Source | Qubit | Public Account QbitAI
The legendary master He Yuming, the latest trend has been settled!
He just officially announced that he will join MIT EECS (Department of Electronic Engineering and Computer Science) in 2024.
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
He Yuming is a leader in the CV field. From CVPR's first Chinese best paper to ResNet, his Google Scholar has been cited more than 460,000 times.
As soon as the news came out, Nvidia AI scientist Jim Fan immediately sent a congratulatory message and said:
His achievements are too numerous to list, and some people are born for AI!
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
In addition, some netizens expressed envy that MIT students are blessed!
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
There are also some netizens who questioned the soul of Meta's brain drain:
Who is left in Meta?
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
The first person cited by MIT
In March of this year, He Yuming gave a job talk on the theme of "Visual Intelligence" at MIT (Job Talks), and revealed that in the future, he will focus on AI for Science, including vision and NLP to do self-supervised X+AI.
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
At that time, according to the description of the netizens at the scene, the venue was overcrowded, and it took a few turns to queue up. For this reason, MIT CSAIL temporarily opened the next meeting room to cast a screen.
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
This incident not only attracted widespread attention in the AI circle, but also attracted many netizens to speculate "whether he will join MIT".
Now, with He Yuming's official announcement, everything can finally be confirmed! And he will also become the most cited person in MIT.
Previously, Robert Langer, a heavyweight professor in the Department of Chemical and Biomedical Engineering, had the most citations in MIT, and the number of times on Google Scholar was 390,000+.
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
And now, He Yuming has been cited as high as 460,000+!
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
In addition, according to the official announcement, He Yuming will join MIT's Department of Electrical Engineering and Computer Science (EECS) in 2024.
EECS is MIT's largest academic department and a world-renowned research department in computing science and AI.
Many well-known figures in the field of computer science and AI have gathered here, including Hal Abelson, the leader of the MIT App Inventor experiment, and so on.
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
CV God
He Yuming studied in the Department of Physics of Tsinghua University for his undergraduate degree, studied for his Ph.D. under Tang Xiaoou, and graduated from the Chinese University of Hong Kong.
In 2009, Tang Xiaoou, He Yuming and Sun Jian won the CVPR Best Paper Award for their paper "Single Image Haze Removal Using Dark Channel Prior", which was the first time it was awarded to an Asian research scholar.
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
After graduating with a Ph.D. in 2011, he joined Microsoft Research Asia as a researcher, and joined Facebook AI Research (FAIR) in 2016 to continue researching computer vision.
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
In 2015, He Yuming proposed ResNet (Deep Residual Network), which not only won the first place in the ILSVRC 2015 classification task competition, but also won the best paper of CVPR in 2016.
ResNet is also currently its most cited research. The number of citations per article will exceed 100,000 by the end of 2021, and has now risen to 170,000.
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
During FAIR, He Yuming and his team achieved many outstanding achievements in the field of computer vision, including a series of researches such as Faster R-CNN and subsequent Mask R-CNN.
Among them, Mask R-CNN solves the problem of instance-level object segmentation in pictures. It can not only detect objects such as people and animals in photos, but also generate a high-quality separation mask for each object instance. This research also Won the best paper of ICCV 2017.
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
His recent main research interest is unsupervised learning. The MAE proposed at the end of 21 uses the mask pre-training method of the language model on the visual model, opening the way for large-scale unsupervised pre-training of large-scale visual models.
Recently, he also introduced the mask method into CLIP, the basic model of many AI painting applications, which increased the training speed by 3.7 times.
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
One More Thing
Recently, there has been a wave of returning to academia, and many FAIR research experts have left.
In September last year, ResNeXt announced that Xie Saining would leave as soon as he made it. He will join New York University as an assistant professor in January this year.
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
Xie Saining's main research direction is also "deep learning, computer vision". Currently, his citations on Google Scholar have exceeded 30,000+.
Before that, some netizens gave their own answers to "How do you think about Kaiming interviewing MIT faculty?"
Edit toggle to center
Add picture annotations, no more than 140 words (optional)
What do you think of this “return to academia” trend?
Reference link: [1]https://kaiminghe.github.io/ [2]https://scholar.google.com/citationsuser=DhtAFkwAAAAJ&hl=en/