Alex Waldherr currently attends the final year at a higher technical college for chemical engineering. She’s passionate about science, especially physics, programming and chemistry. In 2018 she finished the Viennese Physics’ Olympiad best girl, and has worked with the department for material chemistry of TU Vienna and as a developer at AF-Institute Vienna (AI and data management for the medical sector). Her passions generally evolve around the tiny things that matter – a beauty hidden almost invisible in everyday life whether it is now microbiology, genetic code or quantum computers. In autumn 2019 she helped with research at CERN Geneva and was invited to speak at the FifteenSeconds-Festival and introduced quantum computing at WeAreDevs Vienna. In chemistry, quantum mechanics have a long tradition – but how can we harness intriguing calculations and maybe change whole industries? Practical concepts of quantum computing still need lots of research and curiosity. Here’s an interview we did with her.
What do you think how much time will it take until quantum computers can take over correct cryptography?
We may see serious attempts to break encryption (which can even be blockchain) in a few years. However, companies providing quantum services have boundaries set to limit everyday users from being able to do so widespread. The entry level to “hacking” with quantum is much higher.
The interesting topics in my eyes are projects, which strive to utilize quantum to transfer secure keys. The genius thing about quantum is that you can see in the result whether somebody listened in (Bell’s inequality)! A cool experiment is the Micius satellite – coordinating the first intercontinental transmission of proven secure quantum keys aka entangled photons betweeen ground stations in China and Austria (Graz).
So, talking delicate data businesses like telecommunication or banking it may get very important to establish their own satellite-to-ground transmission lines and they already invest. For other things there still seem to be other secure encryption systems e.g. lattice based.
Do you think smaller than 5mm transformators could be used to quantum computing as they do quantum jumps?
Maybe. Quantum annealers, a limited quantum computing approach, exhibit quantum jumps to find minimum energy states. But for universal quantum gates the question is whether flow control can be guaranteed. Random, uncontrolled quantum jumps do not make a quantum computer.
I can’t say if today’s approach of superconducting circuits really becomes the standard built, but there is plenty of room at the bottom to manipulate qubits! Hardware of quantum computers is under development and new approaches are tested constantly. Some are nicely summarized here.
5mm transistors will probably not become quantum components. However, some novel silicon quantum computer approaches, which manipulate doped atoms, show stable and promising alternatives to current superconducting technology.
Is the concept of digital even relevant to quantum computing?
Yes, because data storage will still be 0 or 1. Quantum computers act more like multiple experiments exploiting vast infinities. Like in scientific research it is good, that you can ask many questions during these calculations, but in the end, you will have a result of works this way or doesn’t work this way.
And interestingly what does quantum stand for? Quantised. So even if the calculations utilize the whole Bloch sphere, not every operation may be successfully applied to the system because we have to stick to single photons and energy bunches.
In your opinion, using so many filters doesn’t differentiate the answer that we could receive?
Quantum states are so fragile, that there is hardly any change to get replicable calculations if environmental influences are not filtered out as best as possible. In contrast to these distinct calculations, I support the idea of grasping the whole system as a more complex, “natural” and intertwined unity when it comes to experiments with biological systems or chemistry.
What do you think of google’s claim about quantum supremacy?
It is an awesome step forward to prove that quantum computing investigations are not non-sense or useless! Only, I believe we need to leverage the potential of achieved algorithms to our ultimate goals e.g. sophisticated protein folding problems. Compare – current AI is still far away from proposed general AI in many fields.
But even you can be better in quantum computing than classical computers: apply the Hadamard gate to as many qubits as you can and measure the outcome. Congrats! You made the most random 01-number-string possible. 😊
How far it is from today’s accessible solutions, e.g Digital Annealer from Fujitsu?
It definitely depends on the application. If we look at the most widely used quantum annealers (> 1000 qbits) by D-Wave, optimization and sampling problems, important for e.g. machine learning, can already be scaled up much more efficiently. These rely on finding the minimum energy of an initialized quantum system.
The goal are gate model quantum computers and the ability to change and manipulate these energy states. The difference is very well explained in the video below. IBM has some ten qubits of superconducting circuits to calculate with. And in these systems, it especially depends on coherence time and error correction of the hardware how far we can go. Besides superconducting circuits, ion traps or silicon-based options may become a valid option. We are currently figuring out, what is the way to go. All of them are already in research or operation.
How can quantum computers help me in everyday life?
Think medicine, finance, science! Today’s computers are readily available to end-consumers and we all carry smartphones with us. In my eyes, quantum computing should be more seen as delicate experiments, which facilitate much more powerful calculations and optimizations as today’s technology.
So, you might not carry around a quantum computer in the future, as you also do not carry around a laboratory. But if the research and development is done correctly, quantum computers will help us understand natural mechanisms much better. In healthcare they help predict the body and molecule interactions much better and find new drugs. We may even get faster database search, much more accurate weather forecast or (stock) market predictions.
What do you think, how much resources and time needed to check the correctness of a qbit? 100’s of qbit?
Error correction be done on classical bits after storing parallel measurements. On the software side, the fewer gates and the shorter an algorithm, the better the results on quantum computers. Wherever possible, intermediate results can also be stored on regular bits, if the following steps do not call for superposition.
There are 1:3 ideas, but more 1:1000 currently. Decoherence time is a very big problem and in the end quantum computers aspire to work with Schroedinger cats. “Some observer” can easily wreck the calculations or thermal turbulences collapse the quantum state (this is why ion traps are more stable than superconducting circuits). By time this will improve. I like to compare it to the reliability of vacuum tubes in early computers and our billions of highly reliable transistors nowadays.
When would it be possible to use quantum calculations in real life on your mind?
Quantum annealing is already used for big optimization problems, but gate model quantum computers propose the most desirable use cases: In my eyes, chemistry and healthcare will be very interesting fields to investigate. Especially protein folding and hormone interactions are poorly understood and cannot be calculated precisely nowadays.
Speaking of chemistry, the biggest molecule with precise energy states calculated on a quantum computer is beryllium hydride. It is still way to go from 3 atoms to many hundred! However, the variational quantum eigensolver approach used promises a method which scales up.
Applications are coming and I would give reliable hardware some more years to be implemented for finance and business problems. Distinct calculations like Grover’s algorithm (fast search) and Shor’s algorithm (break RSA encryption) can already be run on the IBM machines.
Did you already use some quantum algorithms? If so, which ones?
Of course! Hands-up for hands-on learning right?
I started with some basic algorithms where nice instructions were provided by qiskit (e.g. Bernstein-Vazirani to guess secret numbers or Simon’s algorithm) and of course, as on-going chemist I had to go over the variational quantum eigensolver for BeH2. The
https://quantumalgorithmzoo.org/ is a very nice resource to get creative and learn a lot!
For me, it was very important to visualize and grasp the matrix calculations. So in the beginning I built very simple circuits (one or two gates) and plotted the results as histograms and on the Bloch sphere.
Fun fact #realworldapplication: For our chemistry quiz-advent calendar in school I quickly programmed the simplest quantum circuit to randomly pick the winner out of all correct answers. 😊
Can we defeat the AI uprising with Quantum computing?
What to defeat, when AI is not necessarily bad? On large datasets quantum will become much more powerful than current computers – yes. But scientist already utilize this for good and learn to better train AI with quantum computers.
AI improves, if intelligence and the brain is better understood. To build better quantum hardware, one learns to understand the behavior of elementary particles much better. These quantum behaviors might be crucial in biological systems like the brain … and eventually understanding quantum could correlated to understanding neurons better and enhancing AI!
Like for everything: Quantum computing is a tool. It is up to us, to use it responsibly. A smartphone can connect you to the biggest library in the world or waste your time. Quantum is evolved by a highly curious and supportive community. Let’s continue this way.
Where to start?
- Quantum computing for everyone – Chris Bernhardt
- IBMQ – introduction: https://www.youtube.com/playlist?list=PLOFEBzvs-Vvp2xg9-POLJhQwtVktlYGbY
- Ion traps: https://www.youtube.com/watch?v=_dvByciEwT0
- Google: https://ai.google/research/teams/applied-science/quantum/
- QISKIT Slack channel
You can hear Alexandra’s talk at our upcoming WeAreDevelopers World Congress 2020. See you in May!