Chapter 60: Give Words a Vector

(A notice that the novel's timeline has been changed to November 2017 at the request of everyone, and the plot time is now March 2018.) It was originally set at the end of November 2009 and began not to copy the existing technology for the copyist, but to stretch the timeline a little longer. The revised protagonist's family income, housing prices, mobile phones and computers used in the previous article are in line with the current era, and there is no other change and does not affect the following text. If you find that there is anything I have not modified before, please let me know, and the explanation does not occupy the number of words. Above. )

Xiao Ming read the advantages of the Pangu language in detail, and the biggest point is that it integrates all the logic and rules of human Chinese, and can directly tell the machine the meaning of human language expression.

The Pangea language would have worked more effectively if it had been used on Panshaxing's biological thinking computer, but now it can only be said to be barely functional when used on a binary computer.

Next, Xiao Ming redeemed the app package he purchased for translation software.

Detailed programming steps for the translation software appear on the computer.

The redeemed application package will not directly send you the app, but let Xiao Ming actually operate it and make up a program by himself.

Teaching a man to fish is not giving a man a fish, and this is the truth.

Xiao Ming thought, in fact, his English has improved a lot, the grammar problem is not big, the biggest disadvantage is that the vocabulary is small, the listening and speaking skills are poor, and he can't read and understand or speak.

This is also the dilemma of most Xia Guo students learning English.

Does English matter? At this stage, it definitely does. At present, a large number of technical sciences of mankind come from the West, and if you are not good at English, you can't even read SCI papers, let alone do academics.

Xiao Ming's biggest change in the past six months is that he will self-reflect, and he will also reflect on his poor English.

He can draw treasure chests and exchange them for technology, but these technology products are also based on basic scientific and technological knowledge.

If Xiao Ming didn't understand the basic biological knowledge, he wouldn't be able to cultivate the Devouring 1 fungus, and similarly, if he didn't understand the knowledge of logic, he would definitely not be able to program. If you don't know English, you can't understand foreign academics, and you won't make progress in science and technology.

There is no free lunch in the world, and no matter how much you have a plug-in in life, you have to work hard.

Back on the computer screen, Xiao Ming had a bold idea, what he needed was not only a translation software, but a software that could intelligently communicate with him in English and improve his English listening and speaking skills as soon as possible.

According to the programming instructions of the Pangu language, Xiao Ming began to do it.

First of all, on the programming page, Xiao Ming writes the general description of the application software - it can intelligently and accurately translate English and Chinese to each other, and can talk to users.

The next step is to write the program.

Xiao Ming's English vocabulary is insufficient, but there is no problem with grammar.

Xiao Ming summarized that there are two biggest flaws in translation software and translation machines on the market today.

One is that the words don't make sense. Whether it's English or Chinese, there are usually multiple meanings and different interpretations in different contexts, but machine translation doesn't fully understand what humans mean. Many times the words have the exact meaning of the words, but they are full of jokes when put in sentences.

Another is the inability to recognize human speech. This mainly appears on the translation machine, everyone has many kinds of accents, there are a lot of slang in the dialogue in life, and it is absolutely impossible for people to talk to each other like CCTV anchors when every sentence is complete with language elements and pronunciation standards.

In many cases, machine translation will pick up the translations that you can understand, and the translations that you can't understand will be messy. This is also the reason why many brands of translation machines make customers feel uncomfortable when they are applied abroad, and the translation machine cannot be used as a simultaneous translation for meetings.

In order to deal with the above two main problems, Xiao Ming edited according to the suggestions in the manual.

Xiao Ming uses mathematical thinking to set each word as a vector and classify it into nouns, verbs, and so on.

The advantage of setting words as vectors is that long and difficult sentences are dismembered, and the translation software will accurately translate each word when processing it.

The next step is to filter and combine different words according to the context of the language, combine different words according to the grammar and meaning required by the target audience, and make up for the missing grammatical elements.

Under the prompting of the Pangu programming language, Xiao Ming knew that his programming logic was correct.

However, logical correctness is only the first step, and how to make words with vectors syntactically combine into new sentences is difficult, which is also the biggest difficulty of modern translation software and machines.

It doesn't matter, that's what Pangu does best.

Pangu gave Xiao Ming a few access points.

Xiao Ming will import a large number of Chinese and English materials, including not only famous books, but also online novels, Tieba Q&A, Weibo, Twitter articles and so on.

In the future, this information will be uploaded by users themselves to optimize the accuracy of the program.

Pangea's database will integrate this information, familiarize yourself with the context of each sentence, and then collate the data model (a model that simulates the way human minds are expressed in Chinese and English).

This data will help the "word vectors" appear in the right places in different contexts and grammars, so that the translation will be more accurate.

The biggest difficulty of this work is that the compilation volume is very large!

Therefore, the existing translation software on the market today uses a grammar library summarized by linguists, and even if a small number of software has a self-learning function, it cannot understand and count all the current language habits and analyze them. This is also the reason why existing translation software is not mechanically intelligent and full of errors.

The remaining interfaces, Xiao Ming will access them to free live broadcast rooms across the country, and the anchors in the live broadcast room have Mandarin and local languages, which are also oral expressions, which are the most representative.

The Pangu language collects the sounds and tones of various places, classifies and compiles them, and finally forms a phonetic database corresponding to the text database.

The way of using Pangu programming is very simple, you don't need to enter code, you only need to tell the logical intent, and when Xiao Ming tells the logic and the method, the programming language can run explicitly.

And then......

Then the computer is stuck, and the card is hot!

Xiao Ming wasted half a day's hard work and hard work came to naught.

Labour......

The notebook that Xiao Ming bought is an ordinary ASUS notebook, using an i7 8550U processor.

The processor of the laptop processes such a large amount of data, it is no wonder that it does not get stuck!

Xiao Ming looked at the time, it was already three o'clock in the morning.

"I need a set of servers. Xiao Ming said, and then he lay on the bed and continued to think about the logic of language translation and English learning assistant software.

For the next few days, Xiao Ming was silent at school.

In addition to doing the necessary math and science exercises, I spend most of my time doing English reading and Chinese reading.

The members of the school group all knew that Xiao Ming was not in a good mood after the teacher from Mizuki University left, so they didn't bother him.

A few days later, Xiao Ming asked his father for 100,000 yuan, and directly purchased four sets of server hosts built by Intel Xeon E5-2603v4 chips online, and contacted them to install them. It also spent money to open an enterprise optical fiber dedicated line.

Xiao Ming looked at the silver of the white flowers, and the time was gone today, and it was also painful.

The two industries are definitely profiteering!

One is Intel's chip industry, and the other is telecommunications' communication industry!

The young master who installed the server looked at Xiao Ming, who hadn't slept well for a few days, and said secretly: "Build a live broadcast website?

Xiao Ming was speechless, "Then do you want to charge a member first?" ”

Master laughed twice and said, "No, no, no." ”