Ohne Angst vor Werbung von anderen Firmen zeigt Google ein Gruppenfoto seiner IMO-Medailleträger und markiert nacheinander deren Kontaktinformationen.

Google DeepMind doesn't believe in evil omens.

Not long ago, Google's team won the gold medal at the IMO, but then Mark Zuckerberg quickly poached three core team members.

Such a dramatic and frustrating situation! Many people advised Google DeepMind to be more cautious. Some even joked that they should sign their papers anonymously in the future. If that doesn't work, they could follow Chinese Internet companies and use nicknames...

Unexpectedly, Google's IMO gold - medal team, not afraid of the odds, directly posted a group photo of the team. It seemed like a response and also a bit of a provocation.

Thang Luong, the leader of Google's IMO gold - medal team and the head of DeepMind's Super Reasoning Team, even tagged everyone in the photo.

This is like handing a knife to Mark Zuckerberg... (Just kidding)

In addition, he also shared that they gathered from all over the world before the opening of the IMO to witness this "miraculous moment".

Interestingly, the three Chinese members poached by Mark Zuckerberg were not in the group photo.

Google's Large Model Races for the IMO: A "Mathematical Miracle" in One Week

Team leader Thang Luong further revealed the details of their preparations this time.

On the eve of the IMO 2025, Google DeepMind's scientists did something "legendary".

Core members from places like Mountain View, New York, and Singapore gathered at the London headquarters for a final push to win the IMO gold medal. As core member Thang Luong said:

I believe that week was when all the miracles happened.

They integrated all the "formulas" they had discovered before and carried out an all - out intensive training.

It's worth mentioning that to run this model, they even had to borrow computing resources from multiple teams, which left Thang Luong quite helpless.

Two days before the end of the IMO, the team finally completed the final training of Gemini Deep Think, bringing its performance to its peak.

Thang Luong commented on it:

Its performance is not only amazing in mathematical reasoning but also incredible in code generation and other complex reasoning tasks!

Core Members of the IMO Gold - Medal Team

In the just - announced list of 16 members, there are 4 Chinese members, and the three who were poached by Mark Zuckerberg are not among them.

Here is a detailed introduction to some of the core team members.

Yi Tay, the co - leader of the Deep Think IMO gold - medal team, was an important participant in Google's large models such as PaLM, UL2, Flan - 2, and Bard.

Interestingly, at the end of March 2023, he announced his departure from Google to start a business at RekaAI and served as the company's chief scientist.

Before that, he worked at Google Brain for 3.3 years, participated in the research of many large models, and wrote about 45 papers in total, 16 of which were first - authored, including UL2, U - PaLM, DSI, Synthesizer, Charformer, and Long Range Arena. He was involved in about 20 product releases. He is truly a top expert.

Unfortunately, due to his wife's pregnancy at the beginning of his startup, he had to take care of both the business and his family, which took a great toll on his physical and mental health. For various reasons, he decided to return to Google.

Quoc Le, one of the founders of Google Brain, graduated from the Australian National University with a bachelor's degree and then pursued a Ph.D. in computer science at Stanford University, under the supervision of Andrew Ng.

Andreas Kirsch, currently a research scientist at Google DeepMind, obtained his bachelor's and master's degrees from the Technical University of Munich and then his Ph.D. in computer science from the University of Oxford.

Theophane Weber holds a Ph.D. in operations research from MIT. In the early years, he also worked as a research scientist at Lyric Labs of Analog Devices.

Dawsen Hwang obtained his bachelor's degree from National Taiwan University and his master's and Ph.D. degrees from MIT. He has been interested in solving mathematical problems since childhood and participated in the IMO in 2003 and 2004, winning a silver medal and a gold medal respectively.

Jieming Mao graduated from Tsinghua University with a bachelor's degree and then obtained a Ph.D. from the Department of Computer Science at Princeton University. He is a post - doctoral researcher at the Warren Center of the University of Pennsylvania.

Jonathan Lee obtained his bachelor's degree in electrical engineering and computer science from the University of California, Berkeley, and his Ph.D. in computer science from Stanford University, specializing in reinforcement learning.

Vinay Ramasesh obtained his bachelor's and master's degrees from MIT and a Ph.D. in experimental physics from the University of California, Berkeley, in 2019. He has been dedicated to researching quantum processors based on superconducting qubits.

Lei Yu completed her undergraduate education at the University of Nottingham, obtained a master's degree in philosophy from the University of Cambridge, and finished her doctoral research at the University of Oxford. She joined Google DeepMind as a research scientist in 2017.

Zicheng Xu studied at the High School Affiliated to Beijing Normal University, with a perfect GPA. Then he obtained a double degree in computer science and mathematical economics from Rice University and plans to pursue a Ph.D. at Johns Hopkins University.

Now that so many talented people have emerged, is Mark Zuckerberg ready to make a move?

Reference link: https://x.com/lmthang/status/1948458590492393834?s=46

This article is from the WeChat official account "QbitAI", author: Shiling. It is published by 36Kr with authorization.