Chapter 45: Director Ao at Full Firing (Part )

The next day, while Lin Feng was quietly developing the recommendation algorithm, he was waiting for Daxia Ordnance Industry and overseas military companies to come to him for negotiations.

By noon of the next day, many UP hosts within Bilibili who were unhappy with or jealous of Lin Feng had already started to take action, and the leader was the famous Director Ao.

Speaking of it, Director Ao is really a veteran of Bilibili, although his core video website is Youku, after all, Youku has more users.

But he has also been paying attention to Bilibili and will simultaneously post his videos on Bilibili. It can be said that his contribution to Bilibili is really great.

In the past, when netizens of Bilibili evaluated the five UP hosts who had made great contributions to the development of Bilibili, Director Ao was one of them, which shows how great his contribution to Bilibili is.

Let's get back to Director Ao. Previously, Director Ao saw a new UP host gain 40 fans in just one day, which was faster than his rise. This made him very jealous.

But because he was a well-known public figure, he was reluctant to speak out when others posted videos making ghost videos of Raymond Lam and accusing him of doing so.

Seeing that Lin Feng was doing better and better, and was even about to obtain 30% of Bilibili's shares with the help of a bet agreement and a recommendation algorithm, Director Ao felt more and more unbalanced in this situation.

Comparing yourself with others will only make you angry. I have more experience than you, so why have you risen so quickly and gained so much attention and fans?
With such jealousy and unwillingness in his heart, Director Ao started staying up late that night to collect information on various recommendation algorithms. In the end, the more he collected, the bigger the smile on his face became.

So after most of the night and a morning of hard work, a carefully produced video was officially released on video website apps such as Youku, Tudou, Penguin, A站, and B站.

"Hello everyone, I am Director Ao, we meet again.

A lot of things have happened to Bilibili in recent days. First, the website and APP were revised, graphic ads were launched, and finally a live broadcast platform was launched...

What surprised me the most was that Lin Feng actually signed a gambling agreement with the shareholders of Bilibili. Director Ao also watched the live broadcast last night.

During this process, Director Ao noticed that Lin Feng claimed that the recommendation algorithm was worth 20% of Bilibili’s shares, and thought that everyone should be very curious about this matter.

So, after staying up all night to collect information and organizing it the next morning, Director Ao lived up to the trust and produced this video. "

When the video reached this point, the comments on Bilibili were filled with comments like "Director Ao has worked hard", "Director Ao is so dedicated", and "Director Ao should take care of your health".

It can be seen that Director Ao still has a lot of fans on Bilibili. Those fans like his videos very much and also care about and love him very much.

"Okay, let's get to the point.

Our main purpose in this video is to introduce the recommendation algorithm, and then to deduce whether the recommendation algorithm mentioned by Raymond Lam is really worth 20% of Bilibili’s shares.”

After hearing this, Ao Changzhang's fans became more and more excited. They were very curious about the mysterious recommendation algorithm.

On the other hand, the proud director in the video continuously introduced the recommendation algorithm:

“First of all, the recommendation algorithm originated from the Palo Alto Research Center in 1992, where they developed a system based on collaborative filtering algorithm and used it for spam filtering.

Then in 1994, the University of Massachusetts and NISU launched a recommendation system for news based on the collaborative filtering algorithm, which can recommend news based on users' news ratings.

However, the recommendation algorithm was not really applied to the Internet until 2003, when it was applied by e-commerce platforms, which used collaborative filtering algorithms to recommend similar products. Then in 2006, Douban and Netflix launched the world-changing matrix decomposition, where users and products were given corresponding latent vectors, with strong generalization capabilities.

From then on, the products recommended by e-commerce companies are more targeted, and the probability of purchasing successfully recommended products is greatly increased.

Finally, in 2010, the FM model (factorization machine), a machine learning model, was proposed by Osaka University, and the recommendation algorithm officially entered the era of machine learning.

It is particularly suitable for dealing with feature interaction problems in sparse data sets. Since then, the push of products and advertisements has become more accurate and efficient. After two years of development in 2012, it officially dominated the field of recommendation systems.

It is worth mentioning that with the launch of AlexNet convolutional neural network last year, many people have also set their sights on neural networks.

They want to rely on neural networks to develop new recommendation algorithm models and open up new technology development lines.

But let’s not talk about the recommendation algorithm model based on neural networks for now. This is a brand new technology line and no finished product has been launched yet.

Just taking the now quite sophisticated FM model as an example, it can actually achieve what Raymond Lam said, that is, accurately deliver videos that users may be interested in to consumers.

So the recommendation algorithm is not something magical. It has existed for more than 20 years and has become quite mature now.

Among these, I guess Lin Feng should have developed the recommendation system of Bilibili based on the FM model, and ultimately achieved the goal of recommending videos.

This kind of technology is actually not that exaggerated. To be honest, there are at least hundreds of people in China who can develop such recommendation system technology, and there are even more overseas.

Lin Feng claimed that his recommendation algorithm is worth 20% of Bilibili’s shares. I think this is actually a bit too confident. Sometimes confidence is a good thing, but being too confident is arrogance.

Of course, we can understand why Lin Feng is so confident. After all, Lin Feng said last night that his algorithm research and development has not officially started yet.

It's like when we were in school, we always said that we would have an annual salary of $100,000 in the future and would buy a car and a house. But after we entered society and started working, we realized how naive we were.

At this time, Lin Feng’s algorithm has not been formally studied yet. It is normal that he does not know about the algorithm. We need to understand Lin Feng.”

Director Ao has been speaking well of Raymond Lam, and there is not a single dirty word in the entire text. However, from the perspective of the video viewers, everyone knows that Director Ao is secretly mocking Raymond Lam for being too arrogant.

This naturally led to polarization in the video barrage and comment sections, as both sides have huge fan bases.

Raymond Lam gained fame through the invention of a writing robot, a coin-fired electromagnetic rifle, two highly-anticipated documentaries and the film I Love Inventions, and has attracted millions of followers.

Although Director Ao's experience in recent years is not as exciting as Lin Feng's, he has accumulated millions of followers in the few years since he started making videos in 2009.

The number of followers on both sides is not much different, but in terms of the number of loyal fans, Director Ao has more. After all, after watching hundreds of videos, people have already become Director Ao's die-hard fans.

So Lin Feng’s fans were quickly defeated and were completely unable to defeat Ao Changzhang’s fans. After all, how can you defeat others with just one mouth when they have several mouths?
　I received the second round of recommendation notification this afternoon. The author was so happy that he decided to add more chapters to thank his brothers for their support.

　　I'm also asking for recommendation votes and monthly votes.

　　
　
(End of this chapter)