.

Perhaps LIN HUI will prove that there are no low-level patents, only low-level vision.

　The reason why the role of Eve is arranged is because there must be someone who can understand and implement the purely technical tasks explained by the protagonist. Uh, Huang Jing feels that this cannot be done.Don't pay too much attention to the characters who appeared in the book before and don't appear now. The existence of some people seems to only confirm the growth of the protagonist.

　　
　
(End of this chapter)