Quantum computing master Aaronson joins OpenAI! He is also the mentor of Chen Lijie, the top student in Yao class

Big bosses from major domestic companies have left their jobs, switched to universities, and started their own businesses.

What is different is that quantum computing guru Scott Aaronson is now moving from the teaching position to a large company!

Today, Aaronson stated on his blog that he will take a year off from the University of Texas at Austin (UT Austin) and go to OpenAI next week Work.

His job responsibility is to think about the theoretical basis of artificial intelligence safety and alignment (AI Safety and Alignment).

This includes thinking about things like "the principles of computational complexity for how to make artificial intelligence do things we want to do, rather than do things we don't want to do." What’s your understanding?”

Jan Leike, a machine learning researcher at OpenAI and leader of the Artificial Intelligence Alignment Team, said, "I'm really looking forward to working with the legendary Scott Aaronson."

Open AI welcomes a quantum computing expert

It can be said that OpenAI welcomes a "work from home" person BIG Name.

how to say?

Scott Aaronson said, “For family reasons, I will be doing this primarily from my home in Texas, but will also travel from time to time. OpenAI office in San Francisco."

While working at OpenAI, Aaronson will spend 30% of his time continuing to manage Austin University's Center for Quantum Information, working with his students and postdocs.

By the end of this year, Aaronson plans to return to full-time teaching, writing, and thinking about quantum problems. In other words, he went to OpenAI just to experience working life for a year.

#For Aaronson, quantum issues remain at the forefront of his life even as artificial intelligence rules the world in ways that none of us can ignore. Hobby. Artificial intelligence was an area Aaronson began researching as a doctoral student before he turned to quantum computing.

In other words, what kind of project does Scott Aaronson want to do at OpenAI?

He admitted that he had no idea for the time being, so he needed to spend a whole year thinking about it and put forward several possibilities. .

First, he might arrive at a general theory about sample complexity for learning in dangerous environments.

Secondly, one might work on interpretability of machine learning: when given a deep neural network that produces a specific output The network provides an explanation for why it produced that output; what can we say about the computational complexity of finding that explanation?

Third, the ability of weak agents to verify the behavior of strong agents may be studied.

Some netizens asked directly, should you be worried that OpenAI is just hiring you to say, "Look, we have Scott Aaronson solving this problem, and What did security researchers come up with who didn’t really care about it?”

Scott Aaronson said, "I can't prove the issue you are worried about. No matter what work I do on this topic, I have to speak for myself. .”

Scott Aaronson is currently the David J. Bruton Jr. Centennial Professor of Computer Science at the University of Texas at Austin and serves as the founding director of the school’s Quantum Information Center. .

Aaronson received a bachelor's degree in computer science from Cornell University, a Ph.D. from the University of California, Berkeley, and worked at the Quantum Computing Institute at the University of Waterloo in Canada. Postdoctoral researcher.

#Previously, he taught electrical engineering and computer science at the Massachusetts Institute of Technology (MIT) for nine years.

He taught at MIT from 2007 to 2016, serving as an assistant professor in the fall of 2007, and was promoted to associate professor in the spring of 2013. Until 2016, he taught at the University of Texas at Austin as a full professor.

Chen Lijie, a top student in Yao class, studied under Aaronson during his exchange at MIT.

##Photo source: Tsinghua University

Scott Aaronson is no ordinary person.

#In 1981, Aaronson was born in the United States.

# His childhood experience was relatively rich. Although he has lived in the United States since he was a child, his father was sent to work in Hong Kong when he was a child. As a result, Aaronson also spent some time in Asia.

#At that time, he showed his intelligence in schools in Asia - he skipped a grade.

Unfortunately, due to acclimatization or some other reason, his path to study became very bumpy after returning to the United States.

# He often clashed with teachers, and his grades became unsatisfactory.

Finally, he enrolled in Clarkson School, a program for gifted young people run by Clarkson University, which allowed Aaronson to apply to college during his freshman year of high school.

It was also because of this opportunity that he was admitted to Cornell University and received a bachelor's degree in computer science in 2000.

#After receiving his degree, he did not give up his studies and continued to study for a doctorate at the University of California, Berkeley. Finally, in 2004, he received his PhD from Professor Umesh Vazirani.

#Actually, Aaronson’s skill points have been maxed out since he was young. His mathematical ability was unusually good compared with his peers. He taught himself calculus at the age of 11.

Even after he discovered computer programming when he was 11 years old, he regretted that he had not started to get in touch with it earlier. He felt that he had already become familiar with "programming". Many years of "peers" have fallen behind too much.

#After that, he clicked on the branch of quantum computing on the higher-level skill tree. At Cornell, he worked on two components: computational complexity and quantum computing.

# His hard work and talent also gave him enough rewards.

In April 2021, the Association for Computing Machinery (ACM) awarded the 2020 ACM Computing Award to Aaronson in recognition of his contributions in the field of quantum computing contribute.

Specifically, his research areas include the performance and limitations of quantum computers, and the broader computational complexity theory wait.

ACM introduced that the goal of quantum computing is to use the laws of quantum physics to construct devices to solve problems that cannot be solved by classical computers or cannot be solved in any reasonable time. The problem.

And Aaronson shows us how findings from computational complexity theory can provide new insights into the laws of quantum physics and clearly illustrate how "quantum computers What can and cannot be done."

#Not only that, Aaronson also helped develop the concept of "quantum supremacy." Quantum supremacy refers to a "milestone" being reached when a quantum device can solve a problem that cannot be solved by a classical computer in a reasonable amount of time.

Aaronson established the theoretical basis for many quantum supremacy experiments. This kind of experiment allows scientists to give compelling evidence that quantum computers can provide exponential speedups without having to first build a complete fault-tolerant quantum computer.

ACM President Gabriele Kotsis said, "Aaronson's contribution is not limited to quantum computing, but has also had a significant impact in fields such as computational complexity theory and physics. ."

#It is worth mentioning that Aaronson is also the author of "Quantum Computing since Democritus".

His personal blog "Shtetl-Optimized" often answers some questions about quantum computing from a popular science perspective. Very popular.

# He wrote "Who Can Name the Greater Number?" , which has been widely circulated in the computer science academic community, uses the concept of Busy Beaver Numbers described by Tibor Radó to illustrate the limitations of computability in teaching settings.

#Now that such an outstanding person has come to OpenAI, it can be said that a talented general has arrived.

Hot comments from netizens

After Scott Aaronson announced that he would join OpenAI, many netizens sent him blessings for a smooth work.

To commemorate this moment, netizens used DALL·E to generate a painting for Scott Aaronson.

Some netizens discussed the issue of artificial intelligence alignment with him,

Please explain how artificial intelligence aligns with human values ​​when humans themselves identify with what those values ​​are? And more often than not, humans are not consistent with the values ​​they profess.

To me this proves that the fundamental category of thought about alignment or safety or whatever is not moral philosophy, nor computational complexity, but Evolutionary Theory. That is natural selection. Since evolution has a mathematical basis (see John Baez), I think there may be some intersection with comparative complexity.

While I am not an expert in either field, the question of whether artificial intelligence has agency or consistency may be important from an evolutionary perspective, but Not the core.

## Another netizen who admires Aaronson’s course asked, “Will you also teach quantum at UT in the 2022-2023 academic year? Information science course? I’m really looking forward to taking this class!」

The above is the detailed content of Quantum computing master Aaronson joins OpenAI! He is also the mentor of Chen Lijie, the top student in Yao class. For more information, please follow other related articles on the PHP Chinese website!