Millions of dollars in annual salary + double-digit equity? Who is Sun Jian, who is rumored to have joined Face++?

I'm sure everyone has heard about a sky-high transfer case in the AI ​​academic community these days. The rumor goes like this: Sun Jian, chief researcher at Microsoft Research Asia, joined the domestic AI startup Face++ (Megvii Technology) with a sky-high transfer fee of "eight-digit annual salary + two-digit equity". Currently, Sun Jian himself, Face++ and Microsoft have not publicly confirmed the news.

In this regard, Shan Ting, CEO of SmartEye Technology, commented on this in his WeChat Moments:

"We will remain rational, consider issues from the perspective of solving practical problems, and recruit talents. However, in this industry, you can't do without money. It's better to spend money on talents and technology than to spend a lot of money on subsidies."

What this means is that for Shan Ting or Huiyan Technology, the money is worth spending.

Of course, I won’t discuss business matters here. There are too many cases of academic experts switching to enterprises recently, so you can just get used to it. But for the sake of gossip, let’s talk about who this “Mr. Millionaire” is.

Sun Jian

Sun Jian was born in 1977. After being admitted to Xi'an Jiaotong University at the age of 18, he stayed there for a total of ten years and obtained bachelor's, master's and doctoral degrees.

But the most unforgettable and glorious moment for him was when he was studying for his doctorate.

Sun Jian's doctoral major is pattern recognition and intelligent control, and as a joint graduate student of Microsoft Research Asia and Xi'an Jiaotong University, he studied under Dr. Harry Shum.

Speaking of this, we have to mention Dr. Harry Shum, who is a world-class expert in computer vision and graphics, and co-founder of Microsoft Research China (now Microsoft Research Asia).

That’s right, Sun Jian is such a lucky man.

Sun Jian, who has always been diligent and focused, naturally would not miss this opportunity. In the world-class scientific research center, he quickly upgraded from an ignorant student to a researcher in the academic world.

In 2001, the first project that Sun Jian and Dr. Shum did, using belief propagation for stereo reconstruction, was a big success. He became the first person in the industry to use Bayesian belief propagation to solve stereo vision problems, which was the best solution at the time. With such outstanding performance, Sun Jian won the "Microsoft Scholar" award that year.

At the same time, the academic atmosphere of Microsoft Research Asia also successfully captured Sun Jian's heart.

After graduating with a doctorate in 2003, he resolutely chose to continue his scientific research here, focusing on interactive computer vision and network computer vision.

A year later, Sun Jian participated in a graphics research project for the first time as a doctoral student, and gradually gained fame in the industry. Soon he became a senior researcher at Microsoft Research Asia.

In 2010, 33-year-old Sun Jian won the "Global Outstanding Young Innovator Under 35" award from MIT Technology Review, an authoritative technology journal.

For Sun Jian, who is able to endure loneliness, he basically maintains the pace of publishing papers every year. He has published more than 50 academic papers in conferences such as CVPR, ICCV, ECCV, and SIGGRAPH.

So what is the value of these papers?

Take CVPR, ICCV, and ECCV for example. They are the top three conferences in computer vision, and the acceptance rate of papers is basically maintained at around 20%. What is more worth mentioning is that Sun Jian’s published papers have won many awards:

In 2006, it was shortlisted for one of the top 100 doctoral dissertations in China;

Won the CVPR Best Paper Award in 2009;

At the just-concluded CVPR 2016, he won the best paper award for his paper "Deep Residual for Learning Image Recognition".

Here are his research results:

In 2015, Sun Jian, He Kaiming, Zhang Xiangyu and Ren Shaoqing, three researchers, used rectifier neurons to develop a "computer vision system based on deep convolutional neural networks (CNN)" that reduced the error rate in the ImageNet2012 classification data set to 4.94%, while the error rate of human eye recognition was about 5.1% in the same experiment. It is understood that this data set contains about 1.2 million training images, 50,000 verification images and 100,000 test images, divided into 1,000 different categories. This is the first time that the ImageNet1000 challenge has surpassed the ability of humans to identify and classify objects.

At the end of 2015, their "Deep Residual Learning for Image Recognition" system defeated the systems of Google, Intel, Qualcomm and academic laboratories in multiple categories of ImageNet tests, becoming the system with the highest recognition accuracy (classification error rate of 3.5%). The neural network layer of this system reached 152 layers, while the general level is only 20-30 layers.

In addition, he also participated in the development of a famous "age robot "How-Old.Net". I believe everyone has experienced this product. It can quickly locate, detect, identify and judge age based on photos uploaded by users, combining computer vision methods with powerful cloud support.

If Sun Jian, who is such a man, earned tens of millions of yuan in annual salary, would you still not accept it?

BTW, in August this year, Leifeng.com will hold the unprecedented "Global Artificial Intelligence and Robotics Summit" (CCF-GAIR) in Shenzhen , and invite top domestic and foreign artificial intelligence experts and investment, academic and corporate circles to discuss the technology and market trends in the field of artificial intelligence. If you also want to participate, please click Read the original article ! Buy the midsummer ticket now for only 35% off!