The legendary god He Yuming was exposed to return to the academic world, netizens: to fulfill the highest citation of MIT – IT HOME

AI DanielHe KaimingWith the latest trends, and thereturn to academia.

MIT CSAIL The laboratory announced that next Monday, March 13, He Yuming will give an academic lecture at MIT.

This matter has aroused widespread concern in the AI ​​circle.Under related Zhihu questions, Dr. MIT @Charles pointed out that the seminar He Yuming participated in was all about Job Talks(Job Speech).

Chen Yiran, a professor at Duke University, also said that he has been hearing that He Yuming is looking for a teaching position, and now he can finally confirm it.

Job SpeechIt is a great tradition in academia, which refers to applicants for doctoral degrees or faculty members coming to the school to give academic reports, the content of which is their own research results, usually concentrated in March and April every year.

And entering this stage means that He Yuming has at least passed the resume screening, and his research achievements and abilities have been recognized by MIT.

Some netizens predicted that since Li Feifei’s lecture at MIT was full of seats, He Yuming’s lecture may also be very popular this time, becomingDeep learning large-scale star chasing scene.

However, some netizens added in the comment area that this does not mean that He Yuming will choose to go to MIT in the end, or he may go to other schools, but in short, he is basically sure that he is looking for a teaching position.

Or become the most cited person in MIT

Although he is not sure about He Yuming’s final choice, sharp-eyed netizens have discovered that if he really went to MIT,Citationswill beHighest in the school.

At present, Robert Langer, a heavyweight professor in the Department of Chemical and Biomedical Engineering, has the highest number of citations in MIT, with a number of 380,000+.

△ Source: Google Scholar

And He Yuming has been cited as high as 400,000 +.

△ Source: Google Scholar

Although different disciplines cannot be directly compared in this way, the number of citations of He Yuming is really eye-catching.

Some netizens said that although the number of citations may not necessarily explain too many problems, “the one with an astronomical number of citations must be a great god.”

To discuss He Yuming’s most cited research, it must be ResNet. It broke through the 100,000 mark at the end of 2021, and has now risen to 150,000.

Although ResNet itself is a research in the field of computer vision, its core idea isresidual connectionhas crossed over to becomeBasic Components of Modern Deep Learning Models.

What kicked off the last AI boom AlphaGo Zero It is a combination of ResNet + reinforcement learning + Monte Carlo search.

And the one that kicked off the latest AI craze ChatGPTwhere “T” means that the residual connection is also used in the Transformer network.

In computer vision, He Yuming’s main contributions also include a series of researches such as Faster R-CNN and subsequent Mask R-CNN, which have been the mainstream method of target detection for many years.

His recent main research interests areunsupervised learningthe MAE proposed at the end of 21, uses the mask pre-training method of the language model on the vision model, opening the way for large-scale unsupervised pre-training of large-scale visual models.

Recently, he also introduced the mask method into the basic model CLIP of many AI painting applications, which increased the training speed by 3.7 times.

In MIT’s upcoming speech, He Yuming’s topic is also “Visual Intelligence”.

Including his achievements in ResNet, object detection, and the impact of visual unsupervised learning on future research.

As for He Yuming who might join MIT CSAIL (MIT Computer Science and Artificial Intelligence Laboratory), is the largest laboratory of MIT and a world-renowned computer science and AI laboratory.

Many well-known figures in the fields of computer science and AI have gathered here, including Tim Berners-Lee, the inventor of the World Wide Web, experts in distributed computing and concurrent algorithms, Turing Award winners Leslie Lamport and Barbara Liskov, etc.

When Meta AI has product pressure

He Yuming studied at Tsinghua University as an undergraduate and a Ph.D. from the Chinese University of Hong Kong. In 2011, he joined Microsoft Research Asia.

Since 2016 he joined Facebook AI Research, also known as FAIRcontinue to study computer vision.

However, within the last year, there have been two major adjustments in this department, both of which areEnhance applied research and reduce basic research.

In June last year, affected by the company’s transformation of the metaverse, FAIR was downgraded to a subsidiary organization of the metaverse department Reality Labs.

At the same time, the algorithm team supporting its major apps was transferred to the product engineering team. The new architecture gives the product team greater authority and accelerates the application of technology.

At the end of February this year, Zuckerberg made another adjustment in the face of the ChatGPT boom, announcing the establishment of a top AI product team.

Gather all the people involved in AIGC in the company.

Many years ago, when LeCun removed the FAIR management and became the chief AI scientist, researchers such as Tian Yuandong and Wu Yuxin who were still at FAIR said that in this departmentno product pressure.

However, after a series of recent adjustments, Meta has lost many research talents, including at least 4 well-known scholars such as Edward Grefenstette who led reinforcement learning research, and even the AI ​​laboratory in London lost most of its top researchers.

Judging from He Yuming’s return to academia, he is still more inclined to engage in basic research.

One More Thing

During this time, the comparison between academia and industry once again became a hot topic in the AI ​​circle.

On the one hand, some netizens pointed out that big industrial companies are having a hard time, engineering and application-oriented jobs are increasing, and research-oriented jobs are decreasing.

On the other hand, large-scale models dominate, and the cost of AI scientific research has increased exponentially. Even the top journal Science published an article in the main issue stating that the computing power mastered by academia and industry is no longer of the same order of magnitude.

Under the attack of the two, AI experts returning to academia and leaving big factories to start a business have become two trends that exist at the same time.

In addition to He Yuming, representatives of returning to academia include Pete Warden, a founding member of Google Tensorflow. In China, Qi Yuan, former vice president of Ant Financial, joined Fudan University, and Dai Jifeng, former executive research director of Shang Tang, joined Tsinghua University.

The representative figures who left the big factories to start businesses include Google Transformer authors who left to establish Adept AI and Character.AI, and Amazon Li Mu and Alex Smola who just revealed two days ago joined hands to start a business.

Under the Zhihu hot topic of “How do you think about Kaiming interviewing MIT faculty?”, a highly praised answer is:

Reference link:

  • [1]https://www.csail.mit.edu/event/eecs-special-seminar-kaiming-he-pursuit-visual-intelligence

  • [2]https://weibo.com/2199733231/4877346375735190

  • [3]https://www.zhihu.com/question/588205714