Quantum Q&A with PhD Student Hamoon Mousavi
Before he left for the Quantum Information Processing meeting, we asked Mousavi about combining quantum mechanics with computer science and how endurance sports help keep his mind fit.
Hamoon Mousavi is a PhD student who moved to Columbia from the University of Toronto last year with Henry Yuen, whose research group studies theoretical computer science and the differences between classical and quantum computers.
In 2020, Yuen, now an assistant professor in the Department of Computer Science at Columbia Engineering, co-published MIP*=RE, a paper proving that quantum entanglement, the communication-free coordination between spatially separated entities, can help solve incredibly complex problems. Those entities could be quantum particles, or Alice and Bob, the players in what’s known as a nonlocal game. This is a physics thought exercise from the 1960s to prove the existence of entanglement that was confirmed by experimentalists 30 years later.
Computer scientists in the mid-2000s saw parallels with their work on multiprover interactive proof systems; these are theoretical concepts important for cryptography that solve problems but are inherently untrustworthy. In classical computing, provers are isolated and cannot share information, so a verifier—an entity with limited problem-solving ability—will ask multiple provers for an answer and then compare those to verify the correct solution. Playing nonlocal games can in turn help a verifier tackle even harder problems.
Quantum entanglement can increase the power of provers if the untrustworthy entities can be kept from conspiring against the verifier, but it also makes it hard to approximately calculate the probability of winning a nonlocal game. In a recent preprint with Yuen and Seyed Sajjad Nezhadi that builds on MIP*=RE, Mousavi explored how much harder it is to exactly calculate this winning probability. The work reveals new details about competing models of quantum mechanics.
Mousavi was invited to present a Plenary Talk about the work at the Quantum Information Processing (QIP) meeting, the largest conference for researchers interested in quantum information. We caught up with Mousavi before his flight to Pasadena about how quantum mechanics and computer science collide and what he does in his spare time now that he’s in New York.
What are you focused on for your PhD?
With Henry, I study how quantum entanglement affects computation and vice versa, using tools from both mathematics and computer science.
The goal is to create faster optimization algorithms, which are needed to solve computational problems. Many problems that are difficult to solve with classical computer science techniques can be simplified with quantum mechanics, which is strange! Usually quantum makes things more complicated but here, it can help us prove things that you can't prove otherwise.
How did you come to quantum?
I finished my undergraduate degree in software engineering at the University of Tehran in 2011—which makes me a very old PhD student!—but I was always interested in theoretical computer science. From Iran, I moved to Canada for a master's in theoretical computer science but ended up working for five years as a software engineer.
I completed a second master's in quantum information at the University of Waterloo, though I didn’t have the time then to do research. But I knew that’s what I wanted to do.
How does being a theorist compare to being a software engineer?
These days, I get to decide what I think about, and I cannot put a value on that autonomy. You have a much more comfortable life as a software engineer, but this is much more fun intellectually.
What was the main finding of the paper you’re discussing at QIP?
In nonlocal games, researchers assume Alice and Bob are totally separated; this reflects what’s called the tensor product model of quantum mechanics. Another model rejects that assumption and says Alice and Bob are indistinguishable as part of one single larger space.
Our work shows that, perhaps counterintuitively, it’s easier to perform certain calculations across this larger space than smaller ones, which may help researchers find better optimization algorithms.
How are you preparing for the plenary?
By stressing about it! I don’t love public speaking, but this time I’m going to be prepared before I leave.
Why do you say that?
I gave another QIP talk during the first year of my PhD, and I made the mistake of not preparing before I traveled. The conference was in China, and when I arrived I realized I couldn’t access the websites I needed. To make matters worse, my talk was on the last day. Instead of enjoying Shenzhen, I was just stressed out all week.
Advice to others interested in quantum computer science?
Quantum computer science is very different from classical computer science. I love it exactly because of that, but it involves continuous mathematics, with lots of analysis and group theory and algebra—you get into territories well beyond undergraduate math. You'll definitely need to learn a lot of mathematics, but take your time with courses to learn all the different facets of quantum mechanics.
How was moving mid-PhD?
My courses from Toronto transferred, so fortunately it didn’t change much for me academically. New York has just been a lot of fun. When I first came, I saw the Seinfeld restaurant and I was like, “This is where I want to be!”
What do you like to do in your free time?
I am a huge fan of triathlons and running. One reason I love New York so much is the incredible running culture. I ran my first NYC Marathon last year, and I’m looking forward to the next one!
How does endurance training help your science?
It gives me so much discipline. I can draw a lot of parallels between being challenged by mathematics and running, swimming, and biking long distances.
I spend a lot of time thinking about math while training. Last summer, I would go on long bike rides to upstate New York and out in Pennsylvania. I enjoy living in the city, but I just love those bike rides—they really allow you to think.