Thoughts & Experience from Womanium Quantum 2023
I ran into Womanium Quantum by chance and decided to give it a try. The virtual program is condensed, which touches on a variety of essential quantum topics, e.g. quantum computing, sensing, key distribution, communication, machine learning, entrepreneurship, etc.
Over the course of two months, I have not only gained new quantum knowledge but also re-learned basic quantum concepts. In quantum information processing, there are two fundamental processes — Qubit Readout & Qubit Control. They manipulate qubits and extract information from qubits. Specifically for qubit control via microwave pulses, there is something called pulse envelope, created by an arbitrary waveform generator (AWG). The pulse envelope is applied to a microwave signal that controls the qubit and generates qubit rotations. There is another term called gate fidelity, which is to measure how accurate the performance of a quantum gate operation is on a quantum system. In qubit control, we want to maximize gate fidelity to achieve precise and accurate quantum computations. The quantum gates, either single-qubit or two-qubit gates, manipulate the quantum states of qubits. The errors in gates have a significant impact on quantum algorithms, therefore, we need to minimize the errors in these gates, which results in higher gate fidelity.
I also got to know about various hardware modalities that industry and academia pursued, such as superconducting qubits (IQM, Rigetti), trapped ion qubits (Quantinuum, IonQ), photonic qubits (Xanadu, PsiQuantum), etc. For superconducting qubits, they are fabricated into a packaged chip for wiring connection and later put into a dilution refrigerator (Bluefors) to function in an extremely cold environment (10 mK!). Then the system will be hooked up with commercially available hardware to control superconducting qubits (Keysight Technologies, Zurich Instruments), controlled by a normal PC. This series of microscopic levels of a qubit device shows a quick breakdown of what the fabrication of superconducting qubits entails. You will see the Josephson junction, an overlapping region that links two superconducting materials.
Throughout the program, one of the topics that I want to learn more about is Quantum Key Distribution (QKD). I first came across QKD while attending ATxSG 2023 with my manager and learned that Singapore recently launched the National Quantum-Safe Network Plus (NQSN+) to integrate QKD into communication networks across the country. The goal of QKD is to build resistance and remain secure against quantum computer attacks. There are multiple quantum encryption protocols (BB84, E91, BBM92, etc.) to encode information, like what the well-known classical cryptographic protocol RSA does. Particularly in BB84 protocol, it is built upon one of the quantum principles — no-cloning theorem, which means you cannot create an exact copy of an unknown quantum state. The theorem comes into play to help QKD identify a hacker's attempt to eavesdrop and obtain information during a key exchange process. If someone intercepts secretly at any point during the transmission, then it will change the qubits’ state, leaving a detectable trace.
Another interesting field that the program touched on is quantum machine learning (QML). Similar to classical machine learning, QML is about “finding patterns in data”. Apart from classification tasks, QML models can perform other tasks such as regression, clustering, solving optimization problems, etc. QML involves using quantum algorithms (Grover’s Search Algorithm, Quantum Support Vector Machines, Quantum Neural Networks, etc.) to run on quantum computers or simulators. Side note, I previously thought that a quantum computer is that massive octopus-like machinery shown below. It turns out that the shiny machine is called a dilution refrigerator, which is an integral part of a superconducting quantum system, to protect the quantum chip from radiation.
One last thought: Despite being packed, this two-month program showed me how fascinating the quantum field is, from the backend to frontend.
