He Kaiming's MIT job search speech: there was a queue three hours in advance, over 100 pages of PPT reviewing CV development, and "the great god is half a month old"

Yang Jingyang comes from Aofei Temple
Qubits | Public account QbitAI

He Kaiming's MIT job search speech really became a large-scale star-chasing scene in the AI ​​circle!

According to news from audience friends who were present at Qubit, some students were already queuing up at the door three hours in advance.

Half an hour before the speech started, the queue at the door was said to have been bent several times...

△ Subsequent correction, the PPT will be more than 129 pages

Come and feel the atmosphere of the scene:

We have talked about the news of He Kaiming returning to academia before. This time the Job Talk representative He Kaiming has at least passed MIT's resume screening, and his research results and abilities have been initially recognized.

For this speech, He Kaiming also prepared over 129 pages of PPT, reviewing his work in the CV field, covering ResNet, Faster RCNN, Mask RCNN, MoCO and MAE.

When talking about future work directions, He Kaiming also mentioned AI for Science.

In addition to being concerned about the great master’s speech itself, some netizens also noticed He Kaiming’s latest changes

He Kaiming returns to academia

At present, it is not yet certain whether He Kaiming will successfully join hands with MIT. But this speech was the first step in his quest for a teaching position.

Some netizens discovered that if He Yuming finally succeeds in joining the company, he will become the most cited person at MIT.

Currently, the most cited person at MIT is Robert Langer, a heavyweight professor in the Department of Chemical and Biomedical Engineering, with more than 380,000 citations.

He Kaiming has been cited more than 400,000 times .

Among them, He Kaiming's most out-of-the-box research is none other than ResNet , which exceeded the 100,000 mark at the end of 2021 and has now risen to 150,000.

Although ResNet itself is researched in the field of computer vision, its core idea, residual connection, has crossed over and become a basic component of modern deep learning models .

AlphaGo Zero, which started the last AI boom, was completed by combining ResNet + reinforcement learning + Monte Carlo search. In ChatGPT, which started the latest AI craze, the "T" in it means that residual connections are also used in the Transformer network.

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

His main recent research interest is unsupervised learning. The MAE proposed at the end of 21 uses the mask pre-training method of language models on visual models, paving the way for large-scale unsupervised pre-training of large visual models. Recently, he also introduced the masking method into CLIP, the basic model of many AI painting applications, increasing the training speed by 3.7 times.

However, some anonymous Zhihu users expressed their opinions from the scene: The quality of the entire Talk was not good, and He Yuming did not use a good story to connect the work.

Does this mean that even He Kaiming doesn’t know how to tell a good CV story?

So, are you there this time? Let’s take a look at the comments section~

click here