Black Technology Research Center

Although Fang Yuan is on fire, he has not yet "floated".

At his current level, he can be said to be outstanding among undergraduates!

Especially he has practical experience that other undergraduates do not have.

But compared to those real masters, his strength is still not enough.

But being young is his greatest capital, and the next few years will be a period of rapid progress for him.

This period may be the period of accumulation in the fantasy novel.

The foundation accumulated over the past ten years will allow him to make the greatest progress during this time.

Unlike others, Fang Yuan strengthened his own potential with the help of the system while releasing his potential.

In layman's terms, Fang Yuan's rapid progress lasted longer than others.

In other words, Fang Yuan could go further and become stronger on this road.

Other aspects of translation software are very easy to solve, after all, Fang Yuan has experience in writing qoqo.

But one thing is not very easy to solve, and that is the establishment of language database and thesaurus.

Why build these two things?

Let's start with the shortcomings of translation software on the market now.

Existing translation software can only literally translate the text or speech we input, which means they will not translate according to context.

Especially in terms of speech, the recognition rate of some dialects with existing technology is really a headache.

But the software that Fang Yuan wants to do can perfectly solve this problem.

The key is the establishment of the two databases mentioned above.

Not only need to establish these two databases, but also need to associate the two.

In addition to these two databases, there is a matching database that Fang Yuan needs to complete.

The relationship here is very complicated.

This intelligent translation software first identifies your current use scene (the most frequently used scene by default) through the matching library.

Then split the text or voice you input based on the data in the language database.

Use the split small units to match in the lexicon.

And reorder them according to the grammatical rules of the translated language.

In addition, some optimizations have been made in the voice input, and the dialect can be well recognized through the data in the language database.

After all, the most important thing is the establishment of the language library.

In this regard, the system has also figured out a way for Fang Yuan, which is to obtain data from short video websites, live broadcast websites, and self-media websites through crawlers.

Of course, this process cannot be completed in a short time.

Moreover, the requirements for computer configuration are very high, and Fang Yuan’s current X Starren computer simply cannot support the needs of this software.

Fortunately, the last time he went to the "Jiuzhou" company, he heard Liu Zhijun say that the company currently has several servers that are temporarily vacant!

This also saves Fang Yuan a lot of effort.

After picking up the phone and explaining the situation to Liu Zhijun, the right to use these servers was temporarily handed over to Fang Yuan.

In fact, the translation of Chinese is a headache for all translation software.

After five thousand years of history, the semantic meaning of Chinese is the most complicated of all languages.

Especially in the Northeast dialect, which is very popular nowadays, it is very difficult to establish the corresponding thesaurus for many words.

Here I have to feel the power of the system!

In the whole program, the semantic matching algorithm is Fang Yuan's favorite place.

Here, a mathematical model is used to establish a text and semantic matching model.

All words are equivalent to a model, and how many words there are in a sentence is equivalent to how many models are connected.

In this way, you only need to process each model to get the meaning of the word in this passage.

It's complicated to say, in fact it means to associate all the meanings of each word, and then find the corresponding meaning in the thesaurus.

Fang Yuan is not in a hurry for the next job. After training in the AXM competition, his progress in programming is visible to the naked eye.

He put the written crawler program out, and then established a connection with the server.

Then just wait, the data waiting to be collected can support the establishment of the language library.

At that time, it was the time to complete this software.

In addition to this work, the most important thing for Fang Yuan now is to improve the learning algorithm of this intelligent software.

This is not a simple matter. It is compared with qoqo's learning algorithm.

Fang Yuan found that the learning algorithm of this software really needs improvement.

If you improve them one by one, you might as well rewrite an algorithm to save trouble!

Easier said than done. Looking at things that are not troublesome, Fang Yuan took the most time to complete.

For the other parts, Fang Yuan only needs to compile according to the programmers given by the system, and even the learning algorithms given by the system have been written by Fang Yuan.

For nothing else, Fang Yuan should do the same even if this translation software can be launched quickly.

As for improving the learning algorithm, it is not the most important thing at present.

Fang Yuan planned to launch a version first, and then through an update method, the learning algorithm was replaced without any knowledge.

However, there must be no less work. Fang Yuan intends to complete the algorithm upgrade before the language database is established.

...

"Fang Yuan, Fang Yuan! I didn't expect your kid to steal so much light for our 333 bedroom!" said Zhang Hua, who had just returned to the bedroom!

"That's it, don't forget the brothers in our dormitory if there is such a good thing in the future!" Sun Wenzheng followed with a roar!

"That's necessary. If there is a good thing, can I still miss our 333 brothers?" Fang Yuan patted his chest, his face was full of expressions that you lie down and I can carry the audience.

Not to mention Fang Yuan's roommates, the voice assistant qoqo has become one of the most popular applications on the Internet!

The reason is that a guy who made the evaluation video compared all the voice assistants on the market.

Whether it is the SIRI of a certain fruit, the voice assistant of a certain flower, or the Xiaoai classmate of the most popular one on the Internet in the past, compared with this qoqo released by an unknown Kyushu company.

It's like the difference between genius and mental retardation. Don't get me wrong. The mental retardation here is the voice assistant of the top three major manufacturers, and the genius is the less famous but widely acclaimed qoqo.

Not to mention, this guy who did the assessment really has a certain influence.

It may be because his previous evaluation videos were fair!

After his propaganda, the download of qoqo increased sharply, and the result was that the information of this humble small company in Kyushu became a frequent visitor on the desk of some big factory bosses.

After professional testing of qoqo, no matter it is flower factory, rice factory or some other manufacturers, all have great interest in this software.

Even fruit factories, which have never been optimistic about the innovation ability of Chinese people, have begun to pay attention to qoqo.

What they are most interested in is not qoqo's stronger functions than their SIRI, but why qoqo can obtain the same permissions as their own system software.

They had planned to ban qoqo from installing on the fruit system the first time, but their curiosity prevented the fruit factory from making such a decision in the first place.

They decided to ignore the software before getting the other party to get their system permissions.

After all, for a closed-loop fruit system, preventing the birth of software like qoqo is the most important thing.

Qoqo's current situation can only be described by Li's adoration.

For collection and recommendation, thank you all.